Business Analytics in Healthcare Operations and the Use of Mobile Applications for Decision Making by Health Professionals Gale

Business analytics in healthcare operations and the use of mobile applications for decision making by health professionals

Galetsi Panagiota

Ph.D. Thesis

submitted to the School of Humanities, Social Sciences and

Economics, International Hellenic University for the degree of

Doctor of Philosophy

Thessaloniki, 2020

To my beloved husband Thodoris and my children Thanasis, Charalampos and Christos

Acknowledgments

I would first like to express my gratitude to my Thesis supervisor Dr. Korina Katsaliaki for the patient guidance, encouragement and advice she has provided throughout my time as her Ph.D student. I have been extremely lucky to have a supervisor who cared so much about my work, and who responded to my questions and queries so promptly. My sincere gratitude also extends to the other members of my Ph.D. committee, Prof. Sameer Kumar, and Dr. Lampros Stergioulas for the assistance they provided at all levels of my research project. I would especially like to thank Prof. Sameer Kumar who was involved in this research project and his valuable guidance and contribution enabled this research to be successfully conducted and submitted to peer reviewed journals. I must also thank the International Hellenic University for supporting of my research and all the staff of the University for being extremely supportive throughout my studies. I am particularly appreciative to my fellow doctoral students, Antonios Chantziaras and Kleopatra Koulkidou, for the suggestions, feedback and cooperation during this journey.

Finally, I must express my very profound gratitude to my parents and to my family for providing me with unfailing support and continuous encouragement throughout my years of study and through the process of researching and writing this Τhesis. This accomplishment would not have been possible without them. Thank you.

Abstract

The emergence of powerful software in healthcare has created conditions and approaches for large datasets to be collected and analyzed which has led to informed decision-making towards tackling health issues. Big Data Analytics (BDA) in Healthcare, otherwise Health Analytics, express the analysis methods of the wide amount of electronic data related to patient healthcare and well-being that are very diverse and difficult to be measured by traditional software or hardware. This PhD Thesis includes two parts. The first part presents a systematic review using PRISMA of the research activity in Big Data Analytics (BDA) in the field of health and demonstrates the existing knowledge. The objective of this profiling study is to discuss this scientific field through related examples and to inform researchers about the nature and magnitude of the technological innovations in health information analysis tools, its influence, and where and how further material could be searched. With reference to the resource-based view theory this Doctoral Thesis has focused on how big data resources are utilised to create organization and social values, discussing the classification of big data types related to healthcare, the associate analysis techniques, the platforms and tools for handling big health data and the future aspects in the field. In recent years a large number of mobile health applications (mHealth) have been developed for medical practitioners and students that use apps and other digital technologies as part of their practice training and education. These trends have created a new social context in clinical diagnosis process based on technology innovation in the field. Inspired by this, the Second Part of the Doctoral Thesis aims to review available mHealth apps addressed to medical professionals and students designed to assist in the diagnosis process and explore the multiple dimensions of the research subject. Based on three conceptual frameworks, different approaches have been taken intending to investigate the social dimension of the intention of integration of mHealth innovation in the diagnosis process, explore the ethical challenges related to their data governance and reliability and explain how the specific consumers’ behaviour is affected by certain app characteristics and attributes. A special emphasis is placed on mHealth apps that use artificial intelligence and a future agenda is provided for the development of new apps for medical professionals with the use of responsible innovative methods. To investigate the relationships between app quality, downloads, features and users’ ratings multiple linear regression statistical analysis was used. The Thesis contributes to the information

4 systems and operations management research, while empowers mobile health literature providing a better understanding of the matter. This study also provides a multi-layered analysis and aims to assist health professionals and health policy makers with a better understanding of how the development of an innovative data driven strategy can improve public health and the functioning of healthcare organizations but also how such a strategy creates challenges that need to be addressed in the near future to avoid societal malfunctions.

Keywords: Big data analytics; systematic review; bibliometrics, citation analysis, operations research techniques; machine learning; information and communication technologies (icts), organizational and societal values; mobile applications; health diagnosis; artificial intelligence; mHealth apps, consumer behaviour; regression analysis; mobile apps for professionals

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or of any other university or institute of learning.

The author of this Thesis (including any appendices and/or schedules to this thesis) owns any copyright in it (the "Copyright") and she has given the International Hellenic University the right to use such Copyright for any administrative, promotional, educational and/or teaching purposes. Copies of this thesis, either in full or in extracts, may be made only in accordance with the regulations of the Library and Information Centre of the International Hellenic University. Details of these regulations may be obtained from the Librarian. This page must form part of any such copies made. The ownership of any patents, designs, trademarks and any and all other intellectual property rights except for the Copyright (the "Intellectual Property Rights") and any reproductions of copyright works, for example graphs and tables ("Reproductions"), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property Rights and Reproductions cannot and must not be made available for use without prior written permission of the owner(s) of the relevant Intellectual Property Rights and/or Reproductions. Further information on the conditions under which disclosure, publication and exploitation of this thesis, the Copyright and any Intellectual Property Rights and/or Reproductions described in it may take place is available from the Dean of the School of Economics, Business Administration and Legal Studies.

Contents

Acknowledgments…………………………………………………………………….3

Abstract……………………………………………………………………………….4

Declaration……………………………………………………………………………6

PART A………………………………………………………………………………12

CHAPTER A. 1 1. Introduction ...... 12

CΗΑPTER A. 2

2. Literature Review 2.1 Previous Literature ...... 16 2.2. Research Framework ...... 17

CHAPTER A. 3

3. Materials and Methods……………………………………………………………21

CHAPTER A.4……………………………………………………………………….26

4. Results……………………………………………………………………………..26 4.1. Bibliometric Analysis and Descriptive Results ...... 26 4.1a. Years of Publication, Country of Origin, Source of Publication, Subject Areas and Authors’ multi-disciplinarity ...... 26 4.1b. Popular authors and co-cited authors, affiliations and departments ...... 30 4.1c. Citation and Co-citation analysis based on the bibliographic data and the most popular keywords found in the 804 articles ...... 32 4.2. Content Analysis Results ...... 36 4.2a. Medical Specialties ...... 37 4.2b. Stakeholders of Big Data Analytics in HealthCare ...... 39 4.2c.Research approach ...... 42 4.2d. Nature of Analytics ...... 43 4.2e. Types of data ...... 44

4.2f Big Data Techniques ...... 45 4.2g.Big Data Capabilities ...... 47 4.3. The association of the selected analytical techniques with the data types and capabilities ...... 50 4.4. Gained values/capabilities from the use of BDA in the health sector ...... 51 4.5 Types of social value creation in healthcare ...... 54 4.6 Challenges from the implementation of BDA in healthcare industry ...... 60 4.7. The use of Machine Learning in the health field ...... 64 4. 8. Future perspectives as derived through the article content analysis ...... 65 4.9 Big Data Analytics ...... 69 4.10. Recent Examples of Big Data Analytics Tools for use in Healthcare ...... 71 CHAPTER A. 5 ...... 75 5. Discussion and Conclusions…………………………… ...... 75 CHAPTER A. 6 ...... 82 6. Limitations ...... 82 CHAPTER A. 7 ...... 84 7. Future research ...... 84 PART B ...... 86 CHAPTER B. 1 ...... 86 1. Introduction ...... 86 CHAPTER B. 2 ...... 92 2. Literature Review ...... 92 CHAPTER B. 3 ...... 96 3. Methodology ...... 96 CHAPTER B. 4 ...... 99 4. Conceptual frameworks and results ...... 99 4. 1a. 1st Conceptual framework. “Communication Privacy Management” ...... 99 4.1b. Results ...... 102 4.1c. App features and related challenges of digital healthcare ...... 102 4.1d. Health app types ...... 106 4.1e. mHealth apps using Artificial intelligence ...... 112 4. 2a. Conceptual Framework. “Normalization Process Theory” (NPT) ...... 116 4.2b. Results ...... 119 A. Implementation stage: Health apps’ categories ...... 119 B. Embedding stage: App features analysis indicating app trustworthiness ...... 123 C. Integration stage: Quality evaluation ...... 127 C. Integration stage. Health apps popularity and engagement ...... 128 4. 2c. Results from statistical tests: Comparison of Means ...... 129 4.2d. Results from statistical tests: Correlations ...... 131

4.3a. 3rd Conceptual framework ...... 133 4.3b Hypothesis testing ...... 136 4.3c. Regression model variables ...... 138 A. Depended variable ...... 139 B. Independent Variables ...... 139 4.3d. Regression analysis results ...... 145 A. Descriptive statistics of variables ...... 145 B. Regression analysis results ...... 146 CHAPTER B. 5 ...... 148 5. Agenda with valuable insights about mHealth apps for professionals ...... 148 5.1. mHealth apps: What exists ...... 148 5.2 Suggestions for new mHealth apps with increased usefulness ...... 150 5.3 Mitigation of Privacy and Security Challenges ...... 151 5.4 Mitigation of Reliability Challenges ...... 155 CHAPTER B. 6 ...... 157 6. mHealth apps: What is needed ...... 158 6.1 Building Useful and Trustworthy, Quality mHealth apps ...... 158 6.2 Managerial Implications ...... 162 CHAPTER B. 7 ...... ………………………………………………………………..163 7. Discussion and Conclusions…………………………………………… ...... 164 CHAPTER B. 8 ...... ……………………………………………………………….168 8. Limitations and Future Research…………………………………………...... 168 CONCLUSIONS OF THE TWO PARTS OF THE Ph.D THESIS………...... 169 REFERENCES……………………………………………………………...... 171 APPENDIX……………………………………………….………………...... 205

Tables and figures

PART I Figure 1: Schematic Research Framework …………………………………………….20 Figure.2: Scheme of review methodology ……………………………………………..23 Figure 3: Dataset identification process …………………………………………….....26 Figure 4: Distribution of articles per year ……………………………………………..29 Table 1a: Top 10 most popular countries of authors’ origin …………………………..30 Table 1b: Top 10 most popular journals in the dataset with publications in Health Analytics ……………………………………………………………………………….30 Table 1 c: Most popular subject areas………………………………………………….30 Table 2: Top 5 most co-cited authors………………………………..…………………32

Table 3: The 10 most popular author affiliations in the dataset…………………..……32 Table 4: Top 10 most cited articles (as of Nov 2018)………………………………….34 Table 5: Top 10 most co-cited articles……………………………………………..…..35 Figure 5: The most frequent keywords of the 804 articles …………………………....37 Table 6: Classification Framework…………………………………………………….38 Table 7: Distribution of articles to Medical Specialties………………………………..39 Table 8: Stakeholders of Big Data Analytics in Healthcare……………………………41 Table 9: Research Approach and indicative research examples……………………..…43 Table 10: Nature of Analytics and indicative research examples…………………...…45 Table 11: Types of data and indicative research examples………………………….....46 Table 12: Big Data Techniques and indicative research examples ……………………47 Table 13: Big Data Capabilities and indicative research examples…………………….50 Table 14: Classification of dataset articles based on the analytical techniques by data type…………………………………………………………………………………….51 Table 15: Created Values from the use of BDA………………………………………54 Table 16: Classification of dataset articles based on the analytical techniques by created value………………………………………………………………………………….…55 Table 17: Types of Value Creation…………………………………………………..…60 Table 18: Challenges from the implementation of BDA in healthcare……..……….…65 Table 19: Presentation of BDA in healthcare future aspects…………………………..69 Table 20: Description of new tools related to Big Data Analytics in Health………….74 PART II Figure 1: Methodology process……………………………...... 99 Figure 2: Conceptual framework “Communication management privacy theory”…..103 Table 1: Feature relevance……………………………………………………………104 Table 2: Health apps types……………………………………………………………108 Table 3: Frequencies of features per app type………………………………………..113 Table 4: Artificial intelligence apps’ Features…………………………………….….116 Table 5: Artificial intelligence mHealth apps and outcomes…………………………117 Table 6: The Mechanisms of NPT of mHealth in Medicine………………………….119 Figure 3: Conceptual framework………………………………………………..……120 Table 7: Health apps’ categories…………………………………………………...…121 Table 8: mHealth Features’ Analysis…………………………………………………125 Table 9: MARS clusters’ scores………………………………………………………129 Table 10: Description correlation variables…………………………………………..131 Table 11: Hypothesis testing………………………………………………………….132 Table 12a: Correlations of the conceptual model variables and individual app Features……………………………………………………………………………….135 Table 12b: Correlations of the NPT conceptual model elements (SUM of app features)………………………………………………………………………….……136 Table 13: The study’s conceptual constructs of TAM & TRA………………...……..136 Figure 4: Conceptual framework with research hypotheses of mHealth apps adoption intention from professionals…………………………………………………………..137 Table 14: Comparison of our study with similar studies on the hypotheses testing….139 Table 15a: mHealth app capabilities (PU) based on app content and features ……...141 Table 15b: Perceived Ease of Use (PEU) features………………………………...….142 Table 15c: Perceived Trust (PT) features……………………………………………..143 Table 16: Description of regression analysis variables…….…………………………145 Table 17: Correlation coefficients of individual variables……………………………146 Table 18: Descriptive statistics and frequencies of model variables………………….147

Table 19: Hypothesis analysis results ……………………………………………...... 148 Table 20: Pearson Correlations of regression model variables...... 148

PART A

BIG DATA ANALYTICS. A SYSTEMATIC REVIEW OF THE LITERATURE

CHAPTER A. 1

1. Introduction

The healthcare industry is highly data intensive and could use interactive dynamic big data platforms with innovative technologies and tools to advance patient care and services (Ali et al, 2018, Carvalho et al, 2019). Healthcare industry manages every day a wide amount of data from clinical and operational information systems, such as Electronic Health Records (EHR) (Brooks et al, 2015) and Laboratory Information Library Systems (LIMS) (Groves et al, 2013). To illustrate data volume magnitude, the health data explosion from 500 petabytes in 2012 (Feldman et al. 2012) will reach 163 zettabytes in 2025and practitioners are developing new applications in order to assist healthcare stakeholders to increase opportunities for a greater value (Groves et al, 2013). Business Analytics include the techniques, technologies, systems, practices, methodologies, and applications for the analysis of the vast amount of data and help organizations better understand its business, market, and make timely decisions (Chen, Chiang & Storey, 2012; Wamba et al. 2017; Srinivasan & Swink, 2018; de Camargo Fiorini et al. 2018). Big Data Analytics (BDA) in healthcare involve the methods of analysing this wide amount of electronic data related to patient healthcare and well- being that are so diverse and difficult to measure by traditional software or hardware. There are various forms of health data such as, clinical and lab data, medical notes, machine generated data from medical equipment or from sensors at home monitoring, health services financial data, hospital bills, literature data from medical journals, social media posts blogs in health subjects, etc. These data may be available internally in health services (e.g. EHR, LIMS) or come from external sources (e.g. insurance companies, pharmacies, government) and could be in a structured format (e.g. tables

12 with laboratory results) or unstructured (e.g. text of medical notes in EHR) (Raghupathi & Raghupathi, 2013). The term “analytics” pulls together Management Information Systems (MIS), Operational Research (OR), and statistics. It describes the combination of Business Intelligence reporting and descriptive analysis, advanced statistical methods in data mining and forecasting, and other OR methods such as optimization and simulation (Gorman & Klimberg, 2014). OR has been benefited from big data processing and leveraged analytics for advanced problem solving and better decision making (Turaga, 2018). The practice of big data via advanced modeling techniques creates great opportunities (Hazen et al., 2018). Analytics are categorized as descriptive, predictive and prescriptive. This taxonomy refers to the nature of the analysis techniques and the information gained. Descriptive analytics aim to identify problems and trends in existing processes and functions, predictive involve the use of mathematical algorithms to discover predictive patterns and prescriptive determine decisions based on certain objectives for improving performance (Wang et al., 2016). The term BDA includes two perspectives: big data and analytics (Wang et al., 2016).Big data are recognised by four characteristics, the so called 4Vs: volume - due to the incredible size of data, velocity - due to the rapid and real-time accumulation, variety - due to the differentiated formats (structured, unstructured and semi-structured) and veracity, which refers to reliable data (Gandomi & Haider, 2015). The methods of BDA refer to techniques such as forecasting, optimization, simulation, and others which assist decision-making and provide insights to managers and policy-makers (Doumpos & Zopounidis, 2016; Duan et al. 2019). Therefore, the term “big data” does not characterize only the volume but it also highlights the analytical workloads associated with some combination of data variety and velocity, as well as volume (Ferguson, 2012). Thus, health big data are associated with advanced analytics and their potential to analyze complex data to improve operational potential and decision making. As a result, IT professionals constantly develop new applications with big data capabilities to help healthcare stakeholders increase value. One of the most used computing platforms for processing big data in general but in healthcare too, is Apache Hadoop (De Silva, Burstein & Jelinek., 2015), a software framework that allows for the distributed processing of large data sets across clusters of computers which allows for storage, refinement and analysis of vast amount of data. To this extent, remarkable is

13 also the use of the most recently developed analytics techniques of “machine learning”, a data driven computational approach using algorithms that are capable to recognize patterns and make predictions (Gruebner et al. 2017) and “visualization”, which create tables, images, diagrams and other intuitive display methods to understand data (Cheng & Zhang, 2014). Organizations also develop infrastructure with big data capabilities to help improve manager decision-making (Groves et al., 2013). It is said that 80% of the growth of information and communication technology will be about cloud services, big data analytics, mobile technology and social media technologies (Andreu-Perez et al., 2015). In parallel, the emergence of powerful software for data analytics have created conditions and approaches for large datasets to be analyzed and utilized for the social good and has improved the quality of decisions towards tackling global health issues, such as, disease prevention and leverage, public health surveillance, timely provision of essential medical services in emergency conditions. Sociologists argue that data collected from groups of individuals using digital databases, e.g. social media groups discussing health issues, make visible aspects of individuals and groups that could not otherwise be perceptible. Scientists can make assumptions from a vast range of details derived from such diverse sources (Lupton, 2014 debating about issues raised regarding high public concern about big data ownership and health data privacy protection regulation (Mooney & Pejaver, 2018). The effort of the interested parties (clinicians, patients, healthcare organizations, researchers, etc.) to address issues in order to harness and maximize the potential of big data analytics in healthcare is noteworthy. It is argued that only a small part of the available amount of data is currently captured, stored and organized so that it can be processed by computers and analyzed for useful information. Therefore, healthcare organizations need more efficient ways to manage them (Raghupathi & Raghupathi, 2014). Healthcare organizations are also facing challenges, such as reduced fee schemes, demands for faster turnaround times, diminished numbers of qualified technologists, etc. (Horowitz et al. 2005).To meet these challenges, hospitals and healthcare systems rely more and more on automation and management of those data that come from clinical and operational information systems such as Electronic Health Records (EHR) and Laboratory Information Management Systems (LIMS) (Ward et al.

2014).It is suggested that the management of healthcare data could be beneficial with regards to fraud detection and prevention, production of effective drugs and devices for patients’ well-being and improvement of public health surveillance and speed of service (Raghupathi & Raghupathi, 2014). Despite these important observations, scholars have made little progress in the formulation of an approach to broadly identify organizational and social values and the related challenges of big data analytics in health. Although the published material in the field of BDA in the healthcare industry has rapidly increased in recent years (Ivan & Velicanu, 2015), there is yet no study which shows an all-encompassing organizational and social impact - positive or negative - of BDA in healthcare. “Challenging the dominant techno-utopian approach evident in digital health discourse” has been characterized as critical for future research (Lupton, 2014). In the first part of the PhD Thesis, it has been conducted a systematic literature review study to map the scientific field. Therefore, text mining methods have been used to identify the most prominent journal articles and present the described BDA tools according to their specialty, the types of data they process and the capabilities they offer. The theoretical framework draws upon the resource-based theory and aims to identify the created organizational and social values along with a special interest in the used data, the applied analysis techniques and the information technology innovations. It also aspires to fulfill the need for a deeper analysis of the “state of the art status in the subject field” in order to connect the technological accomplishments of BDA in healthcare with the achieved values and the call for future work. Lastly, it aims to outline a critical research agenda in order to conceptualize and explore the impact on health organizations and the society arising from these technological advances (Loebbecke & Picot, 2015). This paper contributes to the global literature because it attempts to classify basic analytic terms followed by representative results and examples. As a result of an increasing interest in health analytics, the synthesis of the current literature, through a theoretical framework and the presentation of its outcomes, is beneficial to researchers and the industry itself.Findings from this work will be of interest to public health practitioners, policy-makers, and researchers that are interested in population health improvement, medicine and social science as well. It will also assist organizations not to lose the chance to understand value creation opportunities due to the unstopped hype of business analytics and data science (Hindle & Vidgen, 2018). This study will allow

15 governments, groups and health policy makers to gain a better understanding of how the development of a data driven strategy can improve public health, reduce the incidence of disease and better inform the populations (Bram et al., 2015), but also how such a strategy creates challenges that need to be addressed in the near future to avoid societal malfunctions. To summarize the existing knowledge, this study, firstly, has recorded the production of articles between 2000 and 2016 and then provided an analysis of the publication activity, to offer unique information through targeted examples in order to explain the use of ‘Big Data Analytics in Healthcare”. The structure of the first part of the Thesis is organized as follows. The next section presents the followed methodology continuing with an extended theoretical framework with an outline of the resource-based view that describes the path to value of big data analytics in healthcare. Chapters 3 and 4 describe the BDA values and issues respectively and their organizational and social impact, presented through indicative research examples. The next chapter presents the future perspectives of the analysis of health big data and the first part of the PhD Thesis concludes with the study limitations and the discussion.

CΗΑPTER A. 2

2. Literature Review

2.1 Previous Literature

Searching in the international literature for systematic reviews in big data analytics in healthcare, it is notable that in the existing literature there is a lack of holistic bibliometric approach towards the characteristics of these publications. There is a number of very informative papers which profile research in the field of BDA (Chen, Chiang & Storey, 2012; Peng et al. 2017, etc.), however only a few focus on the use of big data in the health sector, such as the study of Wamba et al. (2013), which is a review of 215 papers about “RFID-enabled healthcare applications” and the study of

West, Borland & Hammond (2014) that examines 18 articles on the issue of “innovative information visualization of electronic health records”. These studies are of limited spectrum in terms of the number of papers analysed and/or the discussed content. On the other hand, there are reviews and profiling papers in OR healthcare which study the use of techniques to solve complex healthcare problems (Jun et al. 1999; Brandeau, et al. 2004; Katsaliaki & Mustafee, 2011). Several studies have contributed in different ways to the understanding of BDA in healthcare. Baro et al. (2015) and Wamba et al. (2015) are literature reviews that discuss the meaning of big data in healthcare. The studies of Raghupathi & Raghupathi, (2013) and Ward et al. (2014) provide a general overview through the analysis of examples in the health analytics area, concentrating in certain aspects of the field. The study of Zhang & Li (2017) reviewed literature for a specialized healthcare domain, in HIV self-management. Wang et al. (2018) identified the relationships among big data analytics capabilities, IT-enabled transformation practices and benefits, using the health sector as their case study. Jakofsky (2017), in his overview, raised concerns to physicians about the pitfalls of analytics reports from large metadata sets in health care. Although the published material in the field of BDA in health is increasing in recent years (Wills, 2014; Ivan &Velicanu, 2015; Thouin, Hoffman & Ford, 2008), there is no published work that provides both a wide bibliometric and content analysis of this material by categorising at the same time the applications, tools and methodologies of BDA in healthcare for understanding how and why these benefits are achieved over time. The high number of publications on the medical field makes systematic reviews valuable to researchers in order to keep pace with the recent developments.

2.2. Research Framework

The last decades, medical scientists rely more and more on automation and cooperate with IT specialists for the creation of new software solutions to manage the vast amount of patient and other related data. Therefore, the health sector is an appropriate application of the resource-based view theory for examining the value chain created from the analysis of the vast amount of data. “Resources”, such as data and IT infrastructure solutions and “activities”, such as big data analysis, are described as the

17 essential mechanisms that contribute to the value creation of organizations (Lim et al, 2018). It is important though for an organization to recognize and understand the factors of data-based value creation to gain competitive advantage and to provide better services. The resource-based view states that firm, by acquiring valuable resources and synthesizing them appropriately, can create unique values/capabilities that provide their competitive advantage (Barney, 1991). This is the most commonly used organizational theory to big data research (see, Gunasekaran et al. 2017; de Camargo Fiorini et al., 2018). The data gathered from IT infrastructure is reported as an important organizational resource for gaining competitive advantage (Jaklič et al. 2018; Mamonov & Triantoro, 2018). Success in business analytics depends on the firm’s ability to simultaneously utilize multiple resources (including data) and capabilities within a business context, and make decisions to deliver a valued output (Vidgen et al. 2017; Srinivasan & Swink, 2018; Dubey et al. 2019a,b). In the case of the healthcare industry, data comes from clinical and operational information systems. Scientists use this data to address healthcare problems (reduced budgets, demand for faster turnaround times, etc.) and to gain value from better decision making.According to the resource-based view, firms gain a competitive advantage by bundling resources into capabilities to create value (Gunasekaran et al. 2017). IT infrastructure is a major business resource for gaining long-term competitive advantage (Bharadwaj, 2000) along with the data gathered from IT infrastructure. Data resources in healthcare such as clinical, patient, pharmaceutical data, etc., must be appropriately processed and analyzed in order to create capabilities translated into business values, which are going to be thorough discussed in section 4.5. Their analysis is based on OR techniques, such as modeling, simulation, machine learning, visualization, data mining and others (Yaqoob et al., 2016; Chen & Zhang, 2014) . These techniques develop models which are fed with raw big data and to cope with their volume and their processing time utilize computing applications, such as Apache Hadoop. These applications allow the distributed processing of large data sets across clusters using simple programming models. The effective use of data analytics tools or models can reach organizations’ “agility” only when there is continuous cooperation of various bundled resources (Ghasemaghaei et al. 2017). These models are useful for interested parties to offer solutions for observed problems based on quantifiable

18 measures and propose alternatives which can lead to improved performance (Katsaliaki, Mustafee & Kumar, 2014). The study aims to identify the use of the big data resources and their analysis techniques and examine the capabilities and values that are created for the healthcare industry (Wamba, Anand & Carter, 2013) in parallel to the positive or negative impact of the same dynamics in society. These values lead to the need of further developments and therefore future research is essential in terms of technological and organizational improvements that big data analytics will bring in health. According to the above description, there is a summary of the research framework of BDA in health in Figure 1.

Figure 1: Schematic Research Framework I

Describing in detail the “path-to-value” of the Fig. 1, the data resources that the healthcare industry needs to appropriately handle in order to create big data capabilities are categorized according to Groves et al. (2013) in (a) clinical data, (b) patient and sentiment data, (c) administration and cost activity data, and (d) pharmaceutical and

R&D data. Yet, for the transformation of data into capabilities, the process of data analysis is required in-between. Based on the literature (Waller & Fawcett, 2013; Chen & Zhang, 2014), the techniques for the analysis of healthcare data are: modeling, simulation, machine learning, visualization, data mining, statistics, web mining, optimization, text mining, forecasting, and social network analysis techniques. In healthcare organizations, the five most important BDA capabilities developed from data resources and appropriate analysis infrastructure are (Groves et al., 2013): (a) “monitoring”, includes efficiencies (using analytical methods) describing “what is happening now;” (b) “prediction/simulation,” provides information about future outcomes (what will happen); (c) “data mining,” involves methods enabling extraction and categorization of knowledge (what happened); (d) “evaluation” that demonstrates methods for testing the performance of analytical techniques or explains the outcomes of the application of BDA (why did it happen?); and (e) “reporting,” includes methods that shape collected knowledge and provide it in an informational form. By analysing data related to the populations’ health, the health sector could develop better health services by reducing costs, producing more effective drugs and devices for the well-being of patients, improving public health surveillance and the speed of service (Raghupathi & Raghupathi, 2014). The healthcare industry is highly data intensive and there has been an increasing role of electronic data to understand and improve healthcare and healthcare analytics has the potential to create more than $300 billion profit every year and as a result there have been significant investments in healthcare technologies such as mobile computing devices, patient sensors, in-home care devices, etc. (Kambatla et al. 2014). Clinical sciences and administration functions must become information-driven disciplines (Burke, 2013) in order to gain the benefits that other disciplines have acquired through the fast analysis of big customer data. Different kind of stakeholders, e.g. patients, providers, researchers, pharmaceutical companies, medical devices and software companies, governments and insurance companies have different expectations from the evolution of healthcare data analytics (Feldman, Martin & Skotnes, 2012). The resource-based view theory asserts that those capabilities create new organizational values, which maintain their competitive advantage. Moreover, through the same process, the study expands the resource-based view following the “path–to value” approach to also identify the positive societal impact. Therefore, this paper also aims to identify both the organizational values and the societal values that follow from

20 harnessing health big data analysis. Both these values lead to future research in order to further capitalize on these values, create new ones and overcome the challenges. Therefore, an additional path-to-value loop is created, and leads to changes as indicated by the literature. According to the above description, the resource-based view is summarized in an alternative “path-to-value” concept shown in the schema in Figure 1 with regard to the societal values and challenges created and followed by the description of the possible future improvements in technology organizations and research, that will be further discussed in Chapter 4.

CHAPTER A. 3

3. Materials and Methods

The analysis of the elements of the research framework is based on the information gained from the synthesis of the existing literature. For conducting this systematic literature review the key principles of systematic reviews (PRISMA) are followed.The methodology approach includes three stages: 1) Input, 2) Processing and 3) Output stage in order to identify the articles that would be most valuable for the research (Figure 3). Stage 1 involves conducting research into the Web of Science®, a database containing quality impact factor journals and Scopus, one of the largest citation databases (Aghaei et al., 2013; Jahangirian et al, 2012; Burnham, 2006).

Figure 2: Scheme of review methodology

Only papers published between 2000 and 2016 has been reviewed since the term “analytics” was first introduced in the late 2000 (Chen, Chiang & Storey, 2012), the term “business intelligence”, which are considered similar, since they both investigate the capabilities of analytical tools in the business processes (Chae & Olson, 2013) has been established after 2000 (Chuah & Wong, 2011) and the term “big data” started appearing in many well-cited publications even later (Davenport, 2006; Akter & Wamba, 2016). Therefore, for the keyword search it has been used the combination of the terms a) “business intelligence” b) “analytics” and c) “big data”, which are the keywords used in many reviews as “unified terms” (Chen, Chiang & Storey 2012; O’ Conell, 2013; Nie & Li, 2011; Duan &Xiong, 2015) and added the term “health*” and its derivatives (symbol“*”) e.g. “healthcare”, “health sector”, health records”, ‘health datasets”, etc. It has also been used the combination of the terms a), b) and c) with the term “medical” e.g. “medical records” “medical data” etc. (Iqbal et al., 2016) and the term “clinical”. These three terms (health, medical and clinical) are mostly used in scientific papers to describe the nature of the data in the healthcare domain and the purpose of analysis and decision-making (Huang et al., 2018; Schnitzer & Blais, 2018; Kim et al. 2018). To avoid bias, it is not has been used in the search more specific terms to describe: analytics, such as “machine learning”, or health, such as “cancer”. Thus, the database has been searched with the following combination of keywords which could be

22 identified in the title, abstract and/or keywords of any published item in order to download the maximum possible number of papers:

“ANALYTICS” AND “HEALTH*” OR

“BUSINESS INTELLIGENCE” AND “HEALTH*” OR

“BIG DATA” AND “HEALTH*”

“ANALYTICS” AND “MEDICAL” OR

“BUSINESS INTELLIGENCE” AND “MEDICAL” OR

“BIG DATA” AND “MEDICAL”

ANALYTICS” AND “CLINICAL” OR

“BUSINESS INTELLIGENCE” AND “CLINICAL” OR

“BIG DATA” AND “CLINICAL”

The articles were selected based on the following inclusion-exclusion criteria agreed by all authors. The dataset is comprised only of: (1) “articles” and “reviews” (2) studies written in the English language (3) studies relevant to the health sector (4) studies relevant to big data analytics To ensure the relevance of the retrieved articles to the term “big data” included it has been also followed further inclusion criteria based on the above three rules:

a. Papers that tested, even with the use of a relatively small sample, the

capabilities of a proposed new technique or technology for collecting, storing or

harnessing a potentially big amount of healthcare data.

b Papers that refer to biomarkers’ analysis (the detection and measurement of

biological properties or molecules in the blood or tissue which indicate either

normal or diseased processes in the body), even if a small sample of patients is

involved, due to the fact that they use a large number and variety of biological

parameters for testing, which overall may lead to the creation of big datasets.

c. Overviews and case study papers which are descriptive of the existence,

benefits, and use of big data and its tools and technologies in the health sector/

industry.

Only articles and reviews written in English were included in the search, for capturing the full information about a specific study and in particular the results which are usually better presented in a full published article. As indicated in Figure 4, 6817 records were retrieved from the initial keyword search in the two databases. After duplicates exclusion, the article pool ended up with 3241 papers. From the first screening, based on the content of the title and the abstract, 1364 out of the 3241 papers have been excluded as they were not deemed relevant either to the health sector or the field of big data analytics. Nonetheless, papers that tested, even with the use of quite a small sample, the capabilities of a proposed new technique or technology for collecting, storing or harnessing potentially big healthcare data have been included in the article pool. In line with this, have also been included papers that refer to biomarkers analysis and although a small sample of patients may be involved, they use a large number and a variety of biological parameters for testing, which overall lead to the creation of big datasets. Finally, in the dataset are included overviews and case study papers which are descriptive of the existence, benefits, and use of big data and their tools and technologies in the health sector. The broadness of the keywords and the variety of the subjects related to the health domain concluded to a dataset that incorporates papers from the areas of information technology, medical, biology, pharmacology and other disciplines. After having completed the text screening of the 1877 articles, a further number of 1073 papers were excluded for the same reasons, leaving 804 articles in the final dataset which were submitted to content analysis using text analysis software (Zhang, Sun, &Xie, 2015; Mittelstadt & Floridi 2016). The selection procedure took place from September to December 2016.

Figure 3: Dataset identification process

The stage II of the research indicates information extraction, topic tracking and categorisation. In order for the information to derive from the unstructured data of these papers, text mining has been applied (Dinov, 2016). Some of the selected categories and subcategories were based on the existing literature and were enhanced with additional groups from the knowledge generated by reading the articles in the dataset. It must be acknowledged here that many of the 804 studies, during the allocation process, were categorized in more than one subcategory. The second step also refers to the outputs of the classification process. Every category of this classification contains several subcategories. The particular categories and their sub-categories, which act as the guide to the dataset content analysis, were inspired by a number of prominent review and overview papers relevant to big data analytics in general (Wang et al. 2016; Wamba et al. 2015; Chen & Zhang, 2014; Groves et al, 2013; Wallet and Facett, 2013). The content analysis and text mining procedure was conducted with the use of NVivo10 software (Woods et al, 2016) and took place from January to June 2017. After reading the full-text of each paper, the relevant section which signifies and explains its link to a sub-dimension was recognised and coded with the use of the software. From the NVivo menu, all relevant papers to a particular sub-dimension can be retrieved to bring up the highlighted information all at once. In many occasions a paper may fall in more than one subcategories of a certain variable.

The descriptive statistics of this dataset are presented in a number of tables and graphs with regards to the sources of publication, authors, affiliations, citations and others. Following, there is a presentation of a co-citation analysis of references that have been cited in the 804 papers ofthe study in order to capture the high impact publication activity of the broader field. VosViewer co-citation analysis and visualization software (Perianes-Rodriguez et al. 2016) has been used to analyze the data retrieved from the ISI WoS and Scopus databases. Finally, Stage III of the research process presents the outputs of the classification process in the form of tables with article frequencies per dimension and indicative research examples per dimension. As the collected sample of the published work is large, it can be considered quite representative of the health analytics field. Therefore, a presentation of some proportional results of this dataset could shed some light on the research that has been conducted thus far in this area.

CHAPTER A. 4

4. Results

4.1. Bibliometric Analysis and Descriptive Results

For the purposes of the research as stated above, in the below subsections there is a demonstration of the descriptive statistics of the publications in the dataset through several tables.

4.1a. Years of Publication, Country of Origin, Source of Publication, Subject Areas and Authors’ multi-disciplinarity

Figure 5 and Tables 1a-c present the publication movement per year, per country, per source of publication and subject areas.

In Figure 5, it is notable that since 2008 there has been no or little publication related to BDA in healthcare. From 2009 until 2013 a bigger publication activity has begun and after 2014 it is observed an explosion of articles. This phenomenal growth has also been mentioned in other reviews (Wang et al., 2016; Baro et al. 2015; Andrew- Perez et al. 2015 etc.). Of course, the terms of the keywords search (business intelligence/analytics/big data) appeared in the literature after 2000 (Chen, Chiang & Storey, 2012) and therefore a time-lag to the wider adoption of the terms from the academic community was anticipated. The country with the greatest number of publications in the dataset (counting the number of authors affiliated with that country) is the USA with 466 published articles, followed by China (67) and the UK (65), as indicated in Table 1a. Overall, the 804 papers are spread in 460 journals. Table 1b shows the ten most popular journals that have published the higher number of articles from the dataset. The “Journal of the American Medical Informatics Association (JAMIA)” is the journal with the biggest number of publications (31), covering articles in the areas of clinical care, implementation science, imaging, education, consumer & public health and policy and holds an Impact Factor (IF) of 3.698, as of 2016. The next journal is the “Journal of Biomedical Informatics” with 24 articles and IF 2.753, which includes studies in the area of biomedical informatics and a special issue on a field related to health analytics “Methods of Clinical Research Informatics” in 2014. The “PLoS ONE”, which is the world’s first multidisciplinary Open Access journal, comes next with 20 publications. “Big Data” follows with 19 publications, and even though it has only been launched in March 2013, it has already 1.239 IF. In the review of Wang et al. (2016), about the trends of Big Data in social science, the “Big Data” journal was a popular publishing outlet as well. The next popular journal is the “Journal of Medical Systems” (16) with IF 2.456, an established journal in its field which has been published since 1977. The majority of the first ten journals are oriented mostly towards health and medical matters and less towards informatics or engineering. The first OR journal that appears in the list is the “Health Care Management Science” with 5 articles, followed by “Interfaces” with 3 articles and a few others each with one paper. In the future, the growth of BDA in healthcare may lead to the launch of more specialized journals of the field. Currently no OR journal has made it to the top 10. Table 1c presents a description of the broad medical research/subject areas (as of Web of Science and Scopus), covered in the investigated papers. In the general area of

“Medicine” all medical specialties, such as oncology, pathology, cardiology etc. are included. “Medicine” has gathered the majority of papers (293) followed by “Computer Science” (262) and Medical Informatics (124). The study continues with the examination of authors’ multi-disciplinarity. Due to the nature of the scientific field that involves different subject areas and scientists from different affiliations, there has been an investigation upon the level of multi- disciplinarity of the co-authors in the selected articles. Hence, the authors’ affiliated departments have been investigated from the authors’ list of each paper. It has also been counted the specialty/discipline of every researcher that contributed to each article. Only in 11% (91) of the articles the authors come from three or more research areas, in 37.5% (302) of the articles, authors come from two disciplines, with the 72% (216 papers) of these coming from medicine and informatics, and in 411 (51%) articles all authors come from only one discipline, including in this number the 51 single-authored papers. This discipline could be medical, informatics, business, engineering or other. In the dataset there are 138 papers with 2 authors, and 64 articles with more than 10 authors. The remaining 551 articles are written by 3 to 9 authors. The most populous article (Allen et al, 2016) includes 101 authors who form a panel of experts on the evaluation of the predicting performance of Alzheimer’s disease by a computational crowd-sourced project.

Figure 4: Distribution of articles per year

Table 1a: Top 10 most popular countries of authors’ origin

Num 466 67 65 50 48 41 21 21 17 17

Publications

Country US CN UK AU CA DE IN KR ES IT

Table 1b: Top 10 most popular journals in the dataset with publications in Health Analytics

Journals N IF/2016 Journal of the American Medical Informatics Association 31 3.698 Journal of Biomedical Informatics 24 2.753 PLoS ONE 20 2.806 Big Data 19 1.239 Journal of Medical Systems 16 2.456 BMC Bioinformatics 14 2.448 Healthcare financial management: journal of the Healthcare Financial Management 11 0.000 Association Health Affairs 10 4.980 IEEE Journal of Biomedical and Health Informatics 10 3.451 Journal of Medical Internet Research 10 5.175 Indian Journal of Science and Technology 10 2.108

Table 1c : Most popular subject areas

SUBJECT AREAS N MEDICINE 293 TELECOMMUNICATIONS/ COMPUTER SCIENCE 262 MEDICAL INFORMATICS 124 HEALTH CARE SCIENCES SERVICES 117 ENGINEERING 81 MATHEMATICAL COMPUTATIONAL BIOLOGY 62 GENETICS/BIOCHEMISTRY MOLECULAR BIOLOGY 53 BUSINESS ECONOMICS/MANAGEMENT 50 PHARMACOLOGY/ PHARMACY 42 BIOTECHNOLOGY APPLIED MICROBIOLOGY/IMMUNOLOGY 40 SCIENCE/TECHNOLOGY/OTHER TOPICS 36 INFORMATION SCIENCE LIBRARY SCIENCE 33 CHEMISTRY 22 RESEARCH EXPERIMENTAL MEDICINE 20 PSYCHOLOGY/PSYCHIATRY 20 NURSING 19 SOCIAL SCIENCES/ OTHER TOPICS 10 ENERGY/ENVIRONMENTAL SCIENCES/ ECOLOGY 6 INSTRUMENTS/ INSTRUMENTATION 5

EDUCATION/ EDUCATIONAL RESEARCH 3 MEDICAL ETHICS 3

4.1b. Popular authors and co-cited authors, affiliations and departments

The study continues with descriptive results about the authors’ characteristics. The researchers with the highest number of publications (5), in alphabetical order, are:  Lee, Sungyoung from the department of Computer Engineering of Kyung Hee University, South Korea, who is not, however, the first author in any of the 5 papers;  Perer, Adam from International Business Machines (IBM), USA, who is the first author in one of these publications; and  Zhang, Yin from the School of Information and Safety Engineering of Zhongnan University of Economics & Law, China, who is the first author in 2 out of 5 papers. In an attempt to get a broader picture of influential authors in other fields related to health analytics, a co-citation analysis of authors has been performed, using the VosViewer visualization software (Van Eck & Waltman, 2017). 21437 unique authors have been identified in the 26988 references of the 804 papers. Table 2 presents the five most frequently appeared co-cited authors.  Chen, M. from the Huazhong University of Science and Technology, China, holds the first place, as he has authored/co-authored 46 papers of the 26998 references, followed by  Breinman, L. from the University of Berkeley with 43 citations and  Holzinger, A. from the Medical University of Graz, Austria with 36 citations.

Table 2: Top 5 most co-cited authors

CITED AFFILIATION Frequency AUTHORS Chen, M. School of Computer Science and Technology, 46 Huazhong University of Science and Technology, Wuhan, China Breinman, L. Statistics Department, University of California, 43 Berkeley, USA

Holzinger, Institute for Medical Informatics, Statistics and 36 A. Documentation, Research Unit HCI, Austrian IBM Watson Think Group, Medical University Graz, Austria Hood, L. Institute for Systems Biology (ISB), Washington, 32 USA Schadt, EE Department of Medical Biochemistry & 30 Biophysics, Karolinska Institute, Stockholm, Sweden.

The most popular affiliations of authors stemming from the 804 papers are presented in Table 3. Stanford University stands at the top with 20 articles, followed by Harvard University with 17 articles. These Universities excel in many fields of science and are ranked in the top 10 universities worldwide for at least a decade now (Times Higher Education – World University Rankings 2010-2019).

Table 3: The 10 most popular author affiliations in the dataset

Number ORGANIZATION/AFFILIATION OF of AUTHORS papers STANFORD UNIVERSITY 20 HARVARD UNIVERSITY 17 MAYO CLINIC 16 UCLA 15 UNIVERSITY OF MICHIGAN 15 UC SAN FRANCISCO 14 UNIVERSITY OF NORTH CAROLINA 13 UNIVERSITY OF PITTSBURG 13 EMORY UNIVERSITY 12 UNIVERSITY OF WASHINGTON 12

Due to the nature of the scientific field which involves different subject areas such as medicine, informatics, engineering etc. (Table 1c) and scientists from different affiliations (Table 3), there is an investigation of the level of multidisciplinarity of the co-authors in the articles of the dataset. Hence, there is also a description of the authors’ affiliated departments from the authors’ list of each paper. There is also a description of the specialties /disciplines of every researcher who contributed to each article.  Only in the11.32 % (91) of the articles the authors come from three or more research areas/disciplines (based on the name of their affiliated departments),

 in the 37.56% (302) of the articles, authors come from two different disciplines, with the 72% (216 papers) of these from medicine and informatics, and  in 411 (51.12%) articles all authors come from only one discipline, including in this number the 51 single-authored papers. This discipline could be medical, informatics, business, engineering or other. In the dataset, there are 138 papers with 2 authors, and 64 articles with more than 10 authors. The remaining 551 articles are written by 3 to 9 authors. The most populated article (Allen et al, 2016) includes 101 authors who form a panel of experts on the evaluation of the predicting performance of Alzheimer’s disease by a computational crowd-sourced project.

4.1c. Citation and Co-citation analysis based on the bibliographic data and the most popular keywords found in the 804 articles

Table 4 presents the citation report of the 10 most cited articles of the dataset by also providing a short summary of their research. In the summary, an emphasis is given to the nature of the studies which is related to big data. The most cited paper out of a total number of 804 papers is Bates et al. (2014) with 356 total citations and 36.50 average citations per year. It currently holds the highest scores in both categories, and the number of total citations considering the young age of the paper is noticeable. The article presents six cases of high-risk patients as examples of opportunities to reduce costs through the use of big data. It discusses the insights emerging from clinical analytics (types of data, algorithms, registries, assessment scores, monitoring devices, and so forth) for the healthcare organizations, which can lead to better decision making and the implementation of changes that will improve care while at the same time reduce costs. The second most highly cited article is O’Driscoll et al. (2013) with 319 citations and 21.33 average citations per year. It holds the second place in both score categories. This study provides an overview of big data technologies describing the example of the Apache Hadoop software and its current usage within the bioinformatics.

Table 4 : Top 10 most cited articles (as of Nov 2018)

Total Average cit. a/a Articles Summary citations per year 1 Bates, D. W., Saria, S., Ohno-Machado, Presents six examples of high-cost patients and ways to 356 36.60 L., Shah, A., & Escobar, G. (2014). Big reduce risk and costs through the use of big data and data in health care: using analytics to discusses the types of data needed and the infrastructure identify and manage high-risk and high- (e.g. wear devices that monitor in real-time cost patients. Health Affairs, 33(7), physiological parameters and remotely send data to 1123-1131. clinicians; develop machine learning algorithms to learn from previous experience and optimize patient allocation to therapies, etc.). 2 O’Driscoll, A., Daugelaite, J., & Sleator, An overview of cloud computing and big data 319 21.33 R. D. (2013). ‘Big data’, Hadoop and technologies handling biology large datasets, such as cloud computing in genomics. Journal of sequencing million human genomes to understand biomedical informatics, 46(5), 774-781. biological pathways and the genomic variation of a tumor. 3 Quinn, C. C., Clough, S. S., Minor, J. Describes the evaluation of a smartphone diabetes 306 14.18 M., Lender, D., Okafor, M. C., & management software which analyses, through its Gruber-Baldini, A. (2008). WellDoc™ statistical model, users’ logged data, trends, and mobile diabetes management behavior and then, through its therapy optimization randomized controlled trial: change in tools, provides to users real-time advice on diabetes for clinical and behavioral outcomes and better managing their disease. patient and physician satisfaction. Diabetes technology & therapeutics, 10(3), 160-168. 4 Chawla, N. V., & Davis, D. A. (2013). An overview of the role of Big Data analytics and 221 12.83 Bringing big data to personalized computation in personalized healthcare and biomedical healthcare: a patient-centered discovery. Create a personalized disease risk profile for an framework. Journal of general internal individual patient by leveraging the big data resident in medicine, 28(3), 660-665. electronic medical records, patients’ experiences and histories, along with the biological information of diseases and their interactions Deliver a personalized plan to an individual by leveraging similarities across a large group of patient pool, in real-time. 5 Andreu-Perez, J., Poon, C. C., An overview of the progress in biomedical and health 220 28 Merrifield, R. D., Wong, S. T., & Yang, informatics through big data. It explains the benefits for G. Z. (2015). Big data for health. IEEE J medical, sensor and imaging informatics, and translational Biomed Health Inform, 19(4), 1193- bioinformatics from piecing together different personalized 1208. information from unstructured and structured data, such as clinical diagnosis, imaging, continuous physiological sensing and genomics, proteomics, metabolomics. 6 Hilario, M., Kalousis, A., Pellegrini, C., A study on data analytics that takes mass spectra data of 199 8.23 & Mueller, M. (2006). Processing and biological specimens, like DNA microarray data and classification of protein mass discovers patterns between different pathological states spectra. Mass spectrometry applying classification algorithms and reporting predictive reviews, 25(3), 409-449. performance. A mass spectrum contains thousands of different mass/charge ratios. The reduction of the size of the high dimensionality variable set through classification is crucial to biomarker discovery. 7 Zhang, X., Yang, L. T., Liu, C., & Chen, An approach for data anonymization techniques of large 192 17.60 J. (2014). A scalable two-phase top- scale electronic health records that masks sensitive down specialization approach for data information specializing the level of information in a top- anonymization using mapreduce on down manner until a minimum privacy requirement is cloud. IEEE Transactions on Parallel compromised, making it possible to capture, manage, and and Distributed Systems, 25(2), 363-373 process them within a tolerable elapsed time. 8 Castellanos, F. X., Di Martino, A., Focuses on predictive modeling approaches for diagnosis 190 21.17 Craddock, R. C., Mehta, A. D., & through resting state fMRI, a method of functional Milham, M. P. (2013). Clinical magnetic resonance imaging that is used in brain mapping, applications of the functional a tool for brain-based biomarker identification for connectome. Neuroimage, 80, 527-540. neurological and psychiatric illness. The convergence of dimensional approaches and large dataset of images processing and sharing is propitious for improving predictions.

9 Krumholz, H. M. (2014). Big data and Explores the ways in which big data, such as biological, 189 19.60 new knowledge in medicine: the clinical, behavioral, and outcomes data can be analyzed thinking, training, and tools needed for a through advanced methods to predict, discover, and learning health system. Health compare effectiveness to tackle the complexity of patients, Affairs, 33(7), 1163-1170. populations, and health-related organizations in a similar way that it is done in other businesses. 10 Alyass, A., Turcotte, M., & Meyre, D. A review that discusses recent advances in high-throughput 188 20.00 (2015). From big data analysis to large omics (genomics, epigenomic, metagenomics, personalized medicine for all: challenges metabolomics, nutriomics, etc) technologies which have led and opportunities. BMC medical to more precise modeling of complex diseases accelerating genomics, 8(1), 33. the global transition to personalized medicine. It also touches upon ethics and equity issues.

The most co-cited paper (the most commonly referenced paper among the 804 articles) is by Murdoch & Detsky (2013), “The inevitable application of big data to health care”, published in JAMA. It is included in the reference list of 24 out of 804 articles. It is closely followed by the papers of Breiman (2001) and Dean & Ghemawat (2008). Table 5 presents the 11 most co-cited papers. The majority of the papers are related to health with the exception of two: Breiman (2001), which is about a machine learning method often used in health applications and Dean & Ghemawat (2008) which discusses MapReduce (both big data analysis methods used often in the health care sector). There is also the report from McKinsey (Manyika et al., 2011), which is about big data opportunities in general (with a special reference to the health sector in the US), and it is the only one not published in an academic journal. No overlaps are observed between the most cited (Table 4) and the most co-cited (Table 5) top 10 papers, except for two: Chawla & Davies (2013) and Bates et al. (2014) with the latter coming last in the list of the most co-cited papers as it shares the 10th position with Murphy et al. (2010). Overall, the great majority of the most cited and co-cited articles consists overview and review papers.

Table 5: Top 10 most co-cited articles

a/a Citation Summary Frequency 1 Murdoch, T. B., & Detsky, A. S. (2013). The A viewpoint that discusses the applications and 24 inevitable application of big data to health opportunities of big data (deriving from electronic care. Jama, 309(13), 1351-1352. health records) to health care, using an economic framework, to improve quality and efficiency of health care delivery. 2 Breiman, L. (2001). Random forests. Machine Gives insight into the random forests’ capabilities for 23 learning, 45(1), 5-32. classification and prediction. 3 Dean, J., & Ghemawat, S. (2008). MapReduce: Describes the function and success of MapReduce 19 simplified data processing on large clusters. programming. Communications of the ACM, 51(1), 107-113.

4 Raghupathi, W., & Raghupathi, V. (2014). Big data Describes the opportunities in healthcare from the 19 analytics in healthcare: promise and potential. Health analysis of big data, such as electronic health records, information science and systems, 2(1), 3. financial and operational data, clinical data, genomic data, real-time data from health monitoring devises. 5 Chawla, N. V., & Davis, D. A. (2013). Bringing big Look at Table 4 entry 4 16 data to personalized healthcare: a patient-centered framework. Journal of general internal medicine, 28(3), 660-665. 6 Lazer, D., Kennedy, R., King, G., & Vespignani, A. Presents the Google Flu Trends, a flu tracking system 16 (2014). The parable of Google Flu: traps in big data from social media posts, as a case study to provide analysis. Science, 343(6176), 1203-1205. critical lessons for the future of big data analysis. 7 Jensen, P. B., Jensen, L. J., & Brunak, S. (2012). A review focusing on the potential knowledge 15 Mining electronic health records: towards better discovery of genotype–phenotype relationship from research applications and clinical care. Nature integrating EHR data with genetic data and ethical, Reviews Genetics, 13(6), 395. legal and technical reasons currently hindering the systematic deposition of these data in EHRs and their mining. 8 Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, A research of McKinsey Global Institute about big data 15 R., Roxburgh, C., & Byers, A. H. (2011). Big data: analytics in healthcare and other 4 domains focusing on The next frontier for innovation, competition, and the economic impact of the technology. For healthcare productivity. McKinsey & Company it provides examples of health insurance organizations deploying electronic health records, health monitoring data from devises, R&D data and mostly financial and pricing data. 9 Ginsberg, J., Mohebbi, M. H., Patel, R. S., Brammer, Presents a method of analyzing Google search queries 14 L., Smolinski, M. S., & Brilliant, L. (2009). Detecting to track influenza in a population. This approach may influenza epidemics using search engine query make it possible to use search queries to detect data. Nature, 457(7232), 1012. influenza epidemics in areas with a large population of web search users. 10 Murphy, S. N., Weber, G., Mendis, M., Gainer, V., Presents a software, i2b2, that uses large datasets of 13 Chueh, H. C., Churchill, S., & Kohane, I. (2010). patient medical record data, such as diagnoses, Serving the enterprise and beyond with informatics medications, and laboratory values and provides for integrating biology and the bedside (i2b2). Journal clinical investigators with the ability to identify sets of of the American Medical Informatics patients with special health characteristics while Association, 17(2), 124-130. preserving patient privacy.

11 Bates, D. W., Saria, S., Ohno-Machado, L., Shah, A., Look at Table 4 entry 1 13 & Escobar, G. (2014). Big data in health care: using analytics to identify and manage high-risk and high- cost patients. Health Affairs, 33(7), 1123-1131.

The below co-occurrence analysis is based on all keywords (6015) of the 804 papers. The 10 most frequently used keywords are presented in Fig.6. As one would expect, the most popular keyword is “big data” with a frequency of 185 times throughout the whole article pool.

Figure 5: The most frequent keywords of the 804 articles

4.2. Content Analysis Results

The information from the unstructured data of these papers has been derived applying text mining method (Dinov, 2016). For the demonstration of the result, it has been built a text classification that is presented in Table 6. Every category of this classification contains several subcategories. The particular categories and their subcategories, which act as the guide to the dataset content analysis, were inspired by a number of prominent review and overview papers relevant to big data analytics in general, [Wamba et al. (2015) for the category of “research approach”, Wang et al. (2016) for the category of nature of analytics, Groves et al. (2013) for the categories of types of data & capabilities related to health, and finally, Chen & Zhang (2014) and Wallet and Facett (2013) for the category of BDA techniques]. Through the analysis of the selected categories, it has been an attempt to answer questions like: What medical

36 specialties have benefited the most? Who are the stakeholders of the use of Big Data analytics? What is the type and the frequency of the big data which have been used in the healthcare domain?; what big data are mostly used in health analytics techniques?; what is the level of analysis that has been reached (nature of analytics) and what type of research has been conducted (research approach)?; what capabilities have been acquired from the application of BDA in the health sector? In addition, a new category (medical specialties) was added to identify to whom this research is relevant. The selection of this classification attempts to map the knowledge in the field and explain the elements of big data analytics in healthcare through examples. Table 4 presents the classification framework.

Table 6 : Classification Framework

Classification a/a Dimensions Reference Source Tables Questions 1 Medical specialties U.S. Government site of Table 8 What medical specialties have medicare benefited the most? 2 Stakeholders of BDA Groves et al., 2013 Table 9 Who are the stakeholders of the use of Big Data analytics? 3 Research approach Wamba et al., 2015 Table 10 What type of research has been conducted? 4 Nature of analytics Wang et al., 2016 Table 11 What level of analysis have they reached? 5 Types of data Groves et al., 2013 Table 12 What kind of data have been used for the analysis? 6 Big data techniques Wang et al., 2016, Table 13 What techniques have been used? Waller & Fawcett, 2013

7 Big data capabilities Groves et al., 2013 Table 14 What is the purpose of the analysis?

The demonstration of the results has been achieved using some indicative examples of papers based on the classification framework presented in the methodology section. The selection of the particular examples is based on their popularity (high-cited papers in their categories) and the fact that they provide a clear and comprehensive case for each sub-category.

4.2a. Medical Specialties

The first categorization concerns the allocation of papers to the medical specialties. The groupings are based on the official U.S. government site for medicine1. Table 7 shows the relevant specialties and the number of identified papers with an indicative reference. In many cases articles are allocated in more than one subcategory. For example, an article about breast cancer was categorized in “Gynecology” and “Medical Oncology”. The ten most popular specialties are presented according to the number of the allocated articles. The most popular subcategory is the one that includes articles referring to all kind of medical specialties or to “no specific specialty” (49%). The next popular subcategory with 77 articles (10%) combines three medical specialties (Neurology/Neuropsyciatry/Psychiatry) since many articles refer to these categories together. No similar research has been identified for comparison of results, however given that health analytics is a fast- growing field of research it is expected the publication of a great number of studies on specific medical specialties in the near future. The specialties with the biggest number of studies, like neurology, oncology and cardiology, are on the spotlight of the World Health Organization (WHO). Based on World Health Statistics 2016, mental disorders affect one in ten people on the planet and almost 40% of premature deaths are caused by cardiovascular diseases, cancer, diabetes and chronic respiratory diseases. Therefore, the healthcare community is expected to give great importance to the evolution of these specialties with the help of the new capabilities offered by analytics.

Table 7: Distribution of articles to Medical Specialties

Specialties N % Indicative Reference No specific specialty 391 48.9 Althebyan, Yaseen, Jararweh, & Al-Ayyoub (2015) Neurology/ neuropsyciatry/ 77 9.6 Maccione et al. 2015 psyciatry Medical oncology 56 7.0 Miriovsky, Shulman, & Abernethy, 2012 Cardiology 54 6.7 Bardhan Oh, Zheng, & Kirksey, 2014 Infectious disease 23 2.9 Carroll et al., 2014 Endocrinology 22 2.7 De Silva, Burstein,

1https://www.medicare.gov/physiciancompare/staticpages/resources/specialtydefinitions.html?AspxAuto DetectCookieSupport=1

Jelinek, & Stranieri, 2015 Emergency medicine 22 2.7 Baum, 2010 Pediatric medicine 21 2.6 Basole et al., 2015 Radiology 19 2.4 Cook & Nagy, 2014 Pathology 19 2.4 Angelelli et al., 2015 Pulmonary disease 13 1.6 Kenner, 2016

4.2b. Stakeholders of Big Data Analytics in HealthCare

In the second part of the research the allocation of the papers is based on whom their content was relevant and identified seven different stakeholders who might benefit from using big data management systems in healthcare. The most common stakeholders in the use of BDA in healthcare are primarily the clinicians, who appeared in almost half of the articles (392 articles, 48.7% of the article pool). Research has shown that clinicians use new technologies to achieve the highest quality of patient care. A high number of papers focus on the development of decision support systems which can be used by medical staff to take informed decisions about patient diagnosis/care or for developing a better understanding for progression of certain diseases. Researchers appeared in 48.2% of articles. The researcher is the beneficiary when s/he tries to further understand relationships and differences among variables e.g. to cluster data in meaningful categories for achieving better decision making. This research is not yet robust enough to be used by clinicians, patients, policy-makers, etc., but it is very close. This analysis found that patients make up 36.8% of the article pool, as receivers of health information for themselves and the environment they live in. This might be used primarily to optimize their treatment quality or improve their quality of life in general. Administrators make up 20.8% of the BDA stakeholders whereby they successfully lead their organizations through better decision-making, recognize revenue opportunities, reduce re-admissions and costs and improve staff utilization and scheduling, etc. Research for IT specialists is limited (9.8%) and it is aimed at designing the systems analysis and IT platforms for correct and desired data display. Research for Vendors, with 5.72% representation, is targeted at gaining competitive advantage and investing into new products. Finally, policy-makers with 5.1% include governments, ministries of health, international health organizations, with research focused on trying to take advantage of the technological and scientific achievements to provide higher quality

39 public care and protect people from infectious diseases and terrorism. Table 8 summarizes these findings by also providing a definition for each beneficiary and a summary of the two articles given as examples of research targeted to each stakeholder. The selection of the examples is based on their popularity (high-cited papers in their categories) and the provision of a clear and comprehensive case for each sub-category.

Table 8 : Stakeholders of Big Data Analytics in Healthcare

Stakeholder Research Purpose Ν Examples Context s Clinicians To make sense of the vast 392 48.7 Mares et al. Present a system designed to help clinicians to amount of data using new 2014 improve their decision making by combining data of technologies in order to widely varying types and performing analyses which achieve the highest patient facilitates them in making new medical discoveries. care quality McGirt, Offer a predictive analytic tool that combines clinical Sivaganesan, data and patient interviews after lumbar spine , Asher & surgery. Data form 1800 patients were used to predict Devin 2016 12-month improvement in physical disability, return to work, complications, readmission, and need for inpatient rehabilitation. According to the authors, this model benefits physicians to generate knowledge about expectations after surgery and surgeons to have the ability to choose the right intervention, at the right time, for the right patient. Researchers To understand the 388 48.2 Banaee & Developed a BDA model where the temporal patterns differences and provide Loutfi, are mined from physiological sensor data such as outcomes leveraging the heart rate, respiration rate, and blood pressure. The grounded data 2015 proposed approach enables researchers to discover speciﬁc behaviors of vital signs, which are not necessarily recorded in medical ontologies and exploit unseen and distinctive information per patient or condition. After the validation of this approach it can be provided to clinicians to upgrade their decision- making. Researchers could also benefit from the generation of new ideas through their elaboration of the BDA system. Banos et al. Present the framework “Mining Minds”, that enables 2016 the provision of personalized support based on the core ideas of the digital health and wellness paradigms. Mining Minds elaborates technologies, such as Cloud Computing, Wearables and Internet of Things, and concepts such as context-awareness, knowledge bases or analytics, to investigate on people’s lifestyles and provide a variety of smart coaching and support services. The authors present here the conceptual framework and strongly believe that their approach can create new research ideas and also serve as a reference for similar initiatives. Patients To use their own medical 296 36.8 Althebyan,Y Presents a system that aims to improve the health of data in order to optimize aseen, patients and reduce the risk of having medical their treatment quality Jararweh, & problems. They developed a cloud based monitoring Al-Ayyoub, system that targets a crowd of individuals in a wide geographical area and efficiently can integrate many 2015 emerging technologies such as mobile computing, wearable sensors etc. that can offer remote monitoring of patients anytime and anywhere in a timely manner.

Azmak et Introduce the Kavli HUMAN Project (KHP) which al. 2015 aggregated data from 2,500 New York City households (roughly 10,000 individuals) whose biology and behavior is measured using modalities such as environmental conditions, events and geographic information over 20 years. Views were offered from the database of how human health and behavior evolve over the life cycle and why they evolve differently for different people. The authors argue that this kind of approach can improve the health and the quality of human life in urban contexts. Administrat To lead their organization 168 20.8 Brandley & Describe the benefits of predictive analytics from the ors with better decision - Kaplan, perspective of revenue opportunity identification and making increasing staff efficiency to recover the revenue. 2010 Both perspectives are key issues for administrators to run their organizations efficiently Hao et al. Present a decision tree based model with Electronic 2014 Medical Records (EMR) features to estimate readmitting patients. They claimed that their model benefits health care providers for estimating the ED revisit risks at the patient discharge time in order to maintain a perspective of health care economics for the future clinical resource related to ED. IT To design the desired 79 9.8 Klein, 2015 Development of IT infrastructure of Blythedale specialists systems analysis Children’s Hospital, in New York. An Outcomes Committee composed of health care and IT professionals found ways to manage electronic medical record and administrative data and provide mechanisms for real time outcomes. In order to present information to the clinicians when they needed it most, so they trusted IT professionals to create a novel solution in elaborating dynamic dashboards from EMR data Mamary, Describe a new decision support system of 2013 Hunterdon Medical Center, a 178-bed teaching hospital and enhances the importance of IT Specialist in the management team. Vendors Togain competitive 46 5.72 Bhattacharya Provide an overview of business analytics(BA) in advantage and invest in Ramachandr healthcare hosted in the cloud computing new products an, A., environment, offered as Software as a Service (SaaS) &Jha,. 2012 solution. Identify the benefits for healthcare enterprises when making use of a BA SaaS Bose & Das, Develop an innovative visualization tool to help 2012 overcome the potential operational deficiencies of clinical trials by the use of Clinical Trial Management Systems (CTMS) in order to improve managerial control. This solution provides opportunities to reduce costs and time, to stimulate revenue growth, and enables companies to understand the entire clinical trial process from the global organizational perspective Policy- To take advantage from the 41 5.1 Choucair, Describe how the public health departments in makers technology and science Bhatt Chicago are moving from one-time programmatic achievements in order to & Mansour, interventions to sustainable system-level innovations provide improved public 2015 in order to create new strategies to solve old care quality and protect problems by capitalizing on the innovative synergies people from infectious of civic tech communities, health care systems, and diseases and terrorism emerging markets Davidson, A network using data collected by the U.S. Centers Hai, for Disease Control and Prevention (CDC) and &Radin, combine this with Google Flu Trends (GFT) that 2014 predicts infections into the future as well as to identify the regions that are most likely to accelerate influenza spread during epidemics. The findings of this paper have important implications for prevention and control efforts at the local and national level.

4.2c.Research approach

The selected publications have been according to their research approach, as of Wamba et al. (2015) and Chen & Zhang (2014) with some additions (Table 9). For the diversification of the articles according to their research approach, some best examples of each approach are presented here. The results of the categorization indicate that most papers are experimental studies, followed by review studies. In general, papers under the categorization of “experiment” cover both a theoretical contribution (an advancement of an algorithm, experimentation with program running time, etc.) and a part where this theoretical advancement is tested or evaluated in relation to the under- discussion application. In a field with an expected growth in the following years (according to the literature above), it is anticipated that the academia will provide at least this volume of experimental studies. Many articles have been categorized in more than one subcategories [e.g. the research of Barret, et al. (2008) and Blakely et al. (2015)] as they include more than one approach in their methodology.

Table 9 : Research Approach and indicative research examples

Research approach N % Examples Context

Experiment 318 39.6 Barret, The study proposes mixed effect models and Bayesian Incorporates papers that Mondick, forecasting algorithms to develop drug–specific provide experimental Narayan, dashboards for better decision making and education of results of new models Vijayakumar & patient caregivers on clinical pharmacology principals Vijayakumar., which lead to fewer medication errors, reduced 2008 toxicity, reduced length of hospital stay, etc. Data visualization tools summarize patient profiles from hospital electronic medical records of pediatric populations, such as lab values, vital signs, and associated biomarker and interface those data by a web- based decision support system. Review & Overview 178 22.1 Gligorijevic, The study reviewed recent big data integrative methods Includes literature Malod‐Dognin for disease sub typing, biomarkers discovery, and drug review papers that & Pržulj, 2016 repurposing, and listed the tools that are available to present a summary of domain scientists while highlighting key issues in the the research methods context of personalized medicine. and outcomes in a specific field Data analysis 146 18.1 Bello –Orgaz, The study analyzed large scale text related to vaccine Papers that contain Hernandez- opinions retrieved from Twitter for measuring the methods and results Castro & potential influence of these opinions based on the from analyzed data Camacho, 2017 variation in the vaccination coverage rates. This method can be used to detect and locate communities against vaccination that could generate future disease outbreaks in different parts of the world.

Conceptual 140 17.4 Kuiler, 2014 This study presents a conceptual framework for data Studies that provide analytics. An IT-supported ontology-based approach conceptual frameworks for health data to address the semantic challenges and general discussions presented by Big Data sets and discusses architectural on the investigated considerations. Future research will focus on scientific fields developing the specifications for the lexicon, ontology, and other architectural artifacts to support software development. Case study 86 10.6 Chute, Beck, A case study about Mayo Clinic and its “semantically Qualitative research Fisk & Mohr, integrated warehouse of “biomedical data”.An based on a case and 2010 information management initiative that integrates a designed to suit the huge amount of different medical data types. research question Survey 22 2.7 Yildirim, The authors dealt with the analysis of 1941 children Studies that gathered Majnarić, clinical data, in a Health Center of Osijek, Croatia, and and analyzed Ekmekci & interviewed their parents for more details of family questionnaires and/or Holzinger, 2014 history on antibiotics and other allergic and chronic participant opinions diseases with the purpose of investigating reactions and allergy from antibiotics in children. Their analysis involved structure and unstructured data from a big population to present outcomes in biomedical research.

4.2d. Nature of Analytics

The following category allocates the articles according to their descriptive, predictive or prescriptive nature as of Wang et al., (2016). In many cases the articles have been distributed in more than one subcategory because there is evidence of more than one dimension in some papers. The most popular subcategory with 47% (377 papers) is that of “Predictive analytics”. The second subcategory in the classification is “Prescriptive Analytics”, with 33% of papers (263 out of 804) and the last one is “Descriptive analytics” (24% with 190 articles). In the paper of Raghupathi & Raghupathi (2013) descriptive analytics were found to be the most commonly used type due to their explanatory and easy approached nature. However, in healthcare, prediction is more valuable than explanation because the outcomes are measured in lives (Agarwal & Dhar, 2014). The majority of articles in the dataset are published after 2013, and therefore later than the publication of Raghupathi & Raghupathi (2013). Healthcare is a growing sector and as a result advanced technology and skills are needed for the application of models with predictive or prescriptive character. The industry may have a time-lag in the adoption of the more advanced nature of analytics but the research must pave the way (Groves et al., 2013). In the article pool, the majority of papers (40%) included experiments of new models with the hope that these predictive/prescriptive models will become s part of software and will be adopted by analysts for use in the decision-making in health care organizations or systems.

Table 10: Nature of Analytics and indicative research examples

Nature of analytics N % Examples Context Predictive 377 46.9 Bardah et al., Presented a novel model to predict readmission of Involve the use of 2015 patients with congestive heart failure. The model mathematical algorithms to tracks patient demographic, clinical, and discover predictive patterns administrative data across 67 hospitals in North within data and project what Texas over a four-year period. will happen in the future. Prescriptive 263 32.7 Sir, Dundar, Surveyed 2865 patients from the surgery unit and Involve the use of data and Steege & 3241 from the oncology unit and proposed nurse– mathematical algorithms to Pasupathy, patient assignment models to achieve a balanced determine decisions that 2015 assignment workload. Patient metrics used from involve objectives with the QuadraMed AcuityPlus patient classification system aim to improve performance. (which accumulates hospital’s patient indicators over 20 years) to classify patients on nurses’ workload. Descriptive 190 23.6 Basole et al., Presented a visual analytic tool that used clinical Techniques such as online 2015 data from 5784 pediatric asthma emergency analytical processing (OLAP) department patients and reported that asthma is the that aim to identify problems most common pediatric chronic disease and is the and trends in existing third leading cause of hospitalization among processes and functions. children, affecting 9.3% of children in the US. Their results assist in the improvement of health care quality. The data were obtained from Population Discovery, Children’s data warehouse. This included patient and provider information, administrative events, clinical observations, medications, laboratory tests, and charges in a elational database

4.2e. Types of data

For the purpose of this Thesis, it has been adopted a detailed description of the types of primary data pools used in healthcare from the study of McKinsey and Co (Groves et al., 2013). These include: A. Clinical data, B. Patient and sentiment data C. Administration and cost activity data and D. Pharmaceutical and R&D data. The review papers did not take part in this classification. In Table 11 together with the allocation of papers according to the type of the analyzed data, there has also been provided in the second column a definition of these types of data. Overall, the adopted types are in line with the categorization used by other researchers too (Gaitanou, Garoufallou & Balatsoukas, 2014). The most popular data that have been analyzed within the articles are “clinical data” with a 70% (562 articles out of 804) representation. The results are consistent with the literature which has identified that significant research has been focused on electronic health records (EHRs) implementation, but relatively few studies exploited other types of big data (Gaitanou, Garoufallou & Balatsoukas, 2014).

Table 11 : Types of data and indicative research examples

Types of data N % Example Context Clinical data 562 69.9 Forsberg et Collected biomarker and clinical information from 73 Patient data such as EHR and al., 2015 patients who sustained 116 life threatening combat wound by medical images conﬂicts in Afghanistan and Iraq, and tried to determine if those data could be used to predict the likelihood of wound failure. The collected data included clinical information, serum, wound efﬂuent, and tissue and their analysis model indicated that it would improve clinical outcomes and reduce unnecessary surgical procedures. The same approach was also tested and performed equally well with larger samples of patients (67,486 patients with traumatic extremity wounds). Patient behavior and 133 16.5 Boulos, Described the analysis of predictive tools that gather posts sentiment data Sanfilippo, and queries from Social Web (“Web 2.0”) tools such as Data collected from wearable Corley & blogs, micro-blogging and social networking sites to form sensors and social sites Wheeler, coherent representations of real-time health events like flu 2010 out-breaks. Harvested data in the form of human feelings from a large number of blogs and social pages such as those hosted by MySpace. Administrative & cost data 59 7.3 Abbas, Used a vast number of individuals’ administrative and Financial and operational Bilal, Zhang clinical data to create a cloud based solution (Software as a data and patient profiling & Khan, Service) that provides personalized recommendations about data and choices 2015 the health insurance plans according to the user specified criteria. Pharmaceutical R&D data 38 4.7 Calabrese, Described “Pharmachosychrony” as a new concept of Drug therapeutic mechanisms, Minkoff & analytical pharmacy solutions to improved care coordination R& D data from target behavior Kristine, and provided a high quality and patient-centric model of in the body such us effects of 2014 care. Among the data that this solution elaborates, pharmacy toxicity etc. data are included for the effective and safe use of medication. Elaborated claims data from call centers, web portals, mobile technology, and decentralized clinical staff.

4.2f Big Data Techniques

The boundaries of techniques among BDA are difficult to be completely distinguished (Royston, 2013). For better understanding the use of the different techniques, there is a definition in the second column and some indicative examples in the last. The listed BDA techniques have derived from the literature (Chen & Zhang, 2014; Waller & Fawcett 2013) and although some may overlap with each other or consist a sub-category of another, they are as inclusive as possible. For example, in Table 12, “web-mining” is presented separately from “data-mining” although it can be seen as a subcategory of the latter, acknowledging the fact that this mining field is represented by a quite large number of papers and has gathered momentum because of the high usage of internet data in the very last decades. The case with the allocation of papers to the “modeling” and “simulation” techniques is also similar. The criterion for

allocating a paper to the “modeling” subcategory was whether the modeling technique mainly included mathematical formulations of variable relationships presented in a static form, and that for the allocation to the “simulation” subcategory was whether the data variability was addressed by running the model many times with different values taken from a distribution. Willing to address both approaches and present indicative research examples to explain them there are separated techniques. Moreover, in the statistics subcategory, the majority of articles are allocated to another technique too, and overall many of the papers have multiple entries as the handling of big data requires a combination of techniques for their analysis. As seen in Table 12, “modeling” emerges as the most popular technique, as it is also the most general amongst the categories. It is followed by “machine learning” (which includes the design of algorithms), a fast- growing technique with lots of successful cases in the field of health, such as the classification of medical data and symptoms for disease diagnosis and prediction (Chen et al. 2017; Khalaf et al., 2017).

Table 12: Big Data Techniques and indicative research examples

Techniques N % Examples Context

Modeling 344 42.8 Ajorlou, Developed a linear predictive Bayesian model indicating Methods of analytical Shams & that risk adjustment for patient health conditions can mathematical analysis with Yang, 2015 improve the prediction power. Data from 82,000 patients approximate relationships from 888 facilities assembled for a total capture period of between variables (Waller & one year and assessed from the Veteran Health Fawcett 2013) Administration. Machine learning 327 40.7 Dugan, Experimented with six different machine learning methods Artificial intelligence Mukhopadhya to identify the best one for predicting future obesity in aimed to design algorithms y, Carroll & children above two years old with 85% accuracy. Data collected from a pediatric clinical decision support system that allow computers to evolve Downs, 2015 behaviors based on empirical data. (CHICA) and used for the analysis. The data included nine (Chen & Zhang 2014) years of clinical information collected from 4 different community health centers. Data mining 200 24.9 Delen, 2009 Used three popular data mining techniques (decision trees, A set of techniques to artiﬁcial neural networks and support vector machines) to extract information from develop prediction models for prostate cancer survivability. data (Chen & Zhang, 2014) The researchers obtained around 120000 records from the Surveillance, Epidemiology, and End Results Program and formed 77 variables for statistical analysis. They concluded that data mining methods are capable of extracting patterns and relationships but are useless without medical experts’ feedback. Visualization approaches 153 19 Angelleli et Presented a visualization tool “brain atlas” with cohort data The techniques used to create tables, al., 2014 analysis of 100+ participants. The tool, which was assessed images, diagrams and other by neuropsychological testing, genetic analysis and intuitive display ways to understand multimodal magnetic-resonance (MR) imaging, enables a data (Cheng & Zhang, 2014) first quick analysis of the identified hypotheses.

Statistics 132 16.4 Demir, 2014 Proposed a method to compare predictive analytic The methods of organizing capabilities of emergency readmissions. Using data from and interpreting data for exploiting the emergency department from 963 patients with chronic causal relationships between obstructive pulmonary disease and asthma within 45 days different objectives after a patient has been discharged from hospital. This data (Chen & Zhang, 2014) set was divided into derivation and validation samples 1000 times. They actually proved that predictive logistic regression and regression trees could be a valuable decision support tool for clinicians for the prediction of readmissions. Simulation 55 6.8 Liu & Wu, Developed an agent-based simulation model to study Quantitative analysis 2016 accountable care organizations. It identified the critical of a system in a determinants for the payment model design that can stochastic setting motivate provider behavior changes to achieve maximum (Waller & Fawcett 2013) financial and quality outcomes that considers payers, healthcare providers, and patients as agents under the shared saving payment model of care. It constructed a healthcare system analytics model that can help inform health policy and healthcare management decisions. Web mining 54 6.7 Chen & Developed an analytics platform, called “Cytobank”, for The process of information Kotecha, community cytometry data analysis (to track cells and discovery from sources across the 2014 subsets in blood and tissue) using large computing World Wide Web (Cooley et resources for analysis on the Internet. These platforms can al.,1997) simultaneously measure up to 100 parameters. Optimization methods 49 6.1 Katircioglu et IBM Research developed a scenario modeling and analysis Methods that find the minimum or al., 2014 tool, supply chain scenario modeler (SCSM), for maximum of a function, subject to McKesson (the largest healthcare services company) to constraints and solve quantitative optimize its pharmaceutical supply chain policies. SCSM problems, improve the optimizes the distribution network, supply flow and accuracy of forecasting and inventory policies and quantifies the impacts of changes algorithms( Waller & Fawcett 2013) on financial, operational, and environmental metrics. They developed complex queries to generate all input needs and rigorously tested them. The resulting data model has over 200 tables with a combined size of tens of millions of records. Text mining 42 5.2 Holzinger & Presented an overview of some selected text mining Techniques based on Jurisica, 2014 methods, i.e. Latent Semantic Analysis, and Probabilistic machine learning and Latent Semantic Analysis along with examples from the data mining to ﬁnd useful biomedical domain by extracting data from texts patterns in text data (unstructured patient data and, structured patient data e.g. (Holzinger et al. 2014) biometrics or laboratory results), and biomedical images, which will benefit clinical decision support. It provided machine learning solutions for large and complex biomedical data analysis. Forecasting 22 2.7 Toerper et al., Developed and evaluated a web-based forecast tool that Is about predicting 2016 predicts the daily bed need for admissions from the cardiac the future, while also evaluates catheterization laboratory. The forecast model was what could happen derived using a 13-month retrospective cohort of 7029 under different circumstances catheterization patients and included predictor variables using predictive analytic methods such as demographics, scheduled procedures, and clinical (Waller & Fawcett, 2013) indicators mined from free-text notes Social Network Analysis 20 2.5 Abbas et al., Proposed a cloud based framework for BDA in health that A technique that views and 2016 uses the Internet and social media. The framework offers analyses data from social users disease risk assessment and consultation service from networks health experts on Twitter with high accuracy results. It utilizes collective data of people’s health status from whole populations.

4.2g.Big Data Capabilities

The term “big data capabilities” refers to the different organizational competencies created by IT models that analyze vast amounts of complex and different

47 types of data, processed in daily operations (Bharadwaj, 2000) and resulting to better decision making. But, on which big data capabilities healthcare should focus, in order to achieve its goals? To assist managers in better decision-making, organizations must develop infrastructure with essential big data capabilities (Groves et al., 2013). Groves et al. (2013) acknowledged five important BDA capabilities, which have been adopted in this study. These are: a. “monitoring”, which includes articles that present monitoring efficiencies, and collect and analyze data (using analytical methods) describing “what is happening now”, b. “prediction/simulation”, which includes articles that present methods that provide information about future outcomes (what will happen), c. “data mining”, which incorporates articles that involve methods enabling extraction and categorization of knowledge (what happened), d. “evaluation”, which includes articles that demonstrate methods for testing the performance of BDA techniques or explain the outcomes of the application of BDA (why did it happen?) and e. “reporting capability”, which includes articles with methods for organizing the collected data in an informative format. Due to the plethora of capabilities described in the papers, many of them have been allocated to more than one subcategory. The difference between “data mining capability” and “data mining technique” (Table 13) is that the latter applies algorithms and mathematical modeling to perform clustering, etc. while the term capability refers to the process of applying these methods with the purpose of uncovering hidden patterns in large datasets. There is also a link between the BDA capabilities as described by Groves et al. (2013) and the nature of analytics as described by Wang et al. (2016). For example, the prediction/simulation capability is connected to the predictive and prescriptive nature of analytics respectively, and the monitoring and reporting capabilities are associated with the descriptive nature of analytics. However, the capabilities focus on the IT functionalities and the nature of analytics focuses on the technique’s goal. Table 13 provides examples of articles for each subcategory for further clarifying each capability. The most popular subcategory is “monitoring” with 33% of articles (264 papers out of 804), which shows the importance of the use of analytics methods to assist managers to maintain a view of “what is happening now”. The next dimension for the distributed articles is “prediction and simulation” with 32%. Having already demonstrated that in this systematic review predictive analytics is the most exploited type used in the examined articles, it can be justified that a good percentage of all the

48 papers would provide information about predictive BDA capabilities. The next popular sub-dimension is “data mining” with 29% (230 articles), followed by the “evaluation” with 13% (105), reflecting the publishing activity on evaluating the applications of BDA in healthcare. In this subcategory, articles that produce methods to evaluate the performance of other applications are often encountered. Finally, the last capability in Table 13 is “reporting” with a percentage of 9% (72 articles). A closer look at the allocation of articles reveals that the majority of the papers are almost equally distributed in the three first BDA capabilities a. monitoring (33%), b. prediction/ simulation (32%) and c. “data mining” (29%). However, in the reality of the health sector, reporting and monitoring activities are already occurring but predictive modeling and simulation techniques have not been used at scale yet (Groves et al., 2013).

Table 13: Big Data Capabilities and indicative research examples

Big data N % Examples Context capabilities Monitoring 264 32.8 Althebyan et al., Proposed an e-healthcare monitoring system that targets a What is 2016 crowd of individuals in a wide geographical area that happening integrates emerging technologies such as mobile now? computing, wearable sensors, cloud computing etc to offer remote monitoring of patients anytime and anywhere. The monitoring BDA capability providing through this system can enhance the decision support system in order to reduce risk of patient health decisions. Prediction 258 32 Abdelrahman, Proposed a new analytical approach to develop high- /simulation Zhang, Bray & performing predictive models for congestive heart failure What will Kawamoto, (CHF) readmission using an operational dataset with happen? 2014 incomplete records and changing data over time. Data came from 2,787 CHF hospitalizations at University of Utah Health Care Center from January 2003 to June 2013. Data 230 28.6 Chen et al., 2016 Developed a bootstrapping method for global module mining detection on features across breast cancer cohorts. They What did it used electronic medical records’ data from a Medical happen? Center annotated with BioCarta signaling signatures and provided new insights into breast cancer, such as the association of patient’s cultural background with preferences for surgical procedure. The modeling tool demonstrated unique ability to discover clinically meaningful and actionable knowledge across highly heterogeneous biomedical big data sets. Evaluation 105 13 Catlin et al., Proposed a web-based analytics system for conducting in Why did it 2015 house evaluations and comparisons of “infusion pump happen? data” across hospital systems allowing users to select any number and combination of hospital data. Smart pump infusions are customizable libraries with dose limits and administration rates speciﬁc to medications and care areas and provide information in order to avoid medication errors like the delivery of wrong drugs or delivery to the wrong patient or assessing the wrong dose.

Reporting 72 8.9 Curcin, Presented a software -Web Improvement Support in What Woodcock, Healthcare (WISH) – which is a prototype tool that happens on Poots, Majeed, attempts to translate research into practice using local a regular & Bell, 2014 improvement projects. This approach facilitates electronic basis data collection and reporting in health settings and is tested on a Chronic Obstructive Pulmonary Disease improvement project run in Northwest London Hospitals. Data are gathered from a large class of tasks, particularly local ones that cannot be adequately measured by exclusively using routinely collected data residing in hospital’s EHRs.

4.3. The association of the selected analytical techniques with the data types and capabilities

Using the NVivo analysis software, there is a demonstration of the techniques that have most popularly been applied for each data type and presented value. During this procedure, each one of the 804 articles to the data types (as presented in Table 11) and performed a breakdown by technique (Table 12). Table 14 shows, for example, that out of all studies that deal with clinical data (562), 41.5% have used machine learning for their analysis, from studies with patient behavior/sentiment data 51% have used machine learning, etc. It seems that the most popular techniques scientists need or prefer to use are: modeling, machine learning. The same techniques are also popular, in the same order, with the exception that machine learning comes first, for all data types (Table 14). Overall, machine learning and modeling are the most applied techniques amongst all data types with a variance of 29% to 49. Noticing the percentages there is a level of uniform distribution throughout each technique. This is and with most data types as shown in Table 11.

Table 14 : Classification of dataset articles based on the analytical techniques by data type

Patient Administrative Pharmace Data types→ Clinical behavior & (activity & utical &

sentiment cost) R&D data % (n) 100 (562) 100(133) 100(59) 100(38) BDA Techniques↓ % (n) % (n) % (n) % (n)

Machine learning 41.5 (233) 51 (68) 32(19) 31.6(12) Modeling 28.6 (161) 33.8(45) 49(29) 44.7(17)

Data mining 24(135) 21.8 (29) 25.4(15) 15.8(6) Visualization 22.2(125) 15.8 (21) 13.5(8) 5.3(2) Statistical analysis 20.3(114) 11.3(15) 15.2 (9) 13(5) Simulation 7.1(40) 6.7 (9) 8.5(5) 13(5) Optimization 7 (36) 3(4) 6.8(4) 5.3(2) Web mining 4.5(25) 16.5(22) 5(3) 2.6(1) Text mining 4.5(25) 6 (8) 3.4(2) 2.6(1) Forecasting 2 (14) 2.2(3) 1.5(2) 0 Social net. 1(6) 12 (16) 3.4(2) 2.6(1) Analysis

4.4. Gained values/capabilities from the use of BDA in the health sector

The benefits from the analytics in healthcare have been summarized in the ability to provide comparative effectiveness research to determine more clinically relevant and cost-effective ways to diagnose and treat patients (Ragupathi & Ragupathi, 2014). More specifically, in order to identify the full range of the emerging capabilities in the health sector from the use of big data analytics there has been a further classification of them under 10 categories of value creation. Table 15 presents these values sorted by popularity with a short explanation and the frequency of papers from the dataset that refer to one or more of these gains based on the research they present.

The first five values are similar to those identified in the study of Wamba et al. (2017).The most popular value “Better diagnosis for provision of more personalized healthcare” refers to the BDA capability to direct to better case diagnosis from the collection of more data and therefore offer more targeted therapy or health service to the individual. This, for example, could be the analysis of the numerous relationships of specific patient’s biomarkers which can lead disease therapy to precision medicine (Alyass et al., 2015), or the investigation of patient health metrics and behavior through wearables and the Internet of Things leading to specific interventions based on the collected data (Banos et al., 2016).

The second value “Supporting/replacing professionals’ decision-making with automated algorithms” is about mining knowledge from large data sets and training algorithms to pattern matching. This means better automatic categorization of new information entering the analysis process and improved decision-making when it comes, for example, to diagnosis and choice of therapeutic scheme.

The third value “New business models, products and services” refers to the development of new business models, products, and services through the capabilities offered by BDA, such as a new visualization software with real-time

51 statistical analyses of brain images for better patient diagnosis (Angulo et al. 2016) or a mobile application in which people can enter symptoms and get possible diagnoses and recommended medication.

The fourth value “Enabling experimentation, expose variability and improve performance” from the use of BDA, is for researchers to acquire a deeper understanding of all possible interrelationships between variables and develop scenarios for further experimentation with their models and expose new health information.

The fifth value “Healthcare information sharing and coordination” is gained by the coordination and sharing of health information across healthcare services or even countries to improve of health professionals’ decision-making. The sixth value “Creating data transparency” is about the ability of BDA to collect big data and format them in a standardized way. This capability reduces data identification and analysis time and assists the previous value of coordinating meaningful and comprehensive health-related information.

The next value “Identifying patient care-risk” refers to the capability of running the big data in advanced statistical techniques, such as logistic regression models and regressions trees which can identify scenarios of risk patterns and therefore alert for areas of health risk prevention. For example, identifying high risk populations for a particular disease helps policy-makers to decide on giving earlier access to screening to these populations.

The following value “Offering customized actions by segmenting populations” refers to the use of BDA to identify new factors, through clustering and other methods, for segmenting populations differently or in more categories and offer more targeted health services or products.

Value 9, that is “Reducing expenditure while maintaining quality” focuses on the capability of analytics, through process mining, visualization techniques and collaborative tools, to propose ways for reducing health organizations’ costs from better resource utilization, elimination of non-value-added actions, capturing hospital underpayments, etc., while maintaining the quality level. An example could be the use of visualization tools for identifying non-value-added processes in patients with chronic diseases by tracking patient data over time during home, ambulatory and hospital care.

The last value, “Protecting privacy”, is about how BDA can offer data security in ways such as the identification of privacy breaches, the capability to extract data by eliminating ID recognition from electronic medical records and others. This has become a big issue especially for organizations that use cloud computing as their main processing platform in which privacy and security are difficult to be controlled (Larson & Chang, 2016). Overall, the majority of health data analytics studies attempt to direct their efforts to patient benefit. Needless to say, that almost all studies have this ultimate goal but their direct focus may be at the intermediate stage for improving the way of doing it.

Getting an overall picture, the values 1, 2 3 and 7 directly relate to patient wellbeing (P), values 4, 5, 6, 10 relate to analysts (A) for better data handling and values 8 and 9 relate to management (M) for better positioning their products/services and gaining management efficiencies respectively. The identification codes (P), (A), (M) are presented under the “Types” column in Table 15.

Table 15: Created Values from the use of BDA

Value Types Definition N % V1 Better diagnosis for provision of more Analytic approaches for better patient diagnosis which lead to 286 35.6 personalized healthcare (P) provision of more personalized therapeutic schemes or services to the users V2 Supporting/replacing professionals’ Through adaptive rules/algorithms for fast categorization of 206 25.6 decision-making with automated symptoms/medical results and pattern matching, analytics can algorithms (P) provide recommendations for diagnosis and remedy/actions. V3 New business models, products and BDA enables companies to create new products and services. 197 24.5 services (P) e.g. new software for analysis of data/images, enhance existing ones, and invent entirely new business models, new ways of reaching to patients. V4 Enabling experimentation, expose Analytics create conditions for enhanced experimental 144 17.9 variability and improve performance applications of large datasets for testing “what-if” scenarios and (A) assisting performance and decision-making V5 Healthcare information sharing and BDA can organize the selection and sharing of information and 122 15.2 coordination (A) data analysis among stakeholders to gain operational efficiency V6 Creating data transparency (A) BDA can collect/convert data in a standardized format and treat 115 14.3 data in the same way for reducing time, cost of search and processing while maintaining clarity and quality V7 Identifying patient care-risk (P) BDA create enhanced opportunities of health risk prediction for 79 9.8 acting proactively to patient care-risk V8 Offering customized actions by BDA through high exploitation capabilities of big data can 72 9 segmenting populations (M) discover specific segmentations and tailor products and services to meet patients or health professionals’ needs. V9 Reducing expenditure while BDA enables new, cost-effective ways to intervene on the 72 9 maintaining quality (M) determinants of health, aiming at reducing expenditures while sustaining health outcomes. V10 Protecting privacy (A) BDA can identify ways of securing privacy of health-related 41 5.1 data to support the ethical principles and people respect.

Using the NVivo analysis software, there is a demonstration of the techniques that have most popularly been applied for each of the 10 identified values (Tables 15). It seems that the most popular techniques scientists need or prefer to use are: modeling, machine learning, data mining, visualization, and statistical analysis (Table 16). Overall, machine learning and modeling are the most applied techniques across almost all values, with a variance of presence between 32% and 61%.

Table 16: Classification of dataset articles based on the analytical techniques by created value

Values → V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 100 100 100 % (n) 100(286) 100 (122) 100(115) 100(79) 100 (72) 100(72) 100(41) (206) (197) (144)

BDA % (n) % (n) % (n) % (n) % (n) % (n) % (n) % (n) % (n) % (n) Techniques↓

Machine 48(137) 39 (80) 50 (98) 51.4 (74) 47.59(58) 31.3 (36) 38 (30) 32 (23) 26.4 (19) 61 (25) learning Μodeling 46.5 (133) 49 (101) 39.6 (78) 43 (62) 37.7 (46) 54.8 (63) 60.8 (48) 32 (23) 43 (31) 34.1 (14)

Data mining 28 (80) 25.2 (52) 20 (39) 23 (33) 23.8 (29) 23.5 (27) 13.9 (11) 30.5 (22) 33.3 (24) 41.5 (17)

Visualization 20.3 (58) 18.4 (38) 16.7 (33) 20.9 (30) 28.7 (35) 29.5 (34) 17.7 (14) 11 (8) 16.6 (12) 7.3 (3)

Data mining 28 (80) 25.2 (52) 20 (39) 23 (33) 23.8 (29) 23.5 (27) 13.9 (11) 30.5 (22) 33.3 (24) 41.5 (17)

Statistical 19.2 (55) 17 (35) 9.6 (19) 13.9 (20) 13.9 (17) 13 (15) 38 (30) 30.5 (22) 8.3 (6) 12.2 (5) analysis Simulation 6 (17) 8.2(17) 7.1 (14) 14 (20) 1.6 (2) 12.2 (14) 3.8 (3) 7 (5) 9.7 (7) 2.4 (1)

Optimization 5.6 (16) 6.3 (13) 5 (10) 13.2 (19) 3.3 (4) 7.8 (9) 3.8 (3) 4 (3) 9.7 (7) 0

Web mining 5.6 (16) 4.4 (9) 7.1 (14) 9.7 (14) 11.5 (14) 18.3 (21) 7.6 (6) 11 (8) 4.2 (3) 4.9 (2) Text mining 2.5 (7) 4.9 (10) 5 (10) 3.5 (5) 7.4 (9) 1.7 (2) 7.6 (6) 8.3 (6) 1.4 (1) 0 Forecasting 2.09 2 (4) 4.5 (9) 3.5 (5) 0 6 (7) 1.2 (1) 0 6.9 (5) 2.4 (1) Social net. 1.7 (5) 1.4 (3) 2 (4) 2.8 (4) 4.9 (6) 0.9 (1) 1.3 (1) 7 (5) 0 0 analysis

4.5 Types of social value creation in healthcare

As mentioned, aiming to identify the types of value creation in the healthcare industry from the analysis of big data, 10 value categories have been formed. The order of their presentation in Table 15 is based on their degree of popularity among the 804 articles. In Table 17 there is an attempt to denote the social positive effects as well. The majority of researchers expect organizations to gain value from the “personalized innovative medical approaches (35.6%) (V1). For example the analysis of specific patient’s biomarkers guide disease therapy to precision medicine, which brings better diagnosis results tailored to individual perspectives. The capability of BDA to quickly monitor and analyze the health records of large patient cohorts to learn, for example, which individuals respond to certain types of drugs, etc., help in this direction. In sociology, “the concept of identity“ is a highly debated subject and concerns the understanding of a person’s character, situation and experiences, which can be reflected to the individual’s health condition and treatment as health is highly connected with daily life and everyday habbits (Lowton et al., 2017). Personal health-related data, which can easily be acquired from the use of devices e.g. sensors embedded in smartwatches and smartphones log data, enables the continuous collection of behavioral and social data too, such as communication intensity from calls, sms, twits, etc., which characterize social interaction, physical activity (steps count), sleep patterns and heart

54 rate monitoring which affect mood and behavior, identification in real-time hand-to- mouth gestures that characterize smoking pattern, and many others. The associated social value of these new approaches to diagnosis is that utilizing such data on an aggregate level could allow the detection of factors such as, stress, smoking and drugs that have a negative social impact (Kumar et al., 2015). The next popular value type (V2) is defined by the improvement of decision- making through mining all possible knowledge from vast amounts of collected data by supporting or even replacing human decision-making with algorithms (25.6%). Such capability enhances and accelerates health professionals’ diagnosis of patients and possibly mitigates errors in proposed therapy, when algorithmic results are explored and coupled with appropriate medical education. The social impact is related to the replacement of human labor with technology and the development of new employment conditions (e.g. new job positions related to medical technology management, increased specialization of medical staff) with debating results for the society (Loebbecke & Picot, 2015). It might also have an effect on the “sociology of diagnosis” theory related to the authority of the medical expert over the patient, and the changing power in the physician-patient encounter, since the patient will be aware that diagnosis and treatment is mainly based on automated algorithms (Lupton & Jutel, 2015), to which patient may have access too. Third (V3), comes the value of new business models, products, and services, identified in 24.5% of the articles. This includes articles that develop valuable tools and new products/services which assist decision-making for care and therapy for bigger populations. It also refers to the context of a “rapidly developing ecosystem of digital health technologies” including dimensions such as online forums and medical and health-related apps of self-diagnosis (Lupton & Jutel, 2015) that have the potential to be transformed to socially meaningful scientific knowledge conceptualized as a public good (Evangelatos et al., 2016). It is apparent from the research examples that the last two values are closely related. For example, the development of a software/application where the user can input symptoms and get a disease diagnosis and drug recommendation entails the support (or even replacement) of physician’s diagnosis and the offering of a new service available to people for everyday use. The next value type (V4) includes papers (17.9%) that describe researchers’ efforts to enable experimentation and activities to discover needs, expose inter- experiment performance variability and improve the performance of new BDA models that are used in health information systems for decision-making. This will enable large

55 datasets to be analyzed and hopefully utilized for the social good, avoiding people suffering from resources’ misallocation (Amankwah-Amoah, 2016). For example, by quickly collecting and analyzing reported cases of new diseases due to the use of new software applications for retrieving and manipulating health-related big data, health systems of many developing countries can acquire an early-warning system which will assist public policy officials in a timely and efficient allocation of resources and treatments and provide people with essential health services (Amankwah-Amoah, 2016). The following value type (V5) is gained by the coordination of healthcare information (15.2%) which leads to the simplification of the data service. At the organizational level this can bring operational efficiencies (e.g. better resource utilization) and at the society level the effective information sharing across health and public services can improve the quality of decision towards global health issues (Dinov et al., 2016), such as the guidelines issued by the world health organization (WHO) based on health metrics gathered by national governments. The next value type (V6), creating efficiency (14.3%), is about the capability of BDA to collect data in a standardized format for reducing data identification time, analysis time, and the cost of search and processing while maintaining data quality. This will also provide the opportunity to learn, with less cost and in less time, about populations that were invisible only a few years ago, e.g. in developing countries (Grimmer, 2015) as these populations have now access to the information technology necessary for big data collection and sharing (Hilbert, 2014). Avoiding patient risk was mentioned as a value (V7) in 9.8% of the articles. It refers to the techniques (e.g. logistic regression models and regressions trees) that predict risk in patient health. Notable examples are the prediction of populations at high disease risk and the offering of preventive healthcare, or the likelihood of hospital re-admissions of patients and the recommendation of home monitoring, which promotes people’s health but also save money from unnecessary hospitalization and treatment. This value also refers to the prediction of daily events that challenge the health, security and sustainable growth of our society and prevent social and economic vulnerabilities (Boulos et al., 2010), such as violent behavior, terrorism attacks, emerging infectious diseases, etc. The value (V8) of segmenting populations to customize actions comes next (9%). Here, there are the benefits that organizations gain by capturing the share of new markets deriving from differentiating populations’ characteristics through clustering and other techniques and offering products or services tailored to the specific segments’

needs, but also the societal benefits by identifying isolated or socially excluded patients, for example HIV patients, and offering services that can bring them together (Bram et al., 2015). The value (V9) of achieving cost-effectiveness holds 9% and describes the benefit of analytics to offer solutions for reducing organizations’ expenditures (by optimizing sources utilization, or capturing underpayments, etc.) while keeping up the quality level. These solutions, for example, could include business intelligence visualization and collaborative tools for identifying and eliminating non-value-added processes in patients with chronic diseases by tracking patient data during home, ambulatory and hospital care over time. On a societal level, decision support tools, based for example on machine learning, can provide policy-makers with more “granular information about the health of the population, the prevalence and geography of local factors that are shaping community health and where the greatest potential return on investment might lie if confirmatory research supports a causal link” (Lary et al., 2014). The hope is that policy-makers can use the freed resources to interventions that act positively to society or help poor communities buffer the adverse health effects (Lary et al., 2014), such as organize health education programs, and use public money towards high-risk, low-income patients that cannot afford treatment. The last value (V10) focuses on data security (5.1%), which is enabled by BDA. It may include protection from privacy breaches, securing data anonymity of electronic medical records, etc., suggesting the gain for organizations and society to protect people’s privacy. However, the individuals seem to be likely to accept the ‘dark side’ of datafication through digital traces and constant monitoring through sensors, because they are persuaded that the benefits outweigh the costs. Businesses and governments try to send to citizens the message that security is more important than privacy (e.g. for fighting terrorism or an epidemic outbreak) (Newell & Marabelli, 2015).

Table 17 : Types of Value Creation

Organizational Social Value Definition N % examples context impact impact

V1 New approaches to Analytic Enabling the collection 286 35.6 Banos et al. They presented a framework called diagnosis for approaches that of behavioral and social 2016 “Mining Minds” that enables the personalized provide a personalized data too provision of personalized support healthcare personalized and, on an aggregate through prominent digital technologies service to users level, detecting factors from Big Data and Cloud Computing to with a negative social Wearables and Internet of Things, impact (Kumar et al. investigating people’s lifestyle and 2015). human behavior.

Mohan et al. A software in which the user can input 2016 information, such as the symptoms, and get a diagnosis for the corresponding disease/diseases together with the recommended drugs which can mitigate the symptoms. V2 Replacing/supporting Improve 1.Technology replacing 206 25.6 Benharref, Developed a decision–making system, human decision- decision- human labor creating Serhani, & Al the “Fuzzy Expert System” that relies on making with making and new employment Ramzana 2014 data collected from continuous automated reveal valuable conditions (Loebbecke monitoring (health metrics), to produce algorithms knowledge & Picot, 2015) recommendations (related to food-intake, faster which 2. Challenges the medications, and lifestyle) for both can be made power issues in the patients and physicians by mitigating the available to all doctor-patient risks of chronic diseases. stakeholders relationship (Lupton & Jutel, 2015). Chalmers, Hill, A simulation study using prescriptive Zhao & Lou, analytics to recommend optimal in-brace 2015 corrections for braced Adolescent Idiopathic Scoliosis (AIS) patients after predictive modeling outcomes. The computer-generated recommendations (in-brace correction) improved treatment outcomes and safely reduced aggressiveness of treatment in some cases. V3 Innovating new Big data Innovative information 197 24.5 Angulo et al. Developed a new visualization software, business models, enables tools have the potential 2016 the “BRAVIZ”, that provides real-time products, and companies to to be transformed to statistical analyses of brain images for services create new socially meaningful better patient diagnosis. products and scientific knowledge services, conceptualized as a enhance public good existing ones, (Evangelatos et and invent new al.,2016) business Beyan & Developed GO-WELL, which is a models Yeşim, 2014 clinical “envirogenomic knowledge base” to engage patients to predictive care using genomic data of individual health records to calculate risks of groups who have similar characteristics e.g. families or communities. V4 Enabling Provide Create conditions for 144 17.9 Ziuziański Developed a state-of-the-art e-health experimentation to experimental large datasets to be Furmankiewicz, information system based on knowledge discover needs, applications so analyzed and utilized & Soltysik- management and created a performance expose variability, as for the social good Piorunkiewicz dashboard for monitoring epidemic and improve organizations avoiding people 2012 diseases. performance better manage suffering from performance misallocation of resources (Amankwah- Amoah, 2016) You, Shiaofen, Proposed a new network visualization & Jake 2008 technique (GeneTerrainwhere) where differential gene expression profiles obtained from the human brain are rendered for Alzheimer's Disease patients with differing degrees of severity and compared to healthy individuals. The research can lead to innovative biomarker discovery data explorations for other diseases too.

V5 Coordination of Sharing of Sharing of information 122 15.2 Hughes et al. PROACT is an application for healthcare information and across health services 2016 smartphones which receives cancer information data analysis and countries to patients’ clinical information and drug among improve decision- tolerability and identifies and monitors stakeholders to making for global diet and exercise levels to improve gain health issues patient chances of avoiding cancer operational (Amankwah-Amoah, related problems. efficiency 2016) He,Wang, Gao, Introduced an approach to implement & Tang 2015 cloud data services as a PaaS platform, called CDV PaaS to simplify the data service and balance the increasing demand of limited healthcare resources, making the citizen access to the medical care easier. V6 Creating efficiency Collect data in Learn, in a timely and 115 14.3 Azadmanjir et Described the process of creating a map a standardized less costly manner, al. 2015 for laboratory dashboard as a useful tool format for about population for laboratory managers to improve reducing metrics that were decision-making about costs, orders, processing time unthinkable only a few time, and human resources management. and cost and years ago (Grimmer, enhancing data 2015) quality Boytcheva et al. A research for easier knowledge 2015 extraction by storing information from large amount of clinical narratives in a structured format. Approximately 100 million of outpatient care notes in Bulgarian language were used to apply the method. V7 Identify patient care Avoid patient Prevent daily events 79 9.8 Kulkarni et al. The authors explored the effectiveness of risk care risk that challenge the 2016 statistics for predicting readmission rates through the health, security and in different medical departments for development of sustainable growth of identifying specific patients that have applications our society and prevent high risk of readmission. that provide social and economic clinical risk vulnerabilities (Boulos prediction et al., 2010) Kite et al., 2015 They proved that secondary analysis of Electronic Medical Records improves patient management of chronic diseases, such as cardiovascular disease (CVD) for the identification of at-risk populations. V8 Offering customized Create highly Gather vital 72 9 Abbas et al. A cloud based solution (implemented as actions by specific information to draw 2015 Software as a Service) that provides segmenting segmentations important conclusions personalized recommendations about the populations though the on targeted populations health insurance plans according to the exploitation of (such as HIV patients) user specified criteria. big data and (Bram et al., 2015) tailor products Azmak et al. The Kavli HUMAN Project (KHP) and services 2015 aggregated biology and behavior data precisely to together with environmental conditions meet needs. and events from 2,500 New York City households within a geographic information system database and explained how human health and behavior coexist over the life cycle and why they evolve differently for different people. V9 Achieving cost- Discover new Policy-makers with 72 9 Blakely et al. The authors provided health system effectiveness cost-effective evidence from decision 2015 spending estimates based on patients’ ways to support tools can age and proximity to death using new intervene on the allocate money to heath methods of data analysis and more determinants of interventions which can accurate costing data obtained from the health, with the save more lives and updated health information systems goal of help poor communities network of New Zealand able to collect improving buffer the adverse big data and integrate them. health while health effects of reducing poverty (Lary et al., expenditures. 2014)

Bradley & Demonstrated to healthcare financial Kaplan, 2010 executives how the use of predictive analytics enhances their ability to capture charges and identify underpayments of insurance firms or patients to health

services.

V10 Protecting privacy Ethical Protect patient 41 5.1 Zhang et al. They presented a proximity privacy guidelines to confidentiality to 2016 model of multiple sensitive attributes ensure that prevent unethical that improves the capability of defending BDA supports targeting of groups on the proximity privacy breaches, the the principles the basis of race, scalability and the time efficiency of of respect of ethnicity, or local-recoding anonymization. persons, and sociodemographics avoids illegal (Clift et al. 2014) acts. Fabian, Present a novel application which Ermakova & provides a high level of security and Junghanns privacy for patient data in cloud ,2015 computing environments.

4.6 Challenges from the implementation of BDA in healthcare industry

In this section, results are presented concerning the challenges from the use of big data analytics in healthcare as identified in the article pool (Table 18). The Wamba et al. (2013) three review categories of “data management issues,” “technological issues,” and “organizational issues” have been used to form the table (19) and four more have been added: “further evaluation issues,” “regulatory issues,” “limited awareness and support issues,” and “political issues.” Thus, the articles have been allocated based on the identified challenges. Most of the issues deal with data management, security and privacy, with 20% representation (mentioned in 161 articles). In this category, have been distributed all articles that mentioned ethical issues. This important issue is related to informed consent for sharing, aggregating, or repurposing data related to patients. This further relates to the individuals’ right to maintain their privacy, and the right to be forgotten by erasing their personal data from health and other organizations’ databases. Concerns are raised in the literature about the possibility of re-identification of anonymized sensitive information through cross-referencing or about group-level ethical issues from the analysis of aggregate data, as research outcomes may favor populations (usually westernized) from whom data is collected. Furthermore, the capability of the algorithms and monitoring systems to identify relationships between behaviors and particular individuals raises concerns about the possibly use of this data for the “stigmatization” of social groups or individuals (Rich & Miah, 2017). Other ethical challenges involve the increasing complexity of big dataset analysis, the access to the technology/platforms/tools used for this purpose and the ownership of these big datasets, which raise issues of outcomes’ validity and capability of replicating the findings of BDA. An issue of paramount importance is people’s fallacy that BDA

60 diagnostic applications over the internet may replace the need to be seen by a doctor for diagnosis and drug prescription. In the context of a rapidly development of digital technologies in medicine, the existence of self-diagnosis apps will impact several important dimensions of patienthood and healthcare (Lupton & Jutel, 2015). This is both a challenge and a threat to population’s health. Next are “technological issues” (10.2%), which include challenges stemming from the lack of the required infrastructure to achieve the expected outcomes. On the optimistic side, technology brings advances and data that are produced more efficiently (Alyass et al., 2015). However, missing infrastructure is identified in cases where innovative analytics data-gathering platforms are absent, or information technology systems are poorly connected across and within healthcare organizations. The challenges concerning the society are mainly referred to the inequality among those that have the privilege of the “know-how”, such as technical specialists who know how to interpret and use information technology, and those who have not access to more complex systems (Cuquet & Fensel, 2018). Under the category “Further evaluation issues,” (7.1%) are studies in which their authors stated issues concerning their sample, for example the sample is not representative of the general population or the data is inaccurate or with noise, etc. or there is bias in the research with regard to controlling the threats to internal validity or the risk of individual observer bias (Gruebner et al. 2017). Inadequate samples can create confusing outcomes. “Organizational and financing issues” are identified in 25 studies (3.1%). These studies report lack of cost-benefit analysis frameworks for evaluating the worthwhileness of the use of big data analytics for decision-making; lack of training of information analysts to apply big data analytics techniques in healthcare settings and of healthcare practitioners to comprehend the data analysis and results; lack of skills to support IT-enabled healthcare processes, such as telemedicine; and finally, organizational complexity, such as handling personal healthcare data deriving from different sources (labs, home devices, mobile applications, wearables and other) (Carroll et al., 2014). A small number of papers (5 papers with 0.6%) have been distributed in the category titled “Regulatory issues.” These include challenges for BDA due to the lack of new norms to integrate working practices, aligned with the new technologies. For example, for the creation of effective bio-surveillance systems, the governments should

provide to the healthcare agencies “appropriate jurisdiction to exhibit secure, continuous information ﬂow with no latency across jurisdictional boundaries and to enable detection of previously difficult-to-detect events that span public health authorities” (Velsko & Bates, 2016). Four (4) articles (0.5%) have been distributed in the “Limited awareness and support” dimension that include papers that discuss issues about lack of funding for the completion of projects related to BDA in healthcare, and the lack of awareness for BDA and their benefits to healthcare organizations and to their decision- makers. On the society side, dependency on private and not public funding for the execution of health big data projects will create a few big players in the field which will direct research towards their individual goals which not necessarily overlap with society’s goals (Cuquet & Fensel 2018). The last subcategory is “political responsibility issues” (noted in 3 articles, 0.4%), which describes the obstacles of political sources in the implementation of IT technologies, such as the lack of regulatory responsibility in the case of misdiagnosis and who should be medically liable for the adverse outcome: the developers of the software, the technology provider, the hospital that uses the technology, the doctor, or all of the above? (Dilsizian & Siegel, 2014). All these challenges were discussed by researchers as barriers to the implementation of their proposed approaches or to the outcomes of their research. However, more than half of the articles in the dataset do not refer to challenges.

Table 18: Challenges from the implementation of BDA in healthcare

Organizational Challenges Social Impact N % example context impact Data Issues such as data Privacy violation and 161 20 Blobel Lopez & Highlighted security and privacy challenges of big management, integrity and privacy discrimination (Goh Gonzalez 2016 data and analytics for personalized health, such as security and lead to poor data ,Tao, Zhang, & Yong . bio-, nano- and mobile technologies that allow privacy issues management 2016) pervasive computing. Discussed current decision support systems for Goh ,Tao, Zhang, & dentists and the need to be able to fully utilize Yong . 2016 personalized features of a clinical DSS without the risk of compromising the confidentiality of their patients’ information.

Technological Lack of required Social inequality, as 82 10.2 Zhu et al. 2015 Presented current developments in the fields of issues infrastructure cannot data are only open to a sensing, networking, and machine learning and produce safe small elite of technical described a project entitled “SPHERE”:a generic conclusions specialists who know platform that gathers sensor data to generate rich how to interpret and datasets that support the detection and management use it", and to those of various health conditions. The main challenge is who can employ them the design of innovative analytics data-gathering (Cuquet, & Fensel, platforms. 2018). Wilbanks & Langford, A review of dashboards for data analytics in 2014 healthcare where design issues are yet to be addressed.

Further Issues such as samples False alarms from 57 7.1 Dimitriadis et al. 2015 The authors stated that there were bias issues evaluation relationships or bias missing data or not concerning their sample to test their approach (25 issues adequate samples amnestic MCI patients and 15 age-matched controls (Dimitriadis et al. describing it as middle-sized sample of participants). 2015) Goldenholz et al. 2015 They reported concerns about bias in their simulations because they avoided to include patients who lacked data recordings after the end of the testing period. Organizational Lack of cost-benefit Changes in 25 3.1 Barkley Greenapple, Identified the current state of data capabilities among and financing analysis frameworks, employment & Whang , 2013 oncology providers and the impact of data analytics issues lack of training, and conditions could also on clinical and economic decision-making. This study skills to support IT- raise negative social revealed barriers such as lack of staff or skilled enabled healthcare impact (Loebbecke & workforce to incorporate analytics, lack of care processes, Picot, 2015) coordination, and poor internal communications. organizational complexity Barret, 2013 Described new sources of big data in population health and identified as “unresolved challenges” the funding, administration and accessibility to a merged dataset of detailed health, behavioral, and environmental data. Regulatory Lack of connected Lack of regulation 5 0.6 Fleurence Grandas, & Described the large national research network, issues structures and of new about who has control Meyfroidt , 2014 PCORnet that has been launched by Patient-Centered ways to integrate on data creates Outcomes Research Institute to empower patients and working practices confusion to data their families to generate, collect, and use their health across hospitals and analytic processes. information for both clinical and research purposes. community services They argued that among other “if regulatory, (Warrington, challenges can be overcome, e.g. streamlining the Absolom, & Velikova, consent processes while protecting patients’ rights, 2015) PCORnet will allow research to be conducted more efficiently and cost effectively and results to be disseminated quickly back to patients, clinicians, and delivery systems to improve patient health”. Warrington, Absolom, Discussed the value and challenges for using Patient & Velikova 2015 Reported Outcome Measures and e-Health approaches to support cancer patients care during treatment and their role in the development of appropriate and sustainable long-term follow-up models for cancer survivors. The biggest challenge to healthcare systems and professionals will be the need to re-organize the existing structures and create new ways to integrate working practices across hospitals. Limited Lack of funding and Dependency on 4 0.5 Celler et al., 2014 Demonstrated tele-health services for chronic disease awareness and awareness private funding will management in Australia in a range of hospital and support support few big community settings and developed data analytics players that will tools (modeling methods). They reported limited further lead to awareness and support for telehealth services among international economic clinicians, service providers and patients. competitiveness Manchanda & Jacobs, Various genetic testing strategies for gynecological (Cuquet, & Fensel, 2016 cancers, such as population-based approaches and 2018). genomic information along-with biological/ computational tools will be used to deliver predictive, preventive, personalized and precision medicine in the future, but there is lack of funding and awareness among clinicians about Population-based testing of cancer. Political issues Barriers to adoption of There are trust issues 3 0.4 Dilsizian& Siegel, Presented the advances in Cardiac Imaging using IT technologies from from the adoption of 2014 BDA to provide personalized medical diagnosis and political perspectives IT technology in treatment. Political perspectives (such as lack of healthcare liability of misdiagnosis) must be overcome for the adoption of new technologies.

4.7. The use of Machine Learning in the health field

Machine learning, the most preferred of the analytical techniques for the variety of data types, offers immense potential in the healthcare predictive analytics arena for improving outcomes in many domains of research (López-Martínez et al., 2018). It facilitates the development of patient-centric models for improving diagnoses and intervention. Machine learning is a data analysis method that automates analytical model building. As a branch of artificial intelligence refers to analytical algorithms that iteratively learn from data, identify patterns and allow computers to make inferences and find insights without being explicitly programmed where to look (Breiman, 1996). Machine learning techniques can be used to integrate and interpret complex health data in scenarios where traditional statistical methods cannot perform (Shameer et al., 2018). Usually, a plethora of machine learning models for risk prediction are evaluated to choose the most accurate one. The use of machine learning based methods is important during data collection, dimension reduction etc. to achieve different value creation objectives (ur Rehman et al., 2016).

Machine learning algorithms are proving handy in medical diagnosis that require more accurate prognostic models, such as detecting diabetic retinopathy (Gulshan et al, 2016) and in medical disciplines such as oncology, and where pattern recognition is of ultimate importance, such as radiology and pathology (Cabitza et al., 2017).

Through content analysis of the articles, there are some excellent examples of the application of the machine learning algorithm. In relation to the value for the diagnosis of the personalized health (V1), Bertsimas et al. (2016) developed models that use machine learning and optimization which identify a better combination of chemotherapy drugs and improve the outcome of chemotherapy regimens tested in clinical trials without changing toxicity levels. In line with V1 too, Voisin et al. (2013) identified the best performing machine learning algorithm to predict diagnostic error in mammography by merging gaze behavior characteristics from the radiologist and image features.

To support the business value of “supporting/replacing human decision-making with automated algorithms (V2)”, Lary et al. (2014), used the machine learning

64 algorithms to analyze geospatial data of populations (e.g. smoking-obesity rates, education level, air pollution, existing health and social-support services) and to construct tools for public health data-driven decisions (budget allocation on health interventions based on best return on investment).

A paradigm of a new innovative product (V3) that creates value to the healthcare business is the Wiki-Health service platform that collects, stores, and analyses personal health sensor data which are used for tracking existing health conditions and most importantly predicting them, through the use of machine learning algorithms, encouraging a pro-active approach to healthcare (Li & Guo, 2016). Moreover, for improving the performance of the model (V4), Breiman (1996) used new approaches at that time, such as bagging (i.e. Bootstrap Aggregation) to decrease the variance of the prediction.

4. 8. Future perspectives as derived through the article content analysis

Based on content analysis, all the future goals of every article in the dataset are mined and grouped according to the future directions of research in health analytics. Not surprisingly, it is noticed that despite differences in each research’s approach, the future perspectives were in many cases similar. The content has been classified under three main headings: “technological” perspectives (the future technological improvements upon the presented approaches or the expected future technological improvements in the investigated area), “organizational perspectives (the future improvements of the healthcare processes on privacy, training, information management, cost-reduction, etc.), and “research” perspectives (the future research directions that the discussed approach will bring to academia). In Table 19 there is a summary of the classification of all future directions which provide a map of “the road ahead.” Since the volumes of health data will grow globally in an intense manner, and the demand for Information Technology (IT) infrastructure will consequently increase (Abbas et al., 2015), technology has the greatest role in the future of BDA in health. Under the “technological perspective,” many researchers reported their priority to develop the technological approach, as described or evaluated in their study, so that

65 in the future more advanced versions of their methods will become available. Researchers identified the parts of their methods that have the potential for further improvement with experts’ contribution and expect more innovative techniques for large dataset exploitation and new platforms or new mechanisms, to accrue the maximum value of data. They also expect from technology to extend systems’ capabilities and to improve the accuracy of health data advancing risk adjustment. Innovative digital media technologies are positioned forward for healthcare, to provide better, more informed and more economically-efficient medical treatment for certain groups (Lupton, 2014) also reducing the misallocation of resources (e.g. track patients across service sites, aggregate a bigger amount of data, give a more comprehensive view of data, make the medical record accessible to all caregivers, etc. (Barkley et al., 2013). As it has been declared there is a definite need in health care for systems that support or improve the decision-making ability of clinical experts, specifically, to diagnose complex diseases or pathologies (López-Martínez et al., 2018). Lastly several researchers recognized the space for alternation in their computational approach in terms of proposing new modalities to successfully provide more sufficient results. Amongst the “organizational perspectives”, the most relevant are the future changes in the field of personal data privacy by the development of systems which will standardize and secure the process of extracting anonymized healthcare datasets from healthcare organizations (Al-Shaqi et al., 2016). Ensuring privacy and cybersecurity will enable healthcare organizations and researchers to manipulate these datasets for further value creation (Al-Shaqi et al., 2016). Another future direction is the need of healthcare organizations to train and educate both clinicians and the public regarding the new age of datafication or else (using the term by Lupton, 2014 ‘the digitally engaged patient”) the phenomenon in which people have an active role in producing and consuming information about health and medicine by using digital technologies. Organizations also are expected to integrate environmental factors into the analytics process and decision-making in order to detect environmental hazards, such as gas leakage, or to reduce environmental impact by enabling automatic operation of bathroom/corridor lights, reducing trips and minimizing patient falls and other. Using innovative and computer-aided diagnosis systems and questioning old practices will bring new societal norms and will create the need for new protocols for the treatment of patients (Rodrigues et al., 2016) and the relationship between patients and clinicians. Given these created values from the implementation of data analysis in healthcare, it is

expected that more investments will be provided to IT infrastructure and to individuals with appropriate interest/expertise from healthcare organizations or from nations for healthcare monitoring, such as automated bio-surveillance systems at a national scale (Velsko & Bates, 2016). Scientists also stated that it would be useful to understand the extent to which design decisions are made regarding things like drug prescription patterns, which will affect the cost outcomes (Bjarnadottir et al., 2016). As of right now, little attention is given to organizational future perspective in the establishment of broad partnerships between manufacturers, payers, providers, and regulators in the health-care system to demonstrate and communicate the overall value in medicine that big data could bring (Szlezak et al., 2014). The “research perspectives” focus on researchers’ need for more studies to prove their hypothesis and belief that their approach could also “create value” to other healthcare applications. For example, including real-time sensor data from patient’s devices in the BDA process can further empower the clinical decision support, for the diabetes case, but also for other complex medical conditions, such as Alzheimer’s and psychosis (De Silva et al., 2015). Other researchers envision a future where big data will be used to guide clinical decision-making in real-time, based on individual patient characteristics, where large groups of patient data can be pooled from across institutions so that each patient and their clinicians can find ‘patients like me’ to help with real-time clinical decision-making (Broughman & Chen, 2016). Thus, many authors claim their data analytic approaches must be tested with patient demographic characteristics from other regions or countries to expand the scope of their models (Bardhan et al., 2015) and provide assumptions in a worldwide level. Lastly, there is evidence for a future use of the authors’ proposed approach in other scientific areas. For example, the interpretation of postgenomics data using certain algorithms is expected to be the center of knowledge-based innovations in various big data fields such as precision medicine, nutrigenomics, vaccinomics, pharmacogenomics, ecogenomics, and other promising ones on the postgenomics horizon (Ben-Ari Fuchs et al., 2016). Table 19 presents a summary of these future directions as they have been discussed in the article pool.

Table 19 : Presentation of BDA in healthcare future aspects

Organizational and Technological perspectives N % societal perspectives N % Research perspectives N %

Development of the specific More studies to prove the 133 16.5 Need for privacy 37 4.6 57 7.1 approach hypothesis

Propose an approach that Creation of new mechanisms to Training and education of 101 12.6 20 2.5 can be used in other 31 3.9 accrue maximum value of data clinicians and public healthcare applications Integrating environmental Hardware and software development Patient involvement in 87 10.9 factors in analytics for 10 1.2 15 1.9 -extend systems capabilities effectiveness research decision-making Change the protocols and To replicate same methods Improve risk adjustment 41 5.1 9 1.1 14 1.7 define policy purposes in other countries

Identify data elements that can be More investment in Create value to other 21 2.6 5 0.6 12 1.5 automatically corrected infrastructure sciences

National investments on Alternate the proposed approaches 17 2.1 4 0.5 health monitoring Cost effective analysis of the 3 0.4 new tool Create partnerships among stakeholders to establish the 2 0.2 value of BD

In particular about machine learning in health, as it was identified to be the most used technique, future directions that derived from the dataset of the selected articles should focus on the following perspectives: use of unsupervised learning techniques to more precisely phenotype complex disease; the development of automated risk prediction algorithms which can be used to guide clinical care; and the implementation of reinforcement learning algorithms to intelligently augment healthcare providers. From the technological perspective, an important issue is the strain between accuracy and interpretability. Studies should be directed towards the development of machine learning decision support systems which will automatically provide clarifications, and offer doctors’ interactive visualization tools to examine the implications of potential exposure variables (Batarseh & Latif, 2016; Cabitza et al., 2017). Finally, at the organizational level, an important issue, is the training of doctors to assessing the value of machine learning–based aids in practice and avoid the reduction of the skill for diagnosis or the loss in judgment of the accuracy of the decision-support systems results. This further requires knowledge of how these machine learning algorithms work in practice, therefore it requires the acquisition of statistical and data analysis skills. Of course, as the development of technology is a step ahead of its presentation in academic papers, it can be assumed that recent technological innovations in analytical techniques are creating further opportunities deriving from hidden, up to now, information. Novel analytic fields, for instance, the analysis of data gathered from social

68 media or data retrieved from mobile applications, will likely lead to new information systems for the healthcare sector.

4.9 Big Data Analytics

Because big data are large, processing cannot be performed by traditional health informatics such as “a standalone system” with just a simple analytic software. What is required is a more complex, programming intensive system with a variety of skills (Ragupathi & Ragupathi, 2014). That is in many cases the Hadoop open-source platform. Hadoop released by Apache in 2011, consisting of mainly the Hadoop Distributed File System (HDFS/ a way to divide large data sets in smaller types and store them across multiple servers) and MapReduce (a computational paradigm using two sequences of execution - parallel processing) which includes: a) the map phase that produces interposed key value pairs from initial key-value pairs and b) the reduce phase where the interposed key-value pairs are aggregated by a key and the values are combined together to a final reduction output. HBase is a distributed database built on top of HDFS to provide storage for Hadoop Distributed Computing using ZooKeeper as a coordination service (McClay et al, 2015). First, Google introduced MapReduce allowing big data processing on clusters with Mapping and Reducing. Yahoo developed Hadoop as an open source implementation of MapReduce (Van Poucke et al., 2016). Map/reduce jobs on Hadoop, which can also be developed on Hive (a runtime Hadoop support Architecture), provide a mechanism to project structure on this data and query them allowing MapReduce jobs in other languages when required (Van Poucke et al., 2016). Business analytic tools are faced with many challenges and researchers evaluate them in terms of availability, continuity, ease of use, scalability, ability to manipulate at different levels of granularity, privacy and security enablement or quality assurance (Ragupathi & Ragupathi, 2014). For example, in order to overcome the major disadvantage of Hadoop that is tight coupling between the programming model and the resource management infrastructure, a new architecture was developed, called YARN, that decouples the programming model from the resource management infrastructure and delegates many scheduling functions (Van Poucke et al., 2016). Further, the Apache Pig dataflow system was developed to allow users to easily compose multiple data

69 processing functions because Hadoop MapReduce was restricted to practitioners with advanced technical skills due to the complexity of parallel operations and multi-step data flows (Sahoo et al., 2016). So overall, the computing platform most often used for the BDA tools in general and for the healthcare in particular is Apache Hadoop (Dinov, 2016; De Silva, Burstein & Jelinek, 2015). Additionally, MapReduce is a programming paradigm that provides scalability across many servers in a Hadoop cluster with a broad variety of real-world applications (Belle et al, 2015; Berger & Doban, 2014, Khan et al, 2014; Luo et al, 2016). From the screening of the article pool, 36 papers have been published in 2016 that present applications based on the Hadoop ecosystem with different applications and capabilities. From the most recent literature the presented examples come of a few representative studies with a reference to their data types and techniques and the achieved value. Along with this, there is also a discussion of the technical restrictions that each case brings up and attempts to overcome. An example of the use of Hadoop ecosystem is the one presented in the research of Batarseh & Latif (2016), that introduced a “user friendly” tool called CHESS and has been developed in Visual studio for C#, to read EΗR and provide means for analysts to run queries and experiments. CHESS moves the uploaded datasets to Hadoop and aggregated data, with much fewer rows, are settled to an SQL server for analysis. Then, users access them through the statistical software of their choice (e.g. excel, Tableau, R), and after re-organizing the data in the necessary format can run statistical tests to examine, for example, the importance of some factors (e.g. demographics) over certain health conditions. The application relies on Hadoop for handling big data issues, and the users can query only smaller amounts of data to the statistical software. The application could benefit from more advanced clustering methods to allow for running statistical significance tests to identify important healthcare factors in a more automated way. In the post-genomic era, as the focus of biology has started to shift from mapping genomes to analyzing the vast amount of information resulting from functional genomics research Bodenreider & Burgun, (2005), Cui, Tao & Zhang (2016) describe the evolution of using Hadoop and MapReduce in the scalable and computational powerful cloud computing environment to perform biomedical ontology quality assurance (OQA). This capability has made possible to reduce the standard sequential approach for implementing OQA methods from weeks to hours. With this speed, more

70 exhaustive structural analysis of large ontological hierarchies can be performed and also structural changes between versions for evolutional analysis can be systematically tracked. Areas of further research are around the development of better user interfaces for reviewing OQA results and visualizing ontological alignment and evolution while also increasing the performance of the visual interface by automatically pre-computing intensive jobs while in interaction with the user. Istephan & Siadat (2016) presented a new approach of unleashing the content of unstructured medical data and enabling queries and processing of both structured and unstructured health data for the diagnosis of personalized health. This is a step forward as most applications are limited to being able to query only from structured medical data, such as part of the EHR datasets of a population. For example, when it comes to medical image and EHR processing, there are cloud based software and platforms, such as LifeImage, Nuance mPower for sharing and retrieving big data medical images and other health records, but they are limited to using structured data (e.g. run a query on patient gender) to retrieve all relevant images and records and cannot handle unstructured data (e.g. query based on volume of a brain structure). Other developments incorporate models, even in a Hadoop/MapReduce environment (Yao, et al., 2014), that are related to pattern matching in data medical images. This means than an image is uploaded as an input and then feature extraction and similarity pattern matching techniques are used to retrieve similar images (Toews, et al., 2015). Some technology restrictions that are apparent from the selected papers with regards to the Hadoop and MapReduce environment is that they cannot always handle unstructured content from health data and medical images in the desired way. In order to overcome the problem, researchers create customized tools (instead for example of a Hadoop component like Hive) (Istephan & Siadat, 2016). Such approaches further aid medical experts in getting support for decision-making with automated algorithms.

4.10. Recent Examples of Big Data Analytics Tools for use in Healthcare

In this section, there is a short description of new BDA tools to show some of the health decision making support systems that have been developed with the use of BDA. Their description give hints to health information librarians about how decision- making is assisted through the use of BDA with examples for certain medical specialties. For this part of the study and after filtered the 804 articles and selected only those papers that presented BDA tools. 326 papers have been indentified, 33 of which were published in 2016. Because of the plethora of tools in the investigated studies there is an overview of the most recent developments (published in 2016) and present a mix of tools for different medical specialties and a variety of data types and capabilities, easily understood by non-technical readers. Table 20 includes 13 indicative tools and provides the abbreviation for each tool, a short description, the authors’ country (of origin) and the funding source for the financing of the project. Along with this, information is provided for the medical specialty, the developed tool, the data type analysis and the achieved capabilities (base on the categorization presented in the Literature Review section). The Library of Congress classification (LCC) coding system has been used to categorize the articles according to the appropriate medical specialties to assist health information professionals to identify the bibliographic material. Medicine is classified in LCC under the R class with subclasses of the medical specialties. Based on an analysis of all 33 tools published in 2016 it is observed that most of the applications relate to medical oncology and offer predictive capabilities while the clinical data is the most popular data type analysed by those applications. The USA is the most frequently appeared country of authors’ affiliation in the dataset of the 33 tools, with a great frequency difference from the others. Most tools were used for neurology and oncology specialists. For neurology, the most popular analytic capability of the BDA tools is monitoring, using visualisation software (such as BRAVIZ- presented in Table 20, or a wearable system for home monitoring (such as HMSCSE useful also for geriatrics). Whereas medical oncology, largely analyzed clinical data, either with visualized software for reporting or to differentiate cancer from other diseases (such as CAMA). They also use statistical methods to match patients with their ideal therapy (such as AED). For the cardiovascular diseases, the studies mostly use prediction models, using clinical data (such as the Erlang-2/BG/EG). Overall, in the whole dataset, most articles use clinical data for the purpose of monitoring and

prediction, and they are relevant to the neurology/neurosurgery/neuropsychiatry, medical oncology and cardiology medical specialties.

Table 20: Description of new tools related to Big Data Analytics in Health

Specialty Citation Capability DataType Tooloverview Authors Funding Agency affiliation

NEUROLOGY Angulo Monitoring clinical BRAVIZ:a visualization software that Colombia DepartamentoAdministrativ USA Schneider., provides real-time statistical analyses o de Ciencia, Tecnologia e Reporting RC346-429 Oliver, of brain images. Canada Innovacion, Colombia Charpak&Her nande, 2016

Devinslyetal Prediction Clinical/ AED: antiepileptic drug model USA UCBPharma 2016 pharmaceut prediction system that enables Evaluation Israel ical personalized, evidence-based epilepsy care, accelerating the match between patients and their ideal therapy.

NEUROLOGY/ Lazarou et al, Monitoring Patientsenti HMSCSE: Home Monitoring Greece EU FP7 project Dem@Care GERIATRICS 2016 ment System for Care Support of Elders and remote monitoring of daily RC952-954.6 activities.

MEDICAL Bhuvaneshwa Data – mining clinical G-DOC Plus: A bioinformatics USA 1.FDA CERSI Cooperative Agreement ONCOLOGY r, et al, 2016 platform that handles a variety of Reporting biomedical data and medical images 2. NCI ISRCE RD651-678 for multi-omics analysisand clinical 3. NIH/NCATSCTSA information for biomarker discovery award to explore somatic mutations and cancer MRI images. 4.Georgetown University Medical Center

Guo& Zheng, Data - mining pharmaceut SynLethDB is the first database that Evaluation ical harbors a large set of Synthetic 2016 lethality (SL) - a type of genetic interaction- and performs a systematic evaluation of SLs in anticancer drug discovery and development.

Iqbal et al, Reporting clinical CAMA (Cancer Associations Map Taiwan, 1-3. Ministries of Science and Technology,Health and 2016 Animation): An animated China Welfare, Education, visualization tool to chart the Taiwan. 4-5. Taipei Medical University- association of cancers with other Hospital disease over time.

SURGERY/ Luo et al, Data - clinical DRESS: A double-reading/entry USA, LinkDoc Inc ONCOLOGY 2016 miningReporti system for extracting clinical data China ng from unstructured medical records RD651-678 and creating a semi-structured electronic health record database for further processing of cancer patients data in surgical departments.

CARDIOLOGY Bardhan, Prediction clinical Erlang-2 (BG/EG): A predictive USA 1.UT DALLAS Zheng, & analytics model, which predicts the 2. UT Southwestern RC666-701 Kirksey, 2016 propensity, frequency, and timing of Medical Center readmissions of patients diagnosed with congestive heart failure.

PHYSICAL Calyam et al, Monitoring Patientsenti PTaaS (Physical Therapy-as-a- USA 1. National Science Foundation 2. Department ACTIVITY 2016 ment Service): a telehealth eldercare of Energy service that connects a remote RD58 physical therapist at a clinic to a senior at home and shows how the Therapist is able to monitor Patient status, offer verbal, auditory and visual cues for the patient to perform correct exercise movements.

PATHOLOGY Bjarnadóttir, evaluation, clinical EventFlow: a discrete event sequence USA University of Maryland/ Center for Health-Related Malik, reporting visualization software to investigate Informatics and Bio- RB1-214 Onukwugha, patterns of drug prescription fills imaging Gooden utilizing large scale healthcare data. &Plaisant, 2016

EMERGENCY Chen et al, Prediction administrati GIS: A geographic information Taiwan Fire Department of the New Taipei City MEDICINE 2016 ve system that manages and visualizes Government the spatial Taiwan distribution of RA645.5- demand data and forecasting results

645.9 for the pre-allocation of ambulances.

INFECTIOUS Ali et al, 2016 Data – Administrat ID-Viewer: a visual analytics Pakistan 1. ICT-RD Fund (National ive Technology Fund, DISEASES miningMonito decision support system for infectious Pakistan)

RC109-216 ringPrediction clinical diseases (ID) surveillance. It is a 2. DTRA (Defense Threat Reduction Agency, USA) blend of intelligent approaches to make use of real-time streaming data from Emergency Departments for early outbreak detection, health care resource allocation and epidemic response management.

Gale, Simulation clinical IPC: an adaptable learning England, Daily Telegraph Christmas Charity Appeal, United Chatterjee,, Liveria platform using virtual learning and Kingdom Mellor & distributed simulation for training Allan,, 2016 health care workers, across a wide geographical area, regarding infection prevention control (IPC) that can be accessed from a conventional pc.

CHAPTER A. 5

5. Discussion and Conclusions

Given the large numbers and frequently updating healthcare publications, systematic reviews assist healthcare practitioners to make decisions as they provide summarized research on a given topic of interest (Ali et al. 2018). The aim of the first part of this Thesis is to present a systematic overview of the literature in order to determine the way Big Data Analytics have managed to improve the healthcare domain. Resource–based theory is followed to identify the big data sources and the analytics techniques which allow big data capacities to create values which will continue to fuel through new research in the field. The map of the existing literature on the field of BDA in Healthcare has been achieved using content analysis to provide explanatory definitions of the categorisation through representative examples. Specifically, in the systematic review of the literature of 804 the created values from the big data resources and capabilities in the health field have indentified and extended the research framework to further investigate the emerging challenges and the field’s future perspectives. The explanatory definitions have been provided for the categories and the dimensions based on the literature identifying the number of articles discussing each dimension of values, challenges and future perspectives after content

75 analysis using the NVivo software. The results were presented and discussed through representative examples which focussed both on the organizational and social impact of BDA in health. The descriptive characteristics of the 804 articles which have been reviewed in this study reveal an explosion of publications in the field of health analytics the last years. Medicine and computer sciences are the most common subject areas and there is some multi-disciplinarity amongst authors’ backgrounds in less than half of the examined papers, with the authors of 11% of the papers coming from more than three different subject areas and around 27% of the papers being written by scientists from the fields of medicine and informatics.. The findings show that the rate of publications per year, confirmed the assumptions of other research about the radical growth of publications in BDA from the beginning of this decade (Chen, Chiang & Storey, 2012; Baro et al, 2015; Wamba et al., 2015; Andreu-Perez et al., 2015 etc.). The “Journal of the American Medical Informatics Association” has published a good number of papers in the field. Most health analytics articles are published in US journals and most of their authors work in the US. The domination of the US in publishing articles in the field is not surprising and it has also been reported in other reviews of the medical and informatics field (Brown, Gutman, Ho & Fong, 2018). Zhang Yin from the Zhongnan University of Economics & Law, China is the main author who mostly appears in publications (5 articles) and Stanford University is the authors’ affiliation with the biggest number of papers (20). The top author affiliations are also located there and the same applies with the funding agencies which (financially) support the research in the field. For example, the National Institutes for health have funded 54 studies in the article pool and the National Science Foundation has funded 16. Moreover, Bates et al. (2014) is currently the most cited paper of the 804 papers and Murdoch & Detsky (2013) is the most co-cited paper amongst the references of the 804 papers. Both are overview papers, describing applications of big data analytics to health care. Chen, M. from the School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan, China is the most co-cited author. As expected, big data is the most common keyword of the papers in the article pool. Although many different stakeholders were identified in this study, overall, it can be said that the final beneficiaries of BDA research in health are the patients as its future outcome is mainly targeted to ways of improving their health and their quality of

76 care. Most of the examined papers follow an “experimental” approach. Many of the papers in the article pool deal with the medical specialities of neurology, medical oncology and cardiology. Machine learning is a popular big data analytics technique that researchers use in these studies. Almost half of the papers are predictive in nature. One third of the studies develop monitoring capabilities and the 70% of them use clinical data. The indicative examples provided for each subcategory aim to improve the comprehension of the categorization and to further increase the general understanding of the type of research conducted in the field of health analytics. Summarizing with the examples, indicative is the research in the field of population health management in terms of a) disease surveillance by determining disease outbreaks (mostly using social media and web analytics) and ensuring speedy response and needs in new vaccines (Lazer et al., 2014; Boulos et al., 2010; Ginsberg et al., 2009), and mostly b) (chronic) disease management by prediction of disease by patient profiling, in terms of symptoms, lab results, medical images and patient history details, for individuals’ accurate health diagnosis, by applying advanced analysis (such as segmentation and predictive modeling, machine learning, visualization, etc.) (Krumholz, 2014; Delen, 2009). Likewise, medical staff, through decision support systems, can identify the most fitting treatment and medication for each patient to avoid possible complications (Barret et al., 2015) and identify patients with high health risk profiles to offer proactive care options such as screening, brief interventions, etc. (Dugan et al., 2015, Bates et al., 2014) so as to avoid hospitalization or readmissions (Bardah et al., 2015; Demir, 2014; Bardhan et al., 2014). On a similar thematic area, research is concentrated on offering customized e-healthcare solutions, mainly at home, but in hospitals as well, by constantly monitoring and analyzing inbound clinical data from wearables and sensors and alerting health specialists about negative trends for conditions that need attention and possible hospitalization (Althebyan et al., 2016; Baum, 2010). Another very important type of research that emerges from the indicative examples is with regards to the use of big data technologies handling biology large datasets, such as sequencing human genomes to understand biological pathways and the genomic variation of e.g. tumor, which have led to personalized medicine, meaning offering different therapeutic schemes based on a patient’s biomarkers (Alyass et al., 2015; Chawla et al., 2013; O’Driscoll, 2013).

A different type of research is about the provision of cloud services and mobile software that accumulates specific disease-based knowledge deriving from the collection of a vast amount of data from its targeted users and offers them “personalized” consultation for better disease management (Abbas et al., 2016; Quinn, 2008). From the organizational perspective, there are studies that focus on the improvement of health services processes and cost-reduction by evaluating performance of resources, monitoring workload and human error, understanding clinical and other processes and identifying bottlenecks in care quality with the use of modeling and simulation (Sir et al., 2015; Catlin et al., 2015). A similar type of research also focuses on more customized products/services, such as health insurance plan offerings based on the modeling and forecasting of a vast number of individuals’ clinical and administrative data (Abbas et al., 2015). With regard to BDA values, the most highly discussed value is that analytics provide “new approaches for the diagnosis and prognosis of personalized medicine” (35.6%), which has become more broadly applicable across healthcare (Yin & Kaynak, 2015). It includes the process of genetic profiling to offer individual health information for a variety of diseases (e.g., cardiovascular disease, cancers, diabetes, etc.), enabling the use of personalised therapeutic schemes. This value also enables early detection of factors that could create a negative social impact (e.g. people with stress issues, anti- social behavior, etc.). Another important value, is the capability to “support human decision-making with automated algorithms” (25.6%). Researchers need to quantify the real value of the analysis of the available data in everyday clinical practice. This value, on one hand creates new employment conditions and the demand for the acquisition of IT skills which in turn will lead to changing educational curriculums for health professionals, and on the other hand changes the doctor-patient relationship, giving more information to patients to challenge doctors’ knowledge. Some of the findings regarding the values are also supported by the review of Gaitanou et al. (2013), which points out that the main positive effects from big data processing in medicine are: positive behaviour change, improved usability and efficient decision support. As modernity has led to many changes in everyday social life, a remarkable change has been the expansion of medical activity through medical innovation in a variety of new areas (Lowton et al., 2017). In this study, there is also the presentation of the organisational and social values from the analysis of health-related big data through a review of the literature. This research also intends to add more knowledge to the

78 global literature by contributing to the identification of key challenges society faces by the explosion of analytic capabilities of health organizations. The main findings conclude that big data analysis offer new ways for the diagnosis and use of personalized medicine and for supporting decision-making with automated algorithms. However, there are challenges related to data management, security and privacy, relating also to issues of informed consent for sharing or aggregating patient data and issues regarding the effects of technological change to employment conditions (new required capabilities and skills in computer science, statistics, etc.). The development of the discussed approaches reveal the need for more organized ways for securing the privacy of personal health data, for hardware and software investment and development to extend systems capabilities to accrue maximum value from health-related data and the education of health professionals and patients to big data analytics. Data management challenges, which are common among fields using technology (Luo et al., 2016), were further discussed in the papers of the article pool. Thus, the most highly stated challenges are related to “data management, security and privacy issues” (20%), in line with the findings of Wamba et al. (2013). The topic of data privacy has become increasingly important nowadays because of the rapid development of new forms of data, and the ease of transferring and sharing data. Data Protection legislation differs between countries as each country protects medical and health-related data at different level. For example, the rising concerns about data privacy have led to the General Data Protection Regulation (GDPR), enforceable from 2018, which strengthens data protection for all individuals within the European Union, and makes the export process of personal data outside the EU more rigorous. Under this regulation, pseudonymized data are still considered personal data, which means that more BDA health projects that use pseudonymization will now require either consent or authorization (Rumbold, 2017). Issues related to technological challenges come next (10%) and mostly relate to a lack of appropriate infrastructure for supporting big data analytics in health-related organizations and to the inequality among specialists’ knowledge of using big data analytics and their access to more complex systems (Cuquet & Fensel, 2018). In the reviewed papers, a number of issues for future directions emerged. Since the volumes of health data will grow globally in an intense rate, the demand for Information Technology (IT) infrastructure will consequently increase (Abbas, Bilal, Zhang & Khan, 2015). Another field for future research is the assurance of data privacy

79 and cyber security which will enable healthcare organizations and researchers to safely manage/ exploit the big health datasets for further value creation (Zhang et al., 2014). Further research is also desirable towards enabling clinical decision-making in real- time, based on the patients’ individual characteristics, where large groups of patient data can be pooled from across institutions so that each patient and their clinicians can find ‘patients like me’ to help with real-time clinical decision-making (Broughman & Chen, 2016). From the presentation of BDA software, based on Hadoop ecostystem or the MapReduce process, this research confirms that most users use clinical or medical structured or unstructured data for their studies to build new approaches for the diagnosis of personalized health and to invent entirely new business models to reduce time, cost of search or processing while maintaining quality. From the articles in the dataset it is clear, that there is demand for research in health analytics to focus on improving the technological aspects. There is a definite need in healthcare for systems that support or improve the decision-making ability of clinical experts, specifically, to diagnose complex diseases or pathologies (López-Martínez et al., 2018). Progress that has been made via Hadoop and MapReduce has increased performance by reducing time and pre-computing computationally intensive jobs (Cui, Tao & Zhang, 2016). The main difficulty with big data in healthcare is that most data are often unstructured, which means that there are obstacles to computationally process the largest part of them (Dinov, 2016). That is why scientists are in a continuous effort to advance infrastructure in order to achieve the greatest possible analysis and to further develop computational methods in order to extend systems’ capabilities. It is expected that more investment will be given to IT infrastructure and to BDA experts in the health sector, or from nations for health monitoring, or for the development of systems that can track patients’ health-related data across health services and home and make these accessible to relevant professionals. This study aims to verify the status of published research on Big Data Analytics in Healthcare and create a descriptive classification. Trends in big data, and publication statistics have been widely adopted in the scholar community, but such use has yet to be analyzed in depth. This research adds to the existing body of knowledge and provides a more thorough analysis of the field with the use of content analysis, through easily comprehensible information based on examples. It also attempts to continue the effort

80 of other researchers (Waller & Faucett, 2013) to explore the possibilities of big data in OR. This research is multifaceted as it deals with different health issues (different diseases or quality of care), it examines them from a different perspective (for monitoring, reporting, prediction, etc.) and with the use of different types of data (clinical, administrative, pharmaceutical, etc.). Overall, considering the distribution of papers per medical specialty it is noticeable that BDA have a crucial role to play in the research of the most severe diseases that humanity faces nowadays (cancer, Alzheimer, diabetes, etc.) and reveal the importance of innovative technological solutions to unanswered medical questions. Researchers from different disciplines (medicine, information technology, operational researchers, business administrators etc) collaborate to gather and actually use the vast amount of data that cannot be managed from the commonly implemented technology. New modeling and machine learning methods explore new capabilities and reveal hidden information. Since OR professionals are in the front line of offering improved decision-making via innovative modeling tools, this study provides them with the big picture of the big data analytics research that has been conducted in the health sector. The presented overview aims to answer questions like, when (chronologically), where (country and publishing journals) by whom (popular authors, subject areas) and what (medical specialties, research approach, nature of analytics) research is conducted in health big data analytics, with what means this is achieved (BDA data types, techniques) and through what competencies for health-related organizations and information analysts (BDA capabilities). Therefore, firms in the healthcare industry, in the private and public sector, started incorporating big data analytics for strategic decision-making (Gandomi, A., & Haider, 2015). However, the major reason behind big data analytics non-adoption is that first firms do not realize their strategic value and also their managers are not prepared to bring the changes because of technological or organizational difficulties (Gupta et al, 2018). This profiling study has also the ambition to act as a trigger for health organizations to redesign their strategies towards a greater adoption of big data analytics and harness their capabilities to improve service, mitigate risks, reduce costs and grasp new opportunities In the dataset, 326 tools/applications are presented or reviewed. From these tools, some are already known by IT professionals and BDA researchers (e.g. Hadoop)

81 and some are introduced for the first time (e.g. BRAVIZ). From the dataset, it is obvious that in 2016 scientists published articles introducing new applications applied mostly in medical oncology and neurology/neurosurgery, with predicting and monitoring capabilities using, in their majority, clinical data. Many of the BDA tools are related to medical oncology and neurology/neurosurgery. It has been suggested in the literature that for certain diseases e,g. dementia, there is a demand for software engineers to design and develop more applications to help patients. It would, therefore, be beneficial for researchers to provide a road map for the investigation of this area (Asghar, Cang & Yu 2017) by improving current predictive, diagnostic and preventive models, optimizing resource allocation, and delivering more personalized treatments to patients with specific disease trajectories (Ienca, Vayena, & Blasimme, 2018). The ambition of this Thesis is also to present a research that can assist health information professionals in reinforcing their role in the medical and informatics scientific research providing practical steps to explore these in detail, e.g. investigation of the new technology applications of the field and the identification of new patterns for diagnosis and prevention of diseases. In terms of health and well-being, the analysis of big data offers a potentiality for providing enormous prognostic interventions, novel therapies, and shaping lifestyle and behavior, while it is the key to cost efficiencies and sustainability of healthcare infrastructure (Kambatla et al., 2014). Enhancing the breadth and depth of our knowledge about the major and minor aspects of the related field, the healthcare and informatics community can contribute to the creation of new approaches to strengthen outcomes. Methods such as machine learning that has been referred as a necessary tool in problem solving (Obermeyer & Emanuel, 2016), will open-up new perspectives in healthcare since the intelligent elaboration of more and more data will bring new evolution for the prognosis, diagnosis and treatment of diseases.

CHAPTER A. 6

6. Limitations

However, this research comes with limitations. The boundaries between big data analytics and data analytics as well as the boundaries between techniques and other subcategories are not always easily discernible, and therefore a small fragment of the article selection or the categorization may be debatable based on the reader’s point of view. Furthermore, this is a sample of the related literature and by no means is an exhaustive literature review of the field. There seems to be a deficiency of studies relevant to population and public health compared to these related to medicine. Although both keywords “health” and “medicine” were used as a sampling method and the results demonstrate that papers are distributed to a variety of different disciplines (Table 2c), it is true that the majority of the derived papers are more related to medical advancements and clinical decision support. This may imply that more research has been conducted towards this direction. The search on the two databases was conducted in December 2016. Reviewing 804 papers was a very demanding and time-consuming process, which took several months. Moreover, due to the explosion of the publications in the field (for example, only in 2017, 2735 new papers were identified according to the search criteria only from the WoS database without including Scopus), the selection and studying phase of the new material would take other two years. Therefore, it is inevitable to overcome the time lag between the year of publication of the reviewed papers and the time of the presentation of the synthesis of their findings. For this reason, one big limitation of the overview is the lack of more recent material (from 2017-19). Future research would include update advances in information technology in BDA. Furthermore, the classification system is not exhaustive, and it could be expanded to include further categories and subcategories. Only the categories that could clarify certain questions in the area of big data analytics in healthcare are included and specifically with regard to what medical specialties have benefited the most, what kind of data are used for the analysis, what techniques are used, what is the purpose of the analysis (capabilities) and what is the level of analysis that has been reached (predictive, etc.). For the systematic review, articles were only obtained from Web of Science® and Scopus, which are, however, the world's largest multidisciplinary abstract and citation databases and the two mostly used in literature search (Aghaei et al., 2013) and comprise citations from other databases, such as MEDLINE and Biological abstracts. Furthermore, this study uses as a methodology a systematic literature review approach

83 which could be broaden to include many other aspects of health sociology, such as the effects of the commercialization of health data by organization, the changing environment of labor etc. The scope was to offer an overview discussion about the positive and negative impact of the health big data analysis in society along with the values and challenges created under the organizational perspective. For example, sociology can shed light on issues of data ownership, revealing where points of exploitation occur, and on issues of healthcare providers’ responsibilities and their capacity to enforce or discourage certain behaviors (Rich & Miah, 2017).

CHAPTER A. 7

7. Future research

Mapping the existing literature facilitates the health information professionals to draw new opportunities for further development. Future research under this agenda could investigate the benefits and the values created by BDA in healthcare and could focus on the new tools that are used for the analysis of the vast amount of data in the domain along with issues that restrict its extensive use. Therefore, further research could broaden the categories of this overview with more technical content (tools & applications) or with the identification of issues, benefits and future perspectives of BDA techniques and capabilities. Furthermore, it would be very interesting to follow the results of the presented applications over the next years with an updated study that could identify its contribution over the related medical fields. Conclusively, to take advantage of the use of BDA, and reduce risk from “false positives,” it is certainly important to set the focus on the exploration of the information provided from big data (i.e. what information, from which sources, for what purpose it was collected, what is the intention of the analysis, what should be explored) (Strauß, 2015). It is unknown the extent of how digital technology and the associated big data analytics are going to impact society and business in the long term. Methods such as machine learning, which has been an essential tool in problem solving, open up new

84 perspectives in healthcare (Obermeyer, & Ezekiel, 2016). This is because the intelligent elaboration of more and more data brings new evolution for the prognosis, diagnosis and treatment of diseases. After all, the desire of medical experts is to create new decision support systems that reflect upon their “intuitive thinking” and minimize or even eliminate “personal biases” (Oztekin et al., 2018). A stated threat for the future is that algorithmic decision-making may lead to an extremely superficial understanding of why things happen, as answers will be prescribed from a “black box.” This will prohibit decision-makers (such as clinicians) to build cumulative knowledge on phenomena and diseases, which consequently, may cause them to lose their capacity to make decisions on their own (Newell & Marabelli, 2015), and therefore, be fully replaced by big data analytics. Big Data research should begin with a clear understanding of the value it can bring (Flechet et al., 2016). The demonstration of the “bright and the dark side” of the datafication in the healthcare industry can shed light to some of its dilemmas. Mapping the existing literature can facilitate health-related organizations and the society to recognize at first the impact from big data analytics and then revalue strategies, mitigate risks, and draw upon new opportunities for further development. Finally, mentioning that this research focuses on the contribution of big data analytics in the healthcare era a following research could also be oriented in the association between personalization, data quality, and data risk on the adoption decisions and find the link between social connectivity and the adoption decision (Koumpouros & Georgoulas, 2020). As it has been revealed in the first Part of the PhD Thesis the analysis of different types of data generates a lot of opportunities and challenges including dimensions of social, medical, and business services (Sumarsono, Anshari & Almunawar, 2019). Communications promote the development of emerging systems and applications for healthcare called mHealth can capture, store, retrieve and transmit various kind of data to provide instantaneous, personalized informatics and can be useful in monitoring health status and improving patient safety and quality of care (Madanian et al, 2019). For that reason, in order to discuss how health care services can use existing mobile healthcare systems in order to support healthcare services, the second Part of the Thesis proposes an analysis of the model of mobile health (mHealth) accommodating multi sources data channels in supporting mHealth services for better decision making.

PART B

BIG DATA ANALYTICS AND THE USE OF MHEALTH IN HEALTHCARE

CHAPTER B. 1

1. Introduction

Recent trends in healthcare systems have shifted toward eHealth (electronic healthcare) and mHealth (mobile healthcare). The importance of these systems has tremendously increased due to the coronavirus (COVID-19) pandemic outbreak and the extraordinary need for real-time information sharing and fast decision-making (Soltanisehat et al. 2020). Therefore, mobile health applications (mHealth apps) are among the most-discussed issues for healthcare innovation, and they have potential to

86 bring revolutionary insights to the clinical research environment, providing opportunities as well as challenges (Cleary 2018). The term “mHealth” first appeared in 2003 and describes the “medical and public health practice supported by mobile devices” (Kay et al. 2011). The concept comprises a broad range of technologies, including wireless, mobile, wearable and healthcare apps (Vesselkov et al. 2018).In recent years a large number of mHealth apps have been developed and widely used, making mobile technologies emerge as a powerful tool in the health care industry (Sedrati et al. 2016). Better understanding of digital technology creates value for businesses. In particular, mobile apps provide to health professionals with new opportunities to create value, for example by meeting new demands, increasing efficiency, supporting knowledge-sharing and improving competitiveness (Ehrenhard et al. 2017). Medical practitioners have begun using apps and other digital technologies as part of their practice, and it is also common for medical students to use a number of apps- such as anatomical atlases, reference tools, and question banks- for their education (Ellaway et al. 2014). Thus, the new digital-native generation of millennian students that search for and use educational content from digital devices (Montiel et al. 2020). A 2012 survey about mHealth found that 87% of doctors use mobile devices in their workplace and 85% of faculty members and students of medical schools use them in a wide variety of clinical settings ranging from classrooms to hospitals (de Camargo 2012). Even 80% of physicians aged 55 and above own a smartphone (Ventola, 2014), and this percentage will have further increased in recent years. Many positive outcomes of mHealth have been reported, such as faster or more efficient delivery of care to patients, better monitoring of diseases and increased productivity of healthcare service providers (Chatterjee et al. 2009). Mobile technology can have a positive impact on healthcare delivery processes (Free et al. 2013), offering connectivity and accessibility to patients records in addition to a wide range of advanced capabilities (Lomotey & Deters 2018).The benefits offered to medical professionals can be especially valuable for their day to day activities and for making diagnoses (Goldhahn & Spinas, 2018; Fox and Connolly 2018) by leveraging the data that are collected from various sources (Sarker et al. 2020). The most recent mobile applications utilize advanced technologies, such as artificial intelligence (AI) to incorporate expert systems, speech recognition, machine learning, machine vision and others. As the adoption of smartphones increases, mobile apps are a natural way to

87 deliver machine learning algorithms (Zakhem et al. 2018) and there is a rising demand for mobile apps with AI services (Ji Wang et al. 2018).Artificial intelligence systems simulate the human mind by learning, reasoning and performing self-correction, and these systems can even be more accurate than physicians in diagnosis in specialties such as surgery, radiology, dermatology and intensive care (Goldhahn & Spinas, 2018). Several types of AI have already been employed and the applications mostly involve diagnosis and treatment recommendations, patient engagement, education and administrative activities (Davenport & Kalakota 2019). Their development has been lately accelerated also by the need for new diagnostic and therapeutic methods in the 2020 coronavirus disease pandemic (Kernebeck et al. 2020).The most useful application of AI mobile apps for health professionals is to generate a better decision using the AI technology in order to report the status of the patient’s health and provide advanced disease control, customization and personalized diagnosis and treatment (Alotaibi 2020; Siuly & Zhang 2020). Nonetheless, the innovative use of apps for health purposes presents many challenges ranging from ethical issues (e.g. privacy) to credibility issues (e.g. accuracy of content). For example, there are privacy concerns related to ownership of the personal health data collected from app users and the right of mobile app companies to sell or store the data (Galetsi et al. 2019). Reliability issues relate to whether and how the information provided by mHealth apps are evaluated - there are reported cases of mHealth apps offering dangerous and harmful advice to medical professionals (Wisniewski et al. 2019). However, some have argued in favor of replacing doctors with these mHealth apps that provide automated decision-making, which in certain specialties is said to be more accurate (Goldhahn et al. 2018). “IT consumerization” - the adoption of personal IT for decision-making in workplace- may influence how change occurs in organizations (Junglas et al. 2019). Therefore, it is essential to both explore these new working models that medical professionals use to enhance efficiency in the workplace and also to set rules for acceptable mHealth apps that respect privacy (Nerminathan et al. 2017). The doctor-patient relationship (Jutel & Lupton 2015) is also challenged with apps that are marketed directly to consumers for evaluating their own health. But also, from the moment that personal health information is digitized and entrusted to healthcare professionals and the app developers as vendors of the technology that manage health information systems (e.g. electronic health records), questions arise

88 regarding how this information is used and protected (Fox & James 2020).The collection of personal health data from app users also raises privacy concerns related to data ownership and its right to be sold or stored by app companies. A recent example related to the containment of the pandemic COVID-19 is the use of AI apps for digital contact tracing at the societal level for which people actively resist adoption because of privacy concerns (Riemer et al. 2020; Kumar et al. 2020). Therefore, mobile devices present new challenges to the security and privacy of users where sensitive data could be vulnerable to attack by third parties(Ismagilova et al. 2020). Due to the fact that it is difficult to regulate this enormous and growing volume of apps, one practice that requires careful scrutiny is that of developers' privacy policies and data security practices (O'Loughlin et al. 2019). Since 2018, Europe has experienced a major change in data privacy regulation under the EU General Data Protection Regulation (GDPR) (Regulation 2016).More specifically, the “Privacy Code of Conduct” on mHealth apps, facilitated by the European Commission (European Commission 2018), consists of practical guidance to app developers on data protection principles and addresses notably topics, such as purpose limitation and data minimization (data may be processed only for specific and legitimate purposes), data retention(personal data may not be stored longer than necessary), security measures to ensure confidentiality throughout users’ navigation and interaction experience, etc. In the US, the Food and Drug Administration (FDA) has developed three distinct categories of mobile health apps to define the requirements of regulatory oversight. The Federal Trade Commission compiled a list of best practices to advise the development of health apps, such as for minimizing collection of user data and limiting both access and permissions to the users' phone (O'Loughlin et al. 2019). Beyond privacy concerns, there are cases of apps that have been reported to offer dangerous and harmful advice to professionals responsible for decision-making in clinical care (Wisniewski et al. 2019). mHealth apps are of variable quality, ranging from those that appear to have the support and input of valid health organizations, such as the American Heart Association and from distinguished medical experts to those that offer little to support their knowledge claims. The lack of information provided by many app developers also raises questions about how users can be reassured of their quality and safety. A lack of meticulous testing the efficacy, reliability and accuracy of these applications is reported (Wisniewski et al. 2019; Sedrati et al. 2016). Hence, due to the fact that health professionals use mobile devices to enhance efficiency in the

89 workplace, there is a need to investigate these new operational models in everyday clinical practice and a need for guidelines for acceptable and ethical use that respects privacy (Nerminathan et al. 2017). Despite rapid technological innovation and the variety of attractive medical apps, society demands a more regulated environment concerning their use (Cleary 2018). Many studies report the lack of regulation in health promotion (McKay et al. 2018) and the limited research evaluating mHealth apps’ accuracy in giving medical advice and the theoretical foundations that underpin them (Payne et al. 2015). Therefore, further investigation is needed to evaluate the quality and effectiveness of these apps (McKay et al. 2018). According to Sadegh et al. (2018) the outcomes of the use of mHealth can either positively or negatively affect stakeholders, who include the society (patients, their relatives and all people who are affected by the service), health professionals, public or private health organizations service or infrastructure providers, and legislative bodies (Sadegh et al. 2018). To sum up, the mHealth apps for professionals are of multiple utility ranging from apps useful for educational and diagnosis purposes to apps useful for improving operations management in a clinical routine. The apps are also of variable validity, ranging from those that have the support and input of known health organizations, such as the American College of Cardiology Foundation and distinguished medical experts to those that do not provide identification of their knowledge claims. Therefore, certain app characteristics, such as the identification of the information source, may affect the app’s perceived trust by the user. The lack of information provided by many app developers raises questions on how users can be reassured of these apps’ quality, safety, reliability and accuracy (Sedrati et al. 2016). On the other hand, in order to increase revenues from freely available applications, app developers provide in-app purchase options and/or in-app advertisements that may cause annoyance and disapproval, as it displays distracting features, such as animated banners, pop-ups and floating advertisements that could decrease the demand of an app (Ghose et al. 2014), especially when targeted for professional use. Triggered by the rapid development of available mobile apps related to health diagnosis and decision-making, this study’s goal is to analyze available mHealth apps targeted at health professionals and students for supporting the diagnosis procedure. For that reason the selected applications and their intention of adoption were examined by different angles to highlight each time a different way of thoroughly investigating these

90 apps for professionals. Thus, the second part of the PhD Thesis provides a multi-layered analysis consisting of three investigations using three different conceptual frameworks in order to explore and explain the introduction of mHealth in professional clinical routine. The three parts of the research will be structured as above. a. Examination of the ethical issues that are incorporating using Communication Privacy Management (CPM) theory to explain the process of appropriate mHealth apps’ selection based on their privacy, security and reliability features giving emphasis to Health apps that use artificial intelligence. b. Examination of the innovative adoption of digital tools (use of apps) in health diagnosis, using the Normalization Process Theory (NPT) to explain the social context of how users and managers relate to the system, how the system fits with the working environment (Sharma et al. 1991), and how health professionals can engage with certain categories of health apps and use them in clinical practice. c. Acknowledging the tremendous explosion of mHealth apps and the increasing use of mobile devices by medical professionals, it becomes important to understand the underlying drivers of user demand for mobile health apps. Therefore, the objective of this study is also to investigate the factors of successful integration of health smartphone apps in clinical routine and to identify the factors that inﬂuence professional consumer behavior. Multiple regression analysis and inferential statistics was then employed to ﬁnd the association between the groups of variables, such as perceived usefulness, perceived ease of use, perceived trust, app quality evaluation and intention to download the application using a combination of the Technology Acceptance Model (TAM) and the Theory of Reasoned Action (TRA). The study provides several recommendations to both health professionals and app developers. Given the reported deficit of knowledge about mHealth apps’ social dimensions (McKay et al. 2018; Payne et al. 2015; Nerminathan et al. 2017) and the little evidence to demonstrate the effectiveness of the uptake of mHealth technologies (Cegarra- Sánchez et al. 2020), this research is the first study that investigates the social dimensions of the integration of mHealth practices in health environments, such as hospitals, medical schools and clinical workplaces. It presents an analysis of the factors that affect the implementation of innovative practices in health diagnosis, describes the requirements for their use in daily clinical practice and reveals their level of usefulness

91 and trustworthiness for professionals. It also enlightens app developers and information scientists about the importance of the existence of certain features and provided information available in the app stores to help further the engagement of people that intend to use the specific technology in the workplace for decision-making. Additionally, this research informs the stakeholders about popular health app types and the interrelationships of app features with app performance measures that make them more attractive in the specific professional field and what innovations and improvements could enhance the process of engagement. This research also contributes to the need of studies that examine expert systems in mHealth and their general benefits to medicine and the society. The next section provides the literature review upon the scientific field in general and about the related ethical challenges. The methodology section describes the followed methods for the research. Three different studies have been conducting based on three different conceptual frameworks in order to provide a holistic view of the investigated topic. The results of the analysis will be presented along with the conceptual frameworks to reveal the three aspects of investigation of this scientific topic based on the adoption of information technology innovation intervention practices in clinical routines. Therefore, this research refers to various fields ranging from information systems and digital marketing to health and social sciences, contributing towards the need of studies that examine consumer behavior under different information conditions for enhancing our understanding of new e-markets (Grover, Lim, & Ayyagari, 2006).

CHAPTER B. 2

2. Literature Review

In the literature several studies investigate health apps related to patients and consumers. A number of them examine the effectiveness and features of apps related to monitoring and managing chronic diseases (e.g. Donker et al. 2013; Wang et al. 2014)

92 or to health and fitness apps (Cowan et al. 2013; Higgins 2016). However, further investigation into mHealth app features is worthwhile given the inability of the most common features to explain a large portion of an app’s rating (Mendiola et al. 2015). More recently, a survey study found that prior IT experience and perceived self-efficacy positively influence patients’ intention to adopt mHealth apps provided by clinics or hospitals (Balapour et al. 2019). Despite the number of studies evaluating mHealth apps for patients and consumers, there is very little research focusing on mHealth apps for medical professionals. There are a couple of overview and review studies on the subject that conclude that the use of mobile devices by health care professionals will transform many aspects of clinical practice (Cleary 2018; Ventola 2014) and especially these apps that use artificial intelligence and deep learning algorithms will revolutionize the everyday clinical routine (Rajkomar et al. 2019; Alotaibi 2020).A meta-analysis study investigated the effectiveness of mHealth interventions for supporting health-care providers and found that mHealth modestly improved aspects of clinical diagnosis, management and communication between healthcare providers and patients but mHealth attributes, such as diagnosis based on photo sharing, had a negative effect (Free et al. 2013). With regard to studies about mHealth apps there is a main focus on patients or consumers. Several have investigated the usefulness and functions of health and fitness apps (Cowan et al. 2013; West et al. 2012; Lister et al. 2014), monitoring and managing chronic diseases (Donker et al. 2013; Blake 2008; Beratarrechea et al. 2016; Li et al. 2019),or their effect on health behavior interventions (Zhao et al. 2016; Wu et al. 2012).There are also studies about mHealth apps’ acceptability and utility (Robinson et al. 2013), identifying which are suitable and practical for administering health interventions and can benefit users and consumers (Payne et al. 2015; Donker et al. 2013). These studies also identify that more rigorous research is needed to evaluate and determine their efficacy and provide evidence for best practices (McKay et al. 2018; Payne et al. 2015). Some studies have examined user experience and desired functions (Lyles et al. 2011), such as the existence of sufficient information, users’ evaluation (McKay et al. 2018), and features that save time and are simple and intuitive to use (Mendiola et al. 2015). Text explanations, visualizations, examples, simplification, and feature relevance are the most common techniques humans use to explain systems to others, who can then process the information. The selection of the most relevant technique depends on the

93 audience for which the explainability is targeted. Feature relevance explanations methods clarify the inner functioning of a model and the importance it places on each of the such variables when producing its output (Arrieta et al. 2020). Mobile apps can be seen as a kind of consumer product (Hsu & Lin, 2015) and the fact that there are over 300,000 healthcare-related apps in the market (Zulman et al., 2016) makes it impossible for any clinician or patient to remain aware of all health apps. The greater accessibility of consumers to online marketplaces the greater the number of options, thus consumer choice behavior has become more and more important to business managers (McKie et al. 2018). There have been some efforts in the literature to understand the selection criteria of a user when downloading a health app, considering their variety in app stores. A study that investigated 129 urology apps to identify predictors of the number of downloads found that: apps developed with urologists’ involvement, with higher user ratings and number of written reviews, lower or zero price and optional in-app purchases were more likely to have more downloads (Pereira-Azevedo et al., 2016). In another study about mobile apps, not exclusively in the health sector, researchers tracked individual apps and their presence in the top 300 charts in the app store and used a generalized hierarchical modeling approach to measure sales performance and found that attributes such as, free offering, high initial chart rank, continuous feature updates and high user review scores have positive impacts on apps’ sustainability (Lee & Raghu, 2014). Moreover, according to Krishnan & Selvam (2019) high user ratings, frequent updating, long standing market presence, and those offered by US companies are among the factors that boost app downloads (Krishnan & Selvam, 2019). Even though literature indicates that app demand is affected by the price, it has been stated that the option of in-app purchases and in-app ads can affect a user’s decision to download the app and get access to additional features (Ghose & Han 2014). There has also been indicated that app demand is related to the file size, app’s age and the developer’s textual and visual description of an app. These factors can positively contribute to the willingness of users to download an app (Carare, 2012). Sadegh et al 2018 reviewed 76 studies in order to define an evaluation framework for mHealth applications considering the mHealth service evaluation process from analysis stage to implementation stage with regard to different mHealth stakeholder's points of view (Sadegh et al. 2018). Based on the literature, consumer usability evaluation is affected not only by contextual factors, but also by consumers’

94 beliefs about technology, technology readiness and by different belief structures (Massey et al., 2007) and while it has been stated that there is a correlation between the star rating of an app and its number of downloads, there are serious concerns for the lack of quality of apps across all fields of healthcare (Wisniewski et al., 2019). Research has shown that one of the most important challenges when using digital information technology and apps is related to privacy, security and accuracy issues of data management (Galetsi et al. 2019; Walsh et al. 2019; Yaacoub et al. 2020). Thus, ethical debates are raised for developers, regulators, and managers interested in mobile development (Shilton & Greene 2017). For example, self-diagnosis smartphone apps are criticized for their implications to privacy of patients’ personal data and the possible effects on the doctor–patient relationship and medical authority in relation to diagnosis (Lupton & Jutel 2015). The goal in this ethical context is to provide further suggestions for ethical data management, apart from legal necessity, and follow societal tensions regarding the ethics of data-driven products and services that existing norms still fail to intercept (Loi et al. 2019). However, there is still a lack of proper understanding of the required computational and analytical framework tools for these technological developments (Istepanian & Al-Anzi 2018). There have been some efforts in the literature on understanding whether ethical considerations are among the selection criteria of a user when downloading a health app considering the variety in app stores. A recent study used app features to identify 36 criteria that are important to the design, development and analysis of mHealth-related apps. Amongst them, privacy stood out as the compliance with the law and treatment of users’ data, security as the data protection, authorization mechanisms and detection of vulnerability (Llorens-Vernet & Miro 2020). Another study reviewed general app features, in order to investigate the comprehensiveness and quality of information, the study indicated that less than a quarter of the investigated mHealth apps provided a privacy policy and many lacked source citation to help users evaluate their quality (Nicholas et al. 2015). An overview of the literature that discussed the methodologies introduced so far for the identification, characterization and assessment of health apps by doctors outlined that the most significant components of app trustworthiness are privacy of data, reliable app developer, scientific verification, clinical validated information and ease of use (usability) (Paglialonga et al. 2018).

The World Health Organization ran a global survey in 2009 about the adoption of mHealth. The report refers to mHealth apps that support mobile telemedicine for consultation between healthcare professionals and patients; patient monitoring through installed sensors in households; and patient records repositories, appointment reminders and medication reminders to improve treatment compliance and support systems for diagnostic decisions, all of which are relevant to medical professionals (Kay et al. 2011). However, the survey concludes that the level of adoption of these mHealth categories is quite moderate and the lowest adoption is identified in diagnosis decision support systems. A review of health apps assessed by doctors identified that the most important app capability was ease of use (usability) (Paglialonga et al. 2018). It has also been proven that the willingness to adopt mHealth could be reduced by mistrust or beliefs about risk, which have an interconnected influence on adopters (Fox & Connolly 2018). However, the recent literature discusses the opportunities and challenges of AI implementation in healthcare (Goldhahn et al. 2018; Rajkomar et al. 2019; Davenport & Kalakota 2019; Siuly & Zhang 2020), a review paper discusses the AI applications in mHealth (Alotaibi 2020) and there are papers focusing on a single medical specialty that discuss mHealth apps with AI, e.g. (Zakhem et al. 2018; Gamble, 2020). Unfortunately, no studies have been identified in the literature, which analyze the features and challenges of mHealth apps for medical professionals that utilize artificial intelligence technologies. Therefore, another research gap has been identified.

CHAPTER B. 3

3. Methodology

This study aims to describe and catalogue available mHealth apps for medical professionals and/or medical students. This analysis was focused on Android apps, because smartphones using the Android operating system are currently holding the largest market share in the United States and worldwide (Hoeppner et al. 2016).

Therefore, only apps available in Google Play were considered, excluding those that are designed only for iPhones, following the approach of other relevant studies (Jutel & Lupton 2015; Martínez-Pérez et al. 2014; Sedrati et al. 2016). To obtain a sample of such apps, a review has been followed of all smartphone apps in “Play Store” listed under a wide range of search terms. Initially, the search included the terms of “diagnosis” and “prognosis”. Following selection criteria, this search provided us with a sample of 139 related health apps. Next, the search continued with the terms “medical”, “clinical”, “diseases”, “symptoms”, “doctors” and physicians”. After successive searches, it was realized that the returned results repeated the majority of health apps, apart from 29 new apps that have been included in the pool. Each app underwent an initial screening based on the descriptions and associated screenshot images provided by the store. Inclusion of an app in this evaluation required meeting the following criteria: (1) was intended for diagnosis and prognosis, (2) was addressed only to medical professionals or medical students, (3) was not requiring subscription to another program to operate and (5) was in English. After a day of using each app, another screening against the inclusion criteria was undertaken, and those failed to meet the criteria, were excluded. The apps’ selection took about 3 months, from December 2019 to February 2020. After scrutinizing all apps, data from the selected apps were added in an excel file. This process included tasks such as reading in-store descriptions, user reviews and collecting relevant information after downloading the app on an Android smartphone and using it. Figure 1 presents the methodology framework.

Figure 1: Methodology process

A specific taxonomy has been followed concerning the formulation of categories and the grouping of the apps in the sample accordingly. The target was to categorize apps into different types based on their main purpose. First, it has been used deductive reasoning for the content analysis of the apps using the descriptions provided by the 168 apps. The deductive qualitative content analysis was based on the descriptions of the apps and the app types were continually refined and finally identified in full. It has also been used the inductive method by grouping the apps into the identified categories (Ameel et al 2020; Backman et al 2020). A similar process was also followed for identifying all app features relevant to privacy-security and all features relevant to reliability. Therefore, each app was reviewed for its characteristics based on the information provided by its developer and its user comments (Hoeppner et al. 2016; Stoyanov et al. 2015). The collected descriptive information with their technical aspects about security/privacy and reliability dimensions (e.g. log-in, password- protection, brand recognition) is presented in the results section and in the related tables. The taxonomy has been used to identify the features of the mHealth apps and the types of apps for medical professional and then to provide valuable insights into how the taxonomy is relevant to discuss which app types include more certain features and why. This process has been repeated for apps that utilize artificial intelligence along with discussing what medical and operational outcomes attempt to achieve. Triggered by the main results of this research the aim of this study is to compile an agenda for potential useful future apps for medical professionals and for this purpose there has been a second research in the recent relevant literature. This secondary

98 research has been conducted to fill the gap that has been created since the research in the App Store revealed only a small sample of AI mHealth applications for professionals but very useful indeed. The second search has been conducted in the Web of Science database by using the key terms artificial intelligence (AI), machine learning (ML) and health/medicine and focused on important articles from journals such as Nature, Health Affairs, JAMA, etc. Then a quick visit has been followed in a number of online sites of mobile app developers using AI and ML, such as FRITZ AI and the APP Solutions, and read relevant reports on how AI and ML can be used on mobile apps. Therefore, a future agenda for such apps further investigates the road to innovation and clear guidelines are provided for managing the ethical challenges that emerged from the main part of the research. Finally, for the empirical analysis of the research the SPSS statistical software has been used to provide the results for the empirical analysis.

CHAPTER B. 4

4. Conceptual frameworks and results

4. 1a. 1st Conceptual framework. “Communication Privacy Management”

The first study is based on the “Communication Privacy Management” (CPM) theory to explain the way app developers and health professionals behave towards data privacy issues related to a health app and expanded with the inclusion of the information reliability perspective. The CPM theory recognizes that individuals (owners) believe that they own their private information and have the right, on one hand, to share it with other individuals or entities (co-owners) but, on the other hand, need to control it using ‘privacy boundaries’ that define whether their information is protected (Petronio & Child 2020). Users consider that they can decide with whom to share their own private information and their judgment to entrust it to a co-owner involves the risk-beneﬁt dynamic, which is governed by the privacy rules, to avoid

99 boundary turbulence (when privacy rules are not followed by co-owners) (Ngwenya et al. 2016).Hence, this theory can contribute to the investigation of the effects of privacy policy on users’ perception of security in the context of information sensitivity of mobile apps (Zimmer et al. 2020) and provides a way to understand how security of private information occurs in different contexts like using mHealth apps during practicing diagnosis. In the conceptual framework the CPM theory has been expanded considering that, especially for mHealth apps, the reliability status of those with whom the private information is shared (co-owners) makes a difference. Therefore, questions like how reliable is the app developer (brand recognition) and whether the app information derives from credible sources are issues that should affect the selection of these apps and the sharing of private data (McNiel & McArthur 2016; Nouri et al. 2018). Moreover, the reliability aspect adds a trust element on users’ perception which has to do with the app accurately performing the tasks that promises in the app description (Boudreaux et al. 2014). Based on their tasks, which define the app’s purpose and usefulness, the mHealth apps for professionals are categorized into different types. Therefore, the features of privacy and reliability may be assessed for all mHealth apps together but also per their type. A special focus is given in the “smart” app types that use artificial intelligence as there is an evident growing demand for these apps (Wang et al. 2018). In the conceptual framework of Fig. 2 this process has been explain on the basis of the expanded CPM. The first stage of the conceptual schema involves the health professionals who are the information owners (of their patients’ data and their own data) and the decision-makers for considering using an app based on the app usefulness (defined by the type of app and the use of AI technology) created by the app developer (data co-owner). In other words, health professionals download apps to fulfill a variety of needs and use the app tasks for different diagnosis purposes. This need leads to the induction of personal data (their own or their patients’) in the mobile apps that decide to download and use, creating co-owners of the information that in this case are the app developers. The second stage of the process explains the criteria (boundaries) used by the information owner to perform judgments about the app trustworthiness, which is determined by the app features that inspire privacy and reliability. There has been an assessment of the level of the two app trustworthiness tiers that app developers have added to avoid boundary turbulence of medical professionals when selecting the mHealth app. Therefore, this stage of the conceptual framework, describes the added

100 app features by app developers to abide with the user boundaries and remove turbulence (non-selection of the app). During this procedure the health professionals (information owners) decide whether or not an application can be perceived as trustworthy for adoption in the medical practice. The certain app features that have been analyzed are the features related to privacy boundaries and the features related to reliability boundaries. Through the investigation of the health apps addressed to health professionals, there has been an attempt to explore the ethical debate between innovation, privacy and trust and to identify the most innovative of the mHealth apps for professionals by emphasizing the ones that use artificial intelligence methods and the expectation for more such apps with new capabilities.

Figure 2: Conceptual Framework of “Communication Management Theory”

101

4.1b. Results

The results of the study follow the basic elements of the conceptual model. In the first part of the results section, there is a description of the added app features by app developers with regards to privacy and reliability concerns to avoid user turbulence (non-selection of app by users) and the different types of health apps currently available for professionals that define their usefulness. Indicative examples of apps per type, features, and comments from user reviews are provided for better comprehension. In the second part, there has been an identification of the apps from the sample that include artificial intelligence (AI) technology and an analysis of them as per their type and features of privacy and reliability. The third part of the results presents an agenda for future useful mHealth apps for professionals and describes mitigation actions for overcoming privacy and reliability issues concerning the current and future use of these apps.

4.1c. App features and related challenges of digital healthcare

The sample in the current study was analyzed based on the information provided in the App Store. Most of the apps had more than 10,000 downloads which means that a good number of professionals were at least interested in finding a helpful app to assist them in the diagnosis or learning process. In the following tables there is a presentation of the accumulated descriptive characteristics of the sample of 168 mHealth apps targeted to medical professionals. There has been also an identification of the privacy and reliability feature categories available in the apps, which developers have incorporated to avoid turbulence in user judgment about the accepted boundaries of a trustworthy app based on the conceptual framework. Table 1 lists the sum of the app features based both on the description provided in the App Store and by finding its core functions after installing and opening the app. The mapping and measuring of these characteristics provide a first picture of the current attributes that are incorporated into the professional health apps. The description that follows provides an understanding for these features supported also by examples.

Table 1: Feature relevance

102

Features (F) N % a.PRIVACY PrivacyDATNAGEMMAAMANAGEMENT policy 147 87.50 Authorization 66 39.28 Email request 48 28.57 Password protection 44 26.19 Registration 42 25.00 b. RELIABILITY Brand recognition 121 72.02 Bug Fixes 80 47.61 Credible source 76 45.23 Feedback 53 31.54 Help function 52 30.95

a. Privacy is not simple to be analyzed in terms of technical characteristics because many factors affect privacy and some of them concern auditing of the apps while others concern the auditing of the apps' providers (Benjumea, 2019).The declaration of the “Privacy Policy” discloses information about the developer and the terms and conditions of using the app, as the ‘compliance with the law and treatment of users’ (Llorens-Vernet & Miro, 2020). In this paper, the term “Privacy Policy” is considered a feature. The focus has been oriented on the declaration that describes the Privacy Policy (known as ‘data subject’ in the GDPR) and it has been reassured if its internal procedures comply with official regulations (e.g. GDPR). In the case of privacy concerns it has been observed that out of 168 health apps, 147 (87.5%) had a declaration of Privacy Policy and a full description of the policies that the company follows for the interpretation of stakeholders’ data. For example, the Privacy Policy declaration of “Skyscape Medpresso, Inc.” which is the company designer of many health applications, such as Ferri's Clinical Advisor, describes the firm's practices towards users’ personal information. They inform their clients about their policies on: children’s information, users outside of the United States, the personal and non- personal (anonymous) information collected, and the way the company uses the information. Typically, this information tracks users patterns throughout the company’s website, and it may include information such as browser type, operating system, date/time stamp of their visit, IP address, domain name, referring URLs, statistics about the number of visitors to the site, the number of

103 pages visited, the user response rates, etc. The company also informs the app users that it may use or disclose personal information if it detects actions that damage the website or will prevent illegal activities. Lastly it states the security measures (e.g. use of firewalls, secure connections on website, the use of Secured Socket Layers) it takes to help protect and safeguard personal information from accidental loss, misuse, unauthorized access, disclosure, and alteration or accidental destruction. Most importantly, in the declaration form they state the possibility to make available users’ personal information available to third parties for marketing purposes and provide their contact details in case of user disagreement. On the other hand, the auditing of the mobile app provider is related to user’s consent to the Privacy Policy declaration. The user must accept the terms of the privacy statement so that the provider is permitted to manage the requested data, for which access has been requested or to give out data to third parties. There has been also a classification of the apps in the “Authorization” category when users’ acceptance of the company’s terms and policies was necessary for allowing access to their personal data and exploitation based on company’s policies. Only after consent could the app be downloaded. It is noteworthy that only 66 (39.28%) of the apps required user consent of terms and conditions (authorization) before downloading the app. The authorization process includes acceptance of terms for data use for marketing purposes or access to photos and files of the device, which for some apps, like Miniris, constitute the functioning condition of the app. Privacy and security concerns have arisen due to various network attacks. Lack of strong security authentication measures, that many health apps may suffer from, can easily allow attackers to infect devices or bypass personal data safeguards. From the security and protection of cyber-attacks perspective, when this happens to an application used for work activities, enterprise data or credentials are put at risk, along with personal information (Filkins 2016). It has been proven that consumers can take costly actions to protect and anonymize their identities (Valletti and Wu 2020). Sensitive data issued to verify a user’s identity before allowing an individual to login with their name and password in order to fully use the app. Authentication features such as semi registration (just email), full registration, and password protection application are used in this study to describe security attempts of both designers (identify cyber threats) and users

104

(need verification to work with the app). The “Email request” feature was available in 48 apps (28.57%) and requested user’s email before app use, but no other personal information was necessary. The feature “Password-Protection” (44 apps, 26.19%) refers to security and privacy provision with the use of a code for entry to the app after sign-in with email or full registration. “Registration” is the feature that requested users’ personal data in order to become functional. Examples of personal data included name, age, profession, gender, nationality, education, etc. In the sample of 168 apps, 42 apps (25%) requested registration (asking for a full range of personal data) in order to navigate through the app and almost 28% required email (including those that needed registration). b. Reliability is complex as it encompasses diverse elements like adherence to clinical guidelines and recommendations, scientific verification, and clinical validation (to assess apps based on the reliability of scientific sources, as available on literature databases) or the credentials of the developer and the medical professional involvement in the app development (Paglialonga et al. 2018). Reliability also demands transparency, meaning the framework the designer uses in order to describe to stakeholders the mechanisms of the app to process their data and to arrive to specific decisions. Transparency is the right of health professionals to know and understand the aspects of a dataset or an input that could influence their clinical decision-making (Cutillo et al. 2020) avoiding biased and inaccurate results and ensuring the quality and reliability of data. Based on the apps’ specific features there is a description of both the credibility and reliability of information that was provided. The indication of the “Brand recognition” as a feature is relevant to the factors that inﬂuence the trustworthiness of mHealth apps when the developers on their own provide links to end users about the apps’ “manufacturing brand”. In the case that the developer is a well-known app manufacturer or medical organization, users can be engaged by the “Brand familiarity” because health apps from well-known brands/organizations are generally perceived as more trustworthy compared to those of unknown brands (van Haasteren et al. 2019).“Bug Fixes” is counted here as an element of trust, referring to the activity of correcting software defects in order to enhance the provided service and is included in almost 47.61% of the health apps of the sample. The feature describes designers’ technical interventions on app functions that did not work well and were noticed

105

after users’ indication (but not fixed in real time) or by designers’ effort to improve the app by releasing new versions. Under the feature “Credible source,” are classified the apps which ensured the user that their content and provided information have been derived from credible organizations such as FDA (U.S Food and Drug Administration) and EMA (European Medicines Agency) alerts, JAMA Rational Clinical Exam Series, Elsevier, etc. guaranteeing information accuracy. In the sample almost 45.23% (76 health apps) are developed and guided from national institutions and universities or base their content or advice on international peer-reviewed literature (JACC Journals, Lancet etc). The feature “Credible source” is indicative for the evidence base of the information provided from the app. To that extent, and since the nature of the app is basically scientific and medical, are considered the apps with the above characteristics as being more trusted. For example, there are health apps such as the “Diabetes Diagnostics” that is developed by the Diabetes Clinical Research team at the University of Exeter Medical School, University of Exeter, UK, or the “ACC Guideline Clinical App” that is developed by the American College of Cardiology Foundation and gives access to its official guidelines. In the “Feedback” feature, are allocated the apps (31.54%) that give users the possibility to communicate with the app’ developers or suggested health professionals for queries about medical issues prompted from the app. The “Feedback” feature indicates the trust relationship between the user and the app as the user can benefit from the app’s ability to provide relevant feedback by analyzing user data. Under the category “Help function” (52 apps) are being allocated the apps that had a help function tab, including a list with definitions and guidance, and contact details of the designer/company. The allocation of the apps to the 10 identified feature categories revealed that usually apps include only a range of these features. For example, the most common features are declaration of “Privacy Policy” 87.5%, “Brand Recognition” 72.02%, and “Bug Fixes” 47.61%. On the other hand, important features such as “Password- protection” were included only in 26.19% of the apps.

4.1d. Health app types

106

Health apps were also classified based on their type. After screening the content of all selected apps, 9 main categories have been formulated. Table 2 provides this allocation according to the app tasks and purpose it fulfills. Additionally, table 3 presents statistics about the frequencies of features per health type of apps.

Table 2: Health apps types

App Type Description N % a. Handbook/manua Reading assistant tool to increase comprehension 47 28.0 l and memorization of information provided in medical textbooks and study material, designed for educational and training purposes b. Differential Designed to support doctors during the diagnosis 35 20.8 diagnosis process by typing patient’s clinical and lab data in assisting tool order to receive disease matching. c. Clinical Provide evidence-based clinical practice guidelines guidelines, on how to diagnose, investigate and treat a condition dictionaries and or how to quickly examine it. protocols 32 19.0 d. Calculator for Calculate health related scores after the entry of doses, scales patient’s clinical/lab results to help in diagnosis 12 7.1 e. Games - Provide gamified patient cases for practicing simulators diagnosis in a variety of medical procedures. 11 6.5 f. Tool for Include a template or software for patient managing patients management over time by retaining patient records, records health history, visits-appointments, etc. 10 5.6 Multiple choice tests and other type of questions with detailed explanation of correct answers g. Quiz covering all areas of medicine, preparing professionals for exams and for exercising knowledge 9 5.4 h. Scientific Include customized health news and articles news/libraries published by credible experts 7 4.2 Provide communication tools among medical i. Connection with professionals for discussing patient cases and new doctors knowledge about treatments, etc. 5 3.0 Total 168 100 a. “Handbook/manual” is the category that gathers the largest proportion of the selected apps (28%) and includes the apps that are mainly useful to medical students as a reading assistant tool to increase comprehension and memorization of the information provided in their textbooks and study material. They are designed for educational and training purposes and they are popular for helping students reduce the time needed to study to pass exams. The link to the brand webpage has been provided in 61.7% indicating that these apps were mostly developed for marketing purposes. This type of app, most of the times (90%), includes a Privacy Policy, but because they only provide information without any request of data to interpret there is no need for a feature to ensure security.

107

From the 47 apps, 21 use the Authorization feature to receive the user’s agreement for accessing phone’s photos and videos or the unique identifier of their device in order to provide personalized advertisements. There are also 10 apps that request registration in order to login. Most reviewers find these applications very helpful for learning purposes, but some of them wish to see more detailed descriptions of the studied elements e.g. body organs. b. “Differential diagnosis assisting tool” (20.8% of the apps) is the category that comes next in popularity and includes the apps that are designed to support doctors during the diagnosis process by typing patient’s clinical and lab data in order to receive disease matching. It improves health professionals’ differential diagnosis skills and it is useful in daily practice for re-assuring physicians’ choices. Among its applications are some (3) that use artificial intelligence (AI) technology, indicating from one side the need from doctors to have customized results during their clinical practice and from the other side the tension for personalized treatment. Technological advancements are probably the reason for the high score in “Bug Fixes” 60%. The users find the app informative and advice giving, but the recent comments ask for app enrichment with MRI images for diagnosis based on image matching, indicating a new potential to these apps. In this application there is interpretation of sensitive patient data, but there is not a noticeable difference in the scores with other apps (Table 3) in the two tiers (privacy, reliability). c. “Clinical guidelines, dictionaries and protocols” category is next in this classification and holds around 19% of the sample. These apps provide evidence-based clinical practice guidelines following a rigorous development process and guide on how to diagnose, investigate and treat a condition and how to quickly examine it. They also include disease dictionaries with conditions, symptoms, treatment drugs, treatment protocols, medical terminologies, and dosage guides. Most of the apps in this type (75%) had a credible source for the information provided based on published books (Elsevier, Wiley etc) or reliable organizations (American College of Physicians etc). Based on the privacy policy declarations, the information collected included exclusively mobile device information (device name, version, language) and downloadable analysis data by Google Firebase SDK and not sensitive information from users, like contact list, emails, etc. They declared that the collected data is only used for app improvement and monetization. Users’ comments brightly discuss the detailed explanations of diseases and the updated material, which is apparently what is important in this type of app; the quantity and quality of the content.

108 d. “Calculator for doses and scales” (12 apps) category contains apps where medical professionals enter their patients’ clinical or lab results and related scores are indicated that could help the clinical practice. Traditionally, medical calculators are used as clinical decision-making tools (Chen et al. 2019). However, these apps differ in nature from the “Differential diagnosis assisting tool” apps. Their aim is to calculate clinical scores and indices, such as coronary heart disease risk, body mass index (BMI), pregnancy due date, individual drug dosing, etc. (Mosa 2012). This category includes diagnostic criteria, a scoring system for various diseases or provides prognostics for certain diseases based on human measurements. As it is observed, almost all of the apps include a declaration of a Privacy Policy, while only 2 apps ask for users’ agreement to give access to photos and archives and another two ask for a registration and password. The role of these apps is to provide the right result using specific algorithms for each calculation. For this reason, it is very important that the apps provide credible sources for their algorithm justifications. From the findings,6 of the apps provide information about the source of their algorithms (peer reviewed literature, American College of Cardiology, etc.) and the rest are based on widely accepted algorithms, such as the “Glasgow Coma Scale” or the “Body mass index”. It is noticeable that 75% of the app’s present advertisements. Most users found these apps practical and timesaving during the diagnosis process, although some complained about slow response times and inadequate formulas. Therefore, this kind of app must be fast, reliable, and accurate to increase downloads and high ratings. e. “Games/simulators” (11 apps) category includes all the apps that provide gamified patient cases for practicing diagnosis and better training, designed to teach specific techniques for a variety of medical procedures. Email was requested in 7 of the apps along with password protection and full registration in 6. All apps included a privacy policy, but authorization to the archives and photos of the device was requested in one app only. According to the reviews, these apps are good for strengthening clinical knowledge through playfully learning when doctors or students are tired from reading. However, it can also fulfill the need of professionals to gain more experience without spending money and time. Such an app is mostly useful when it covers a variety of medical cases with detailed explanations on the answering part. f. “Tool for managing patient records” (10 apps) is the category that includes a template or software that manages patients over time from first visit, to admission, to discharge and then follow-up and assists professionals to retain patient records and health history

109 as well as other related information. What is useful about this type of app is the management of patients’ information and their medical history, minimizing repeated and unnecessary data entry because it can store all kinds of medical notes and records (like text, audio, video, image, pdf, etc.). The most important characteristic of such apps is the ease of use by doctors with simple screens and with templates that can be customized according to professional’s needs. Users also need flexible and ergonomic solutions that can be accessed from their mobile devices or tablets as well as the ability to be synchronized quickly. They also require space for unlimited patients, appointments, invoices, notes, and immediate technical support. g. “Quiz” (9 apps) is the category that contains apps with targeted questions and detailed explanations covering all areas of clinical lab science, preparing students for exams or helping health professionals to exercise knowledge. All the apps included a declaration of privacy policy but only one requested authorization. According to the comments, many users request a larger question bank that can generate more than one quiz per category and a more attractive layout. h. “Scientific news/literature” (7 apps) is the category of apps that include news and articles in health topics written by experts and published by credible publishers, such as Elsevier or JACC (The Journal of the American College of Cardiology). This kind of app tries to be reliable in terms of information as they all provide their source and developers’ webpage. Their goal is to provide scientific information to users and not to sell or advertise products, thus they have less interest in capturing the personal data of users. According to the users’ reviews, most of the complaints are about technical issues like certificate downloads and requesting regular updates. Doctors are interested in continuous knowledge and skills upgrades and this is what this type of app offers. It is very important that these apps are based on trustful sources of medical information and material. Moreover, professionals ask for user friendly and accurate literature apps. i. “Connection with doctors” (5apps) category offers apps to health professionals and students that provides links to communicate with each other on diagnosis matters or discuss new knowledge about diseases and treatments. These kinds of apps are most useful when providing connection with large communities of professionals, while also achieving high privacy and security safeguards. Additionally, this app type is valuable as it provides the opportunity to learn and grow a professional network. According to reviewers’ point of views, future goals of these applications could include the

110 customization of features based on a user profile, satisfying daily clinical needs, and prognosis requirements. In Table 3, under the privacy tier’s category, it is obvious that the health applications that gather the highest scores are “Connection with Doctors”, which needs a lot of personal user data in order to work properly, “Tool for managing patients” that requests patient data and “Games” that also needs personal data to provide feedback. For the second tier, reliability, high scores again gather the “Connection with Doctors” that need to come from a reliable developer and share reliable material. “Literature” type also scores high because users naturally depend on the validity of the apps’ sources and “Calculators” because these apps use algorithms to measure patient data. It is also interesting that “Handbooks” and “Dictionaries” do not include a high percentage of features relevant to the ethical challenges even though they belong to the most popular type of apps. Overall, from the results it is obvious that privacy is seriously examined by developers when introducing new health apps. Security features are mostly incorporated in apps that require personal data to function. For example, “Games”, “Literature”, “Patient management tools”, and “Connection with doctors” need full or semi registration along with password-protection application in order to navigate the app. Reliability issues that are relevant to “trust”, “transparency”, and “explainability” are also important to these app types as their functions are more complex and users require a deep understanding of the provided information in order for the apps to be useful.

Table 3: Frequencies of features per app type

111

4.1e. mHealth apps using Artificial intelligence

Along with the nine different types of health apps that have been identified, there has been also formulated a separate category of apps that use AI methods. Of the 168 mHealth apps for professionals, 17 used AI methods. There is also a separate catalogue for these apps because due to their technical elements, they are capable of making more complex data matching and calculations and based on queries eventually offer more useful information to medical professionals (Deo 2015).The following mHealth artiﬁcial intelligence includes computerized technologies like natural language processing, which aids in speech recognition, text analysis, translation, and other goals related to expert systems (e.g. fuzzy logic). Based on collections of ‘if-then’ rules that are widely employed for ‘clinical decision support’ purposes, machine learning methods are used to learn from accessed data and with the help of statistics and classification to generate customized results and robotic process automation as if they were a human- user. Following a script or rules that rely on a combination of workﬂow, involving automated planning and scheduling focused on organizing and prioritizing the activities required to improve the efficiency of human procedures (Davenport & Kalakota 2019; Wahl et al. 2018).The examples that follow shed more light. Based on the apps’ content analysis of the dataset, three of the apps that include AI techniques to interpret data belong to the “Diagnosis assisting tools” category, five are “Calculators of doses and scales”, three are “Games/simulators” and six are “Tools for managing patient records”. In Table 5 there is a classification of the 17 mHealth apps based on their type and their medical and operational benefit that is offered to medical professionals. A representative machine learning (ML) tool for “Differential Diagnosis assisting tool” category is “Visual DX”, its purpose is to support physicians and dermatologists as a quick reference tool using cutting edge machine learning at the point of care by uploading an image (e.g. of the patient’s inflamed skin) and getting back a list of a differential diagnosis. The app includes a detailed declaration of a Privacy Policy and requires owners’ information, acceptance, and full registration. The Privacy Policy declaration states that the company is allowed to disclose the users’ information to third parties without prior user consent. VisualDx relies on a worldwide medical

112 editorial board of practicing physician scholars to keep clinical content objective, reliable, and current. In practice, machine learning applications like Visual DX perform automated data analysis by using algorithms that iteratively identify patterns in data and learn from them (Wahl et al. 2018). ML algorithms provide information systems the ability to automatically learn and improve from experience without being explicitly programmed (Kodratoff, 2014). In the case of “Calculators” a representative example is the application “MDCalc Medical Calculator” that could be characterized as a fuzzy logic system, which assists to mimic the logic of human thought because it contains more than two options to provide calculated results by combining the inserted values to offer more than two possible solutions (Guimarães et al. 2018). The app includes a Privacy Policy and full registration but does not include an authorization. MDCalc’ s Editorial Board is comprised of scientists in several medical specialties who hold leading positions in healthcare institutions (e.g. Harvard Medical School in Boston, Massachusetts). This category includes complex calculators that need more than two measures to result in more accurate predictions and follow the process of converting input data into fuzzy components, creating fuzzy sets and the set of rules and inferences (Guimarães et al. 2018). Another artificial intelligence application from the category of “Games/Simulators is “In Simu patient” that intends to practice medical diagnosis from a virtual clinic with simulated patients and a variety of diagnostic methods. Medical teaching simulators may be understood as tools that assist in clinical settings without any potential risk to the patient (Flores 2013). This medical database includes peer- reviewed medical literature (e.g., UpToDate, McGraw Hill, Oxford University Press medical books) and international guidelines. The app developer has a very informative website with details on privacy boundaries and reliability issues. Virtual clinical cases involve robotic process automation that are used for repetitive tasks combined with other technologies like image recognition were users are allowed to follow several tasks according to certain rules. Lastly, a representative artificial intelligence app of the “Tools of managing patients category” is “Doctor Assistant”, which represents an automated planning and scheduling artificial intelligence system that helps health professionals stay organized and improves productivity using a smartphone or tablet while it creates, stores, and retrieves a patient’s health record and schedules the patient’s next appointments on

113 android calendars etc. This app has a Privacy Policy and requires authorization, but not full registration, although it handles doctor’s patients records. The app is focused more on its ease of use rather than setting boundaries to ensure security. It does not refer to a credible source of information because its character is more managerial rather than educational or advisable, but it has been published on the first page of the World Health Organization (WHO) Compendium of Innovative Health Technologies. Table 4 shows the percentages of the features’ inclusion between all 168 apps (as shown in Table 1) and the 17 AI apps. In comparison to the percentages of all apps according to the features, the privacy and reliability tiers of AI are higher. This is explained by the fact that AI apps are designed to accept different kinds of personal data as explained further above and must provide stronger boundaries to ensure selection from data owners (health professionals).

Table 4. Artificial intelligence apps’ Features

PRIVACY Total% AI% RELIABILITYY Total% AI% Privacy policy 28.0 94.11 Brand recognition 72.0 94.11 Authorization 20.8 23.52 Bug Fixes 47.1 47.05 Email request 28.6 52.94 Credible source 39.3 41.17

Password protection 26.2 52.94 Help function 30.9 52.94 Registration 25.0 47.05 Feedback 23.2 23.52

Big Data analytic methods and AI can be proved helpful to organizations for enhancing operational performance (Sivarajah et al. 2015). For example, as shown in Table 5, apps used as “Tools for differential diagnosis” or “Calculators” have implemented algorithms with diagnostic and prognostic capabilities and are used by health professionals as a tool for quickly providing an initial prognosis for certain conditions. The wide use of these kinds of apps with positive outcomes could prompt and further enhance research for the invention of new algorithms for diagnosis and prognosis. Similarly, “Games and Simulators”, that according to reviewers’ comments, are mostly downloaded for learning and practicing reasons could become revolutionary assisting tools in the training of clinicians to perform even complicated diagnosis and optimize treatment decisions and patient’s health outcome predictions. Lastly, there are mHealth apps for managing patients in clinics and care centers, which from the analysis of their electronic health records, could be used as tools for the early prediction of an

114 epidemic with its characteristics (e.g. peak and duration of infection) and therefore could protect the public health from outbreaks of infectious diseases (Noorbakhsh-Sabet et al. 2019). However, taking into consideration that medicine as a professional activity contains operational insights, there is a need to observe these insights under the managerial spectrum. For instance, the “Diagnostic assisting tools” can operationally assist professionals during clinical routine because they can minimize the participation of other specialties for advice and maximize patient engagement for interventions by minimizing diagnosis delays, which usually make patients lose interest when their health problem is not eminent at the time of the doctor’s visit. This can also lead to time savings for patients and medicals and become cost effective for patients and insurance systems. Similarly, the apps that use algorithms as “calculators” can minimize the time of outcome extraction in terms of diagnosis and treatment during clinical trial and enhance better prognosis outcomes. All types of AI solutions have the capacity to provide additional tools, which in many cases led to time reduction for accessing data and generate stronger collaboration between stakeholders (Jones et al. 2017). Leveraging existing remote infrastructures into a common information system (IS) reduces installation, monitoring time and expense and focuses on improving quality. The “Games /simulators” for the prediction of treatment outcomes can lead to knowledge improvement that assists clinicians and clinics to foster operational performance by knowing what comes next as well as to foster innovation by the acquisition of technologically updated equipment for precision medicine. Lastly, artificial intelligence apps used as “Tools for managing patients records” can be proven cost effective because of their predictive ability in managing high numbers of patients, which aids in avoiding congested care centers and frequent clinic visits, leaving space for more people in need to use the health services quicker and a reduction in patient expenditure for health insurance.

Table 5: Artificial intelligence mHealth apps and outcomes Description Health Apps Medical outcome Operational outcome Tools for Visual Dx Disease diagnosis Maximize patient differential Dermion and prognosis engagement diagnosis The Chief Complaint

115

Calculators MDCalc Medical Disease diagnosis Minimize time of Calculator and prognosis clinical trials Calculate by QxMD Medical Calculators Medi Calc® Medical Formulas Games, Human Dx Diagnosis and Knowledge simulators Touch Surgery Treatment improvement, better InSimu Patient optimization and preparation and more Outcome innovation prediction Tool for Medical Records Public health Reduce cost of managing Doctor at work plus managing patients and patients OPD App - For facilities Doctors Doctor Assistant Dentist Manager: patient organizer software List of my patients

4. 2a. Conceptual Framework. “Normalization Process Theory” (NPT)

In IS research and especially in technology adoption research, gaining insight into the factors related to the motivations or goals of smartphone users are key elements to explain important outcomes, such as usage intention, satisfaction and engagement (Jung 2014). In this study it has been used a theory borrowed from sociology the “Normalization Process Theory” (NPT) which is used in the field of technology, and originally in healthcare systems, to explain the adoption of technological and organizational innovations (May 2013). Since, IT-related innovativeness is associated with smartphone adoption and mHealth apps are regarded as an innovative product recently introduced in the market, this theory can explain the formed dynamics through the integration process (Lee & Lee, 2018). Based on the NPT, there is provide a systematic explanation of the dynamics between feature relevance and trustworthiness of mHealth apps and the engagement of

116 medical professionals in using them during clinical practice, which is measured based on the apps’ popularity. In healthcare, the introduction of innovation in an operational routine requires complex organizational processes and involves insights from sociology to offer a more comprehensive explanation of the implementation of new clinical practices (May 2013). NPT explains the process of the introduction (implementation) of new practices in an everyday work routine (embedding) sustaining social contexts (integration) (May & Finch, 2009). More specifically, in the study it has been investigated the usefulness of the health professionals’ apps as innovative practice in medicine and the offered features after experiencing trialability in order to explain their relevance with professional engagement and popularity. Figure 3 explains this process on the basis of NPT. The first stage of NPT involves the “implementation” of the actual innovative intervention (mHealth app). There are four main corresponding components that are followed during the implementation of a new intervention: “coherence” involves understanding the value, importance and the distinctiveness of the actual intervention or else “identification and sense-making of the need”, “cognitive participation” involves the effort of key participants to organize and sustain the new intervention, “collective action” impacts relations between groups of professionals and fits with the overall organizational context including goals, morale, leadership and resources, and “reflexive monitoring” involves the appraisal of the new intervention, determining how effective and useful it is for participants and for others (patients) (Kosse et al, 2020).The mechanisms that occur during the entire process of innovation intervention of mHealth apps are presented in Table 6 to explain the theory’s particular mechanisms and components through empirical contexts.

Table 6: The Mechanisms of NPT of mHealth in Medicine Constructs of NPT Definition (May&Minch, Mechanisms of Implementation (Kosse, 2020) 2009) of mHealth in Medicine Coherence sense Participants’ understanding of Health Categories making the intervention Differentiation Cognitive Participation Participants' commitment to Users' Reviews & Quality effort work with the intervention Evaluation

117

Collective Action Fit with overall context Trialability &FeatureRelevance commitment including goals, morale, leadership and resources Reflective Monitoringa Participants’ evaluations and Users' Engagement ppraisal appraisals of the intervention. &Popularity

The four contracts/components in the case of mHealth adaptation in medicine are mostly explained through the differentiation of mHealth app categories and the needs they are called to fulfill. Based on their capabilities, certain categories of apps are created (the need of intervention). During the process of adapting an innovation, users have the opportunity to validate expectations on how it can fulfill the expected or promising needs experiencing trialability to test the intervention before full adoption. The second stage of the process: “embedding”, explains the mechanisms that are used to perform these new practices, which in this case are the features of the apps (how the need is met).“Embeddness” is comparable with the concept of social capital and includes relational attributes like trust and norms, tiers that are crucial for businesses such as healthcare that implement social innovations, which often require a deep understanding of complex social issues (Lashitew et al. 2020). During the embedding stage and experiencing trialability, users are able to identify three social tiers of trustworthiness (privacy, reliability and usability) that could lead to a greater engagement of innovative practices from professionals or organizations driving them to the adoption of the innovation and the formulation of new policies. In the last stage, the “integration” stage of the process, describes how the new intervention in clinical practice is sustained (Kosse et al. 2020) based on the quality of the intervention and its popularity which ensures its continuation. Based on users’ reviews and researchers’ investigation of the apps, the Mobile App Rating Scale (MARS), a validated tool (Salazar et al. 2018) has been used to measure the quality of each app. Following the MARS methodology, each app was measured against a 5-point Likert scale on 23 individual MARS items clustered under the engagement, functionality, aesthetics and information quality categories. The MARS is targeted to researchers, mobile app developers and mHealth experts to assess technical information and capabilities of mobile apps (e.g., customization, gamification, ease of use). Its rating is based on scientific evidence of the mobile app during clinical trials to determine acceptability, usability, satisfaction, and targeted outcomes (Dawson et al. 2020).

118

The path through the integration process explains the engagement of the innovative practices from professionals or organizations, measured by app popularity (users’ evaluations-ratings and number of downloads). It also reveals that the higher the app quality, the higher the engagement and the popularity of an innovative interaction. Figure 3: Conceptual framework

4.2b. Results

A. Implementation stage: Health apps’ categories

After content analysis of the mHealth apps for professionals, there has been a classification of the apps into three distinct categories based on their purpose and usefulness and identified 9 main categories. Table 7 presents this allocation together with descriptive measures from each category, which are discussed later in sections BandC. The third column includes an indicative app from each category, described in the text in more detail. The selection of the particular app was based on the combination of a high number of downloads and users’ ratings.

Table 7: Health apps’ categories Indicative NFeature MARS Stars Ndownloads App description N % app i. Education mean sd mean sd mean sd mean sd &Training 1. Handbook/manual 47 28.0 Internal 3.1 1.3 2.9 0.6 3.4 1.8 161952 7.4 Organs in 3D Anatomy 2. Guid/es-dict-prot. 32 19.0 Diseases 3.2 1.4 3.1 0.6 4.1 0.9 130346 2.9

119

Treatments Dictionary 3. Games/simulators 11 6.6 Prognosis: 3.4 1.7 3.6 0.6 3.9 1.7 190954 3.3 Your Diagnosis 4. Quiz 9 5.3 BMJ On 2.8 1.3 3.3 0.6 4.3 0.3 32777 3.2 Examinatio n 5. Scientific news/lib 7 4.2 Medscape 3.9 0.4 3.3 0.5 4.4 0.4 41428 4.2 CME Totals & means 106 63.1 3.3 1.2 3.2 0.6 4.0 1.0 111491 4.2 ii Decision-making on diagnosis 6. Differ.diagnosis 35 20.8 Clinical 3.0 1.5 3.1 0.6 3.8 1.4 44917 8.8 Treatment 7. Calculators 12 7.1 QxMD 3.3 1.9 3.6 0.6 4.5 0.3 143500 2.7 8. Connect with docs 5 3.0 Docquity 5.0 1.5 3.6 1.5 4.9 0.5 340200 4.2 Totals & means 52 30.9 4.0 1.5 3.4 2.7 4.2 0.7 176206 5.2 iii. Management 9. Patient Management 10 6.0 Doctor at 3.2 1.1 3.9 0.5 4.3 0.3 30000 3.9 Work Totals 168 100 Mean 3.2 1.5 3.2 0.7 4.0 1.4 117640 4.6 Totals

i. The categorization of the apps based on their usefulness identified that applications designed for educational purposes is the majority of existing apps. These tools help students reduce their time of study and pass exams. Four different app categories intend to fulfill the need of using new technology methods during education in medicine. First the “Handbook/manual” category gathers the biggest proportion of the selected apps (28%) and includes the apps that are mainly useful to medical students as a reading assistant to increase comprehension and memorization of the information provided in their textbooks and study materials. A representative example is the “Internal Organs in 3D (Anatomy)” app, a practical tool that provides anatomical information using 3D pictures with motion to help students understand the anatomy of a human person. Next in the classification is the “Clinical guidelines-dictionaries- protocols” category which holds around 19% of the sample. These apps provide evidence-based clinical practice guidelines on how to diagnose, investigate and treat a condition, as well as how to examine patients. They also include disease dictionaries with conditions, symptoms, treatment drugs, treatment protocols, medical terminologies and dosage guides. A very popular application is the “Diseases Treatments Dictionary” that explains all treatments of diseases providing data for causes, symptoms, prevention, drugs, prescriptions, medical terms, etc. The most common diseases are listed and analyzed under categories.

120

The app has a search feature where the professional can type the disease of interest and relevant findings appear. It is particularly useful to professionals in emergency departments and those giving First Aid. Third is the “Games/simulators” category (11 apps) that includes all the apps that provide gamified patient cases for practicing diagnosis and better training, which are designed to teach specific techniques in a variety of medical procedures. An indicative app is “Prognosis: Your Diagnosis” that offers more than 400 cases, based on actual clinical experiences, throughout a wide range of medical specialties which can be played within minutes. Each case is accompanied by a comprehensive discussion of the diagnostic reasoning and the key learning points. Fourth is the “Quiz” app category that contains apps (9) with targeted questions and detailed explanations covering all areas of clinical lab science, preparing students for exams or helping health professionals exercise knowledge.“BMJ On Examination Exam Revision-Free Questions” is an app that offers mock examinations to help students practice for their real exams. Last, the category “Scientific news/libraries” (7 apps) includes news and articles in health topics written by experts and published by publishers such as Elsevier or JACC (The Journal of the American College of Cardiology). A representative example is “Medscape CME & Education” which offers online medical education with news and thousands of articles’ references which helps clinicians receive or find the latest information. The app offers a customized experience by letting users choose their content by specialty and topic of interest or format e.g. news, videos, expert perspectives and then creates and updates activity lists based on these choices. Accurate material and user-friendly layout are two important features for these apps. The educational category of apps is the biggest of the three in the sample, as it has been a wide expansion in the development and use of portable technologies, such as smartphones, to support learning (García et al. 2019). ii. The next most popular group is “Decision making on diagnosis” which includes mHealth applications operating as an assisting tool for doctors during diagnosis. The category “Differential diagnosis” (20.8% of the apps) is the most popular in this group. It improves medical practitioners’ differential diagnosis skills and is useful in daily practice for re-assuring physicians’ choices. An indicative such app is “VisualDx” that offers customized differential diagnosis through the matching of thousands of images and diagnosis reports by searching symptoms, signs and patient factors. The second app category of this group is “Calculators” for doses and scales with12 apps in which medical professionals enter specific patients’ clinical or lab results and related scores

121 are indicated that could help clinical practices, such as coronary heart disease risk, patient-specific drug dosing, etc. A representative app is “Calculate by QxMD”. Based on what the clinician wants to calculate, certain questions about patient measurements are asked and, after entering the answers, a result is provided. It contains more than 300 calculators. For example, it includes a predictive model for Emergency Heart Failure Mortality Risk Grade (EHMRG), which is calculated after answering 10 questions about certain physiological measurements and other patient conditions. Its aim is to convert the recent research publications into a hands-on tool in order to provide, for example, the risk of stroke in a trial fibrillation based on the latest research. Last, is the “Connection with doctors” category (5apps) that offers health professionals and students links to communicate with each other on diagnosis matters or discuss new information about diseases and innovative treatments.“Docquity-The Doctors' Network” is such an app that offers a network which connects over 100,000 verified doctors with each other to discuss real world medical cases. The specific platform is an Asian medical education and knowledge sharing platform, exclusively for doctors. What is valuable about this app category is the opportunity to learn and grow through a professional network. iii. The last group includes only one app category called “Patient management” (10 apps) for tracking patients over time in doctors’ visits, admissions-discharges, etc. It assists professionals to retain patient records and health history and other related information. It minimizes repeated and unnecessary data entry because it can store all kinds of medical notes and records like text, audio, video, image, pdf, etc. “Doctor at Work (Plus)” is a patient electronic medical record, patient appointment tracker, biller, Rx printer, and sales and income report generator. The app assists doctors to document the history of examination, diagnosis, and treatment of a patient and manages patient appointments. It is also a software that can assist doctors with their billing while maintaining and ensuring the confidentiality of patients’ records. Ease of use, customized templates based on professionals’ needs and space for unlimited patients are important characteristics of such apps as well as simultaneous access to the app’s content from a device with real-time synchronization and technical support. Most of the identified apps (66.6%) cover a wide range of medical conditions as the majority of apps belong to the group of “Education & Training” offering a broad spectrum of knowledge in medicine. The remaining apps are specific to a medical specialty or group of health provisions or health professionals, with 21 identified

122

categories. 9 were specific to cardiology, 8 about emergency care, 5 about nurses, 4 were targeted to dermatologists, another 4 to ophthalmologists, etc.

B. Embedding stage: App features analysis indicating app trustworthiness

To understand the use and safety of professional health apps in everyday clinical routine, certain feature categories has been formulated, in which the apps have been classified and assigned these features to the elements of the integration process (privacy, reliability and usability) of the “Normalization Process Theory”. Table 8 lists the sum of the apps’ features based on the description provided at the App store and after installing and opening the app to reach its core functions. These features are classified in relation to the challenges arising from the introduction of innovative mechanisms in clinical decision processing explaining their trustworthiness.

Table 8: mHealthFeatures’ Analysis Features (F) N % Tier 1. SECURITY 1. Privacy policy 147 87.50 2. Authorization 66 39.28 Tier 2. RELIABILITY 3. Credible source 76 45.23 4. Feedback Contact 53 31.54 Tier 3. USABILITY 5. Guidance 101 60.11 6. Social sharing 68 40.47 7. Tutorial 30 17.85 (Reverse Usability) a. Ads inclusion 82 48.11 b.In-app purchase 65 38.70

Tier 1.Trust is a pressing topic and has been found to be connected with privacy concerns and expectations of users (Martin et al. 2019). Privacy concern is a theoretical control variable that influences the consumer intentions of mobile app use and refers to beliefs about the safety of mobile apps and information disclosure in general (Keith et al. 2015). For that reason, two relevant features has been used to explain security of an app. The “Privacy Policy” feature which describes the developer’s privacy policy

123 declaration, otherwise known as the ‘data subject’ in the General Data Protection Regulation (GDPR). It informs users about the controller/provider and the terms and conditions of using the app, as the ‘compliance with the law and treatment of users’ (Llorens-Vernet & Miro, 2020). Out of 168 health apps, 87.5% had a declaration of privacy policy either at the bottom of the app description in the App store, as a tab after downloading the app or as information in the designer’s website. This describes the policies that the company follows for the interpretation of users’ data and its practices towards users’ personal information, for example policies on personal and anonymous information collected and the way the company uses it (e.g. disclosure of personal information if it detects actions that damage the website, for preventing illegal activities or the sale of users’ personal information to third parties for marketing purposes).It may also state the security measures (e.g. use of firewalls, secure connections on website) it uses to protect and safeguard personal information from accidental loss, misuse, unauthorized access or provide contact details in case of user disagreement with parts of the Privacy Policy. Users’ personal data could include, apart from those in an ID card, information on users’ browser type, date/time of visit, user’s location, IP address, referring URLs, the user responses, etc. Surprisingly, in the examined apps the lowest presence of privacy policy statements was identified in the Differential diagnosis category (71%) and the Patient management category (80%), which both deal with personal data, so many would think that they ought to describe their data use. Second, the feature “Authorization” to describes user’s consent to the Privacy Policy declaration. In the authorization category have been classified the apps, when users’ acceptance of the company’s terms and policies were necessary for allowing access to their personal data and exploitation based on company’s policies, for example, in order to give them out to third parties. Only after consent could the appbe downloaded. 66 (39.28%) apps that required user’s consent of terms and conditions for accessing the consumer’s data before being able to download the app (e.g. “Patients records and Appointments for Doctors” app) or access to users’ device storage (e.g. “Lancet” app) or agreement to both privacy policy and terms and conditions of content usage (e.g. “Common Differential Diagnosis” app). Authorization forms have a high presence (around 60%) in the categories of Patient management, Scientific news/lib and Connection with Doctors, since there is a need for the app to access phone features (e.g. calendar for booking appointments - cookies for storing choices for customized news –

124 audio/camera functions, respectively) and low presence (<16%) in the Calculators, Games/simulators and Quiz categories. Tier 2. The tier of reliability is measured by two app features, with the purpose of gaining users’ trust and engagement. “Reliability” describes the users’ perception that they receive what they believe they have ordered (Agag 2019). Under the feature “Credible source” are the apps which ensured the user that their content and provided information have been derived from credible organizations, such as the FDA (U.S Food and Drug Administration), Elsevier, etc., guaranteeing information accuracy. Around 45% of the apps have been developed and guided by national institutions and universities or based their content or advice on international peer-reviewed literature (JACC Journals, Lancet etc). Higher scores are identified in the Scientific news/libraries (100%) and Clinical guidelines-dictionaries-protocols (75%) categories, but not in Handbooks/manuals (32%) which are also under the education& training group but targeted mostly to students. Moreover, one would expect that Differential diagnosis and Calculators apps would score high in this feature as their role is to provide the right result using specific algorithms for each calculation, therefore, it is very important that the apps provide credible sources for their algorithms justification. However, their scores are 40% and 50%respectively.The “Feedback Contact” feature, available in almost 32% of the apps, gives the possibility to users to communicate with the developer or other health professionals (other app users or professionals in cooperation with the developer for the functioning of the specific app),via email or platforms that support calls or texting, for queries about the app’s info/results and relevant medical issues. It indicates the trust relationship between the app user and the application for the resolution of questions on the basis of feedback provision by analyzing user data. This feature is necessarily present in all Connection with Doctors apps and also has a high presence (55%) in Games/simulators for interpreting results. Tier 3. Usability describes the ability with which users can use an application to achieve a particular goal. This procedure involves the comprehension of how the system functions along with the features that offer a suitable environment for the user to learn (explainability) and attractiveness to keep using it (Zapata et al. 2015). Since a system's usability influences the direct effect on intention to use, business apps' usability is crucial for the engagement of a professional user (Gurtner et al. 2014).Therefore, usability is important for the adoption of these applications because, in many cases, they are used by stakeholders who are not familiar with technology and mobile devices

125

(Zapata et al. 2015). For this category three features have been used to measure an app’s trustworthiness. First, the “Guidance” feature (101 apps) including navigation instructions of how to interact with the app, available in the App store description.60.11% of the apps provided clear instructions of how to navigate the app and how data can be interpreted.“Social sharing” (68 apps) is the tab that gives customers the ability to share on social media (twitter, Facebook, etc), via email, the app itself or the app info/results. This dimension has been classified as a metric that could provide usability in terms of communication availability and the fast exchange of information. It provides a quick and easy way for professionals to communicate results and cases to their patients, admin teams or referral doctors by simply using this feature. All Connection with doctors apps included this feature. The lowest presence was observed in the Scientific news/libraries apps, although this feature could also be useful to those using these apps for a quick share of an interesting finding with other health professionals. Lastly, the feature “Tutorial” is an element of explainability as it provides a tutorial video for explaining the app’s functions at the App store description or app launch. Few apps included this feature (17%). The highest presence was identified in Patient management (55%) as these apps incorporate many functions and have a more complex use case. This was also the category with the highest presence of the Guidance feature (90%). In Table 7, column 4, titled “NFeature“, by comparing the mean of the sum of features per mHealth category, it indicates that in the Education & Training group, the health app category with the most features to indicate trustworthiness is “Scientific news & libraries” with a mean of 3.85 features, while the highest scores among all app categories gathers the “Connection with doctors” with a mean of 5 out of 7 features. These apps need to provide connection with large communities of professionals and accomplish worldwide regulations on privacy and security issues. Overall, the privacy policy feature is seriously considered by developers when introducing new health apps that require personal data to function as in many cases, nowadays, it is enforced by law to inform users about their data use and acquire their consent. The term “Reverse targeted display promotions as a business model of the mHealth apps. In the pool of apps, all applications are free of charge and potential sources of revenue comes from (pop-up or other kinds of) advertisements, in-app purchases (Brouard et al. 2016)to unlock extended app functionality or even from selling users’ data to third party companies, as explained in the Privacy Policy feature.

126

Earlier generation apps depend mostly on advertisements and in-app purchases to produce revenue while more and more publishers are building in-app purchase functions as a primary means of monetizing their work (Hsu & Lin 2016). The question is if this business model affects an app’s trustworthiness. Many developers provide what they call freemium apps, with free accessibility to limited functioning of the app in order to be engaged and then ask for paid full-access (premium). Other applications include advertisements or authorization of the user to have access to personal data for marketing reasons (cookies) increasing the insecurity of individuals who are seeking more for reassurances against potential risks by increasing the complexity in data governance. There is evidence in the literature that the use of these methods weakens the popularity of the apps and annoys users (Pujol et al., 2015). As advertisement and in-app purchases diminishes the usability and trustworthiness of users and most probably move directionally opposite to the other features, denoting them as a reverse category of features. Advertisements are apparent in 60% of the apps and in-app purchases in 42.2% and 27% of apps use both funding methods. In the Scientific news/libraries and Handbooks/manual app category, there is a high combination of both ads’ inclusion (89%and 40% respectively) and in-app purchases (89%& 62% respectively), indicating that these apps were mostly developed for marketing purposes. On the other hand, a low ads presence is reported in Connection with doctors with zero ads and 20% in-app purchases.

C. Integration stage: Quality evaluation Continuing to the implementation stage of the NPT, it has been used the MARS tool (Salazar et al. 2018) to measure the quality of each app. The reviewer scrutinized each of the 168 apps against the 23 MARS items using a 5-point Likert scale (1- Inadequate, 2-Poor, 3-Acceptable, 4-Good, 5-Excellent) and concluded to the given score for each app on each item. Then the mean scores for the engagement, functionality, aesthetics and information quality clusters were calculated together with the overall MARS score. The MARS tool has been used to evaluate the apps after experiencing trialability, taking into consideration the reviews of the users that were available in the Play store. Thus, the MARS evaluation score is the result of a combination of the researcher’s personal experience after using the app and the users’ comments when they were available.

127

The results from the Quality Evaluation column in Table 2 and their allocation according to their category indicate that Patient Management (3.87), Games/simulators (3.61) and Calculators (3.61) gather the highest MARS scores, meaning that in the quality measure these health app categories are stronger than the rest. Looking in Table 9 at the individual scores of each MARS cluster, the highest score is achieved with regards to apps’ information quality and that there is still room for improvement in the apps’ aesthetics and relatively room for improvement in all MARS clusters. “Doctor At Work (Plus)-Patient Medical Records” app of the Patient management category holds the higher score (4.9) in MARS, which is also the indicative app of this category. “VisualDx” app and “MedShr: Discuss Clinical Cases” come next, both with a score of 4.62, and belong to the Differential diagnosis and Connation with doctors’ categories respectively. Table 9: MARS clusters’ scores MARS clusters Mean Sd Engagement score 3.29 0.83 Functionality score 3.47 0.82 Aesthetics score 2.96 0.93 Information quality score 3.49 0.85 Total MARS score 3.30 0.86

C. Integration stage. Health apps popularity and engagement The popularity of an app is related to the user satisfaction, which can be measured by the average user rating and the number of app downloads (Krishnan and Selvam 2019).For an app to be successful, users need to download it, use it, rate it and provide useful reviews to other potential users who would be triggered to use it. This circle increases engagement with the app and the number of users (as long as the ratings are good) (Krishnan and Selvam 2019). In Table 10, the last two columns provide descriptive statistics about the 168 apps’ popularity based on their category and, in particular, show the number of rating stars (scale:1-5) and the number of downloads that were indicated in the app description at the App store at the period of the investigation (Jan/Feb 2020). Only 16 apps (9.5%) had no rating and 23 apps received an average rating below 4*. The remaining apps (73%) have been rated with 4 stars and over (>4*), meaning that users were satisfied. On the other hand, most of the apps(almost 70%)have been downloaded between 10,000 and 100,000 times and around 26% of them less than 10,000.Observing simultaneously the average downloads and the mean features it seems that users mostly

128 download the “Connection with doctors” app category which holds the most features across the trustworthiness tiers. “Internal Organs in 3D Anatomy”, which belongs to the “Handbook/Manual” category is the app that has been downloaded more than 5 million times and rated with 4.5 stars on average from more than 35,000 users at the time of the investigation. This application was released in the Play store in 2014. A data comparison between downloads and the years since the first release of the apps shows no evidence that the popularity of apps is related to the period available in the Play store (p<0.000). According to the results, the highest scores in the stars rating is from the Quiz and Scientific news/libraries categories and in downloads the highest score is from both the Scientific news/libraries and Connection with doctors. However, the MARS score has identified the Patient Management app category as the category with the higher quality.

4. 2c. Results from statistical tests: Comparison of Means In order to investigate the dynamics and associations among the variables that consist the elements of the conceptual model for mHealth apps integration in the clinical practice (Fig. 3), there have been conducted statistical tests with the use of the SPSS statistical software. Table 10 entails the description of the variables that have been used in the research framework.

Table 10: Description correlation variables Variables Description 1 Health types Based on the health type category that belongs each app takes one of the following values: 1. Education & training, 2. Decision making on diagnosis, 3. Management (measurement scale 1-3) 2 NFeatures The sum of the 7 features as described in Table 3 without the features of Reverse Usability (measurement scale 0-7) Tier1 SECURITY 3 Privacy policy Dummy variable, the app takes the value 0 to indicate the absence of the feature or 1 to indicate the presence of the feature (measurement scale 0, 1) 4 Authorization Dummy variable (0,1) Tier 2 RELIABILITY 5 Credible source Dummy variable (0,1) 6 Feedback Contact Dummy variable (0,1) Tier 3 USABILITY 7 Guidance Dummy variable (0,1)

129

8 Social sharing Dummy variable (0,1) 9 Tutorial Dummy variable (0,1) Reverse Usability 10 Purchase Dummy variable (0,1) 11 Ads Dummy variable (0,1) 12 MARS Evaluation quality score MARS (measurement scale 1-5) 13 STARS App’s average star rating (measurement scale 1-5) 14 LgDOWNS Log number of downloads (0 < “1” < 1,000, 1,000 ≤ “2” <10,000, 10,000 ≤ “3” < 100,000, 100,000 ≤ “4” < 1,000,000, “5” ≥1,000,000) (measurement scale 1-5) 15 NDownloads Number of app downloads as described in the app store (Feb 2020)

Based on the research framework, the “integration” stage of the process of the described innovation is based on the quality evaluation of the apps (MARS), the stars rating and the number of downloads. To identify whether the scores of these three measures are affected by the existence or not of the examined app features (described in tiers 1-3) there is a comparison of means using statistical hypothesis testing. The hypothesis have been formed between the three “integration” measures and the 9 app features, which are all presented in Table 11 together with the results. In order to determine if there is a significant difference between the means of the examined independent samples initially, are performed normality tests to the variables (Kolmogorov-Smirnov test) and found that these are normally distributed but in three cases. Thus, the t-test was used in the statistical analysis for all variables and Mann– Whitney U test for the three exceptions of non-parametric data, as indicated in Table 11. It is clear that the only three hypotheses that are rejected are these that prove a relationship between the app quality score (MARS) and the existence of credible source, feedback and guidance features. The mean MARS scores are significantly higher when these features are present in the apps. Accordingly, the test of mean differences indicates that STARS rating is unrelated to the existence of the certain app features but for tutorials. Finally, the average number of downloads is only related to the existence of social sharing features and tutorials with a higher number of downloads when these features are available.

Table 11: Hypothesis testing The average MARS score of the Hypothesis apps is not related to the existence of: Mean scores p-value Decision NPT model Existence of feature: NO YES TIER I SECURITY Privacy Policy 3.08 3.22 .418 Accept

130

Authorization 3.18 3.22 .658 Accept TIER II Credible Source* 3.03 3.40 .001 Reject RELIABILTY Feedback* 3.06 3.50 .003 Reject Guidance* 3.00 3.33 .005 Reject TIER III Social Sharing 3.16 3.26 .353 Accept USABILITY Tutorial 3.16 3.38 .088 Accept In-app purchase 3.17 3.25 .472 Accept Reverse Usability Ads 3.28 3.12 .126 Accept The average STARS rating of the apps is not related to the existence of: Privacy Policy 4.00 3.99 .864 Accept TIER I SECURITY Authorization 4.03 3.79 .241 Accept TIER II Credible Source 3.91 3.94 .873 Accept RELIABILTY Feedback 3.80 4.19 .085 Accept Guidance 3.72 4.06 .105 Accept TIER III Social Sharing 3.90 3.97 .742 Accept USABILITY Tutorial 4.03 3.45 .031 Reject In-app purchase 4.02 3.79 .282 Accept Reverse Usability Ads 3.99 3.87 .550 Accept Hypothesis The average number of app downloads is not related to the existence of: Privacy Policy 32288.1 129832.7 .334 Accept TIER I SECURITY Authorization 148017.2 74054.5 .276 Accept TIER II Credible Source 113475.5 122680.4 .891 Accept RELIABILTY Feedback 111943.2 130000.0 .802 Accept Guidance 135735.2 105635.6 .660 Accept TIER III Social Sharing 58651.6 204386.7 .031 Reject USABILITY Tutorial 85074.3 267440.0 .036 Reject In-app purchase 72993.7 188386.3 .092 Accept Reverse Usability Ads 64415.2 173460.4 .102 Accept * Non parametric data: use of Mann–Whitney U test Reject: p <0.05

4.2d. Results from statistical tests: Correlations

In addition to the previous analysis, it has been further tested the existence of intercollerations, and their level of significance, between any two variables that take part in the research framework including the mHealth apps’ categories (i, ii & iii), the individual app features, and the dimensions of integration of the NPT model as described in the research framework (MARS quality score, STARS rating and number of downloads).Table 12a presents the results of the Pearson Correlation matrix. To satisfy the linearity assumptions of the correlation model, the number of downloads was converted to a log number of downloads for each app as shown in Table 5.The highest positive significant correlation (0.558) is observed between log number of downloads and star rating, and between log number of downloads and the quality score (MARS) (0.373).MARS is the variable that presents the highest number of significant

131

correlations with other variables. Apart from the downloads it is also correlated with the STARS rating (0.244) and also with the features of the “reliability tier”, namely, the existence of credible source (0.262) and feedback mechanism(0.291) as well as with the guidance element(0.232), as also observed in Table 12b (comparison of means). Additionally, it is also correlated to the health type (0.239). The latter means that apps belonging to the categories 2 (decision-making) and even more 3 (patient management) score higher in MARS. With regards to the app features, privacy policy is positively related to the existence of feedback (0.179), in-app purchase and ads (0.189). This is logical since the privacy policy needs to inform about communication rules and the existence of promotional material. App credible source is related to the existence of the authorization feature (user consent) (0.165) and most importantly to the MARS quality score (0.262) meaning that the apps with credible sources are considered of better quality. Moreover, the existence of the social sharing tab is positively associated with the existence of the feedback mechanism (0.301) and ads (0.165), which latter makes sense since a marketing strategy would aim at increasing the ads visibility through app sharing. Lastly, ads and in-app purchases are also weekly positively correlated with each other (0.178), so one would expect that an app which includes in-app purchases will also include ads, and vice versa.

Table12a: Correlations of the conceptual model variables and individual app features

IMPLE MENT. EMBEDDING INTEGRATION NPT Var Tier I Security Tier II Reliability Tier III Usability ReverseUsability Health Type Privac Autho Credible Feedback Guide Social Tutor Purch Ads MARS STARS LgDOWN HealthTyp 1 -.019 .0561 -.007 .151 .187* -.014 .083 -.024 -.114 .239** .13 .03 Privacy -.019 1 .0229 .018 .179* .133 .128 .035 .189* .189* .063 -.013 .043 Auth/tion .056 .023 1 .165* .006 .013 .076 -.105 .057 0,000 .0344 -.090 -.083 Credible -.007 .018 .165* 1 .129 .105 -.018 -.111 .064 -.145 .262** -.125 .118 Feedback .151 .179* .006 .129 1 .291** .301** .118 .040 -.125 .291** .133 .098 Guidance .187* .133 .013 .105 .291** 1 .102 .094 -.026 -.031 .232** .125 .089 Social -.015 0.128 0.076 -.018 .301** .102 1 .059 .092 .165* .0721 .026 .010 Tutorial .084 0.035 -.105 -.111 .118 .094 .059 1 .108 -.019 .132 -.166* .001 Purchase -.024 .189* .057 .063 .0398 -.026 .092 .108 1 .178* .056 -.083 .163* Ads -.114 .189* -.040 -.145 -.125 -.031 .165* -.019 .178* 1 -.119 -.046 .083 MARS .239** .063 .034 .262** .291** .232** .072 .132 .056 -.119 1 .244** .373** STARS .13 -.013 -.090 .012 .133 .125 .026 -.166* -.083 -.046 .244** 1 .558** LgDOWN .03 .043 -.083 .11 .099 .089 .010 .001 .163* .083 .373** .558** 1

**p <0.001, * p <0.05 Number of obs = 168

132

Table 12b demonstrates the intercorrelations between the research dimensions and the sum of the 7 features of the embedding stage (tiers 1-3), instead of the individual features of Table 12a. Therefore in addition to the initial correlation table, Table 12b indicates the dynamics of apps that gather more than one feature. The most interesting observation is the association of the number of features and the MARS quality score (0.346) indicating that the apps with more features are considered of higher quality. Moreover, the number of features of an app has a week positive correlation (0.166) with the health type it belongs. So, if the app belongs to categories 2 and 3 will most probably incorporate more features overall compared to the educational category apps.

Table 12b: Correlations of the NPT conceptual model elements (SUM of app features)

IMPLEMENTATION EMBEDING INTEGRATION NPT Var HealthType Nfeatures MARS STARS LgDOWN HealthType 1 .166* .239** 0.126 0.030 Nfeatures .166* 1 .346** 0.029 0.120 MARS .239** .346** 1 .244** .373** STARS 0.130 0.029 .244** 1 .558** LgDOWN 0.030 0.121 .373** .558** 1 **p <0.001, * p <0.05 Number of obs = 168

4.3a. 3rd Conceptual framework . Τhe literature addresses factors that lead to mobile app downloads, but researchers have yet to develop a comprehensive theoretical approach that explores these factors for professional users. This study, also attempts to fill this gap for health professionals’ mHealth apps. In detail, this study investigates both the role of usefulness, usability and trustworthy characteristics of mHealth apps, and the influence of evaluation attitudes and of subjective norms, which encourage medical professionals in the adoption of apps in their clinical procedure. For this purpose, there has been developed a theory-based structural model that identifies the consumption patterns and determinants of choice preference across health professional apps.

133

To achieve this, a combination of two theoretical models has been used to investigate the consumer behavior intention in the process of using mHealth apps in the everyday clinical practice. The Technology Acceptance Model (TAM) (Davis, 1989) uses key variables of user motivation such as, perceived usefulness (PU) which is translated to users’ belief that technology would enhance their job or task performance, perceived ease of use (PEU), meaning users’ belief that using technology would be free from effort, and other perceived attitudes toward technology (ATT), such as perceived trust (PT), to explore the behavioral intention (BI) of technology use (Scherer, 2019). The TAM has been widely used in several technology adoption related studies indifferent mobile service contexts and is the most reliable for measuring technology adoption intention (Shankar & Kumari, 2019). Therefore, in this study TAM has been used as a grounded model, where as a core variable of PU is used the health app capabilities, of PEU the app features that enhance usability and of ATT it has been included perceived trustworthiness (PT), for making it suitable to mHealth context. The model has been further extended using the “Theory of Reasoned Action” (TRA) to also explain the behavior beliefs that could influence the behavior intention of a certain demographic sample, this of health professionals, examining the influence of personal determinants and social reflections (Ye et al., 2019). The “Theory of Reasoned Action” is based on the concept that humans during forming an attitude towards a behavior are constantly evaluating the relevant negative or positive behavior and behavior intention is a function of two factors, attitude towards performing the behavior itself (AB) and one's subjective norm (SN) (Moore & Benbasat, 1996). Table 13 describes the constructs that participated in the combination of the two models affecting the process of integration intention of professional mHealth in medical practice along with the created external variables. According to the constructs of TAM model, perceived usefulness (PU) has been translated as the sum of different kind of health app capabilities that each mobile app offers and health professionals are willing to use believing that they will increase performance in the workplace, perceived usability (PEU) gathers the sum of features that assist users to reduce time and effort while using them, and trustworthiness (PT) are the sum of app features that increase users’ trust. The latter has been added as a moderating factor due to its direct or indirect mediation effect on user intention or adoption of new technology (Ye et al., 2019). Continuing with the TRA constructs, subjective norm (SN) has been translated by social influence (SN) measured by the

134 stars’ rating evaluation and attitude towards performing the behavior (AB) has been based on the observability after trial of the apps before adoption measured by the individuals’ quality evaluation (MARS tool).

Table 13: The study’s conceptual constructs of TAM & TRA Theory Constructs Key variables TAM Perceived usefulness mHealth app capabilities (PU) TAM Perceived ease of use Feature Relevance (PEU) TAM Moderating factor Perceived trust Feature Relevance (PT) TRA Subjective norms Social Influence (SN) TRA Attitude Towards Behavior Quality Evaluation (AB)

Fig. 4 presents the conceptual framework of the TAM & TRA model following the process of mHealth app adoption intention of professionals in healthcare. All the above variables are further explained in Table 14. Through the review of the health apps addressed to medical professionals, there attempt was to provide an investigation of their purpose and features and to explore the relationships between the constructs that participate in this process. Thus, 5 hypotheses have been formed to understand the relationships between the variables of the framework, which are depicted in Fig. 4.

Figure 4: Conceptual framework with research hypotheses of mHealth apps adoption intention from professionals.

135

4.3b Hypothesis testing

The current research proposes a conceptual path from perceived usefulness, perceived usability and trust to the motivation of a health professional to download an app (Donker et al., 2013; Gagnon et al. 2016; Venkatesh et al. 2002). There is still increasing debate among clinicians as to whether this market is driven by the technological novelty or by clinical usefulness (Istepanian & Al-Anzi, 2018). In order to test the potential effects on user behavior and given this conceptual linkage between perceived usefulness (Donker et al., 2013; Gagnon et al., 2016) the hypothesis are: H1: Usefulness (PU) is positively related to Behavior Intention (BI) to download the app. The existing literature has also indicated that usability and ease of use has become an important topic for smartphones, since it is necessary to prevent applications from being difficult to use (Llorens-Vernet, & Miró, 2020; Venkatesh et al., 2002; Zapata et al., 2015). App’s clear functionality is one contributor to success as complex functionalities tend to confuse users in decision-making (Alnsour et al. 2016). Therefore, the hypothesis is that:

136

H2: Ease of use (PEU) is positively related to Behavior Intention (BI) to download the app. In addition to perceived usefulness and usability, perceived trust should also act as a significant determinant of behavioral intention, consistent with the findings across researches and since a failure to protect customer information can lead to problems which would threaten the trust relationship with its customers (Byambasuren et al., 2019). Therefore, the hypothesis is that: H3: Trust (PT) is positively related to Behavior Intention (BI) to download the app. Researchers highlight the moderators of the effects of online ratings and whether they have become ubiquitous (Kübler et al., 2018) indicating that download choices are shaped by other participants preferences (Carare, 2012). Based on that, the next hypothesis is: H4: Social influence (SI) of app star rating is positively related to Behavior Intention (BI) to download the app Similarly, extensive literature has explored the relationships between app characteristics and quality mostly measured via user reviews (Lee, & Raghu, 2014; Kim et al. 2018; Krishnan and Selvam 2019) but also with specific tools like MARS (Knitza et al. 2019; Stoyanov et al. 2015). Adjusting these propositions, the hypothesis are: H5: Quality evaluation (AB) is positively related to Behavior Intention (BI) to download the app. The research model of Fig. 4 links the hypotheses with the moderators of user behavior intention, which in this case is targeted to health professionals/students, for creating knowledge about the role of app features’ relevance during mobile app choice for work related information and management. Moreover, Table 14 depicts the similarities and differences between the most relevant empirical studies and our research approach for the specific hypotheses. Many studies have explored the relationship between user rating (social influence) and downloads (behavior intention) but not as many have focused on the other aspects of the model.

Table 14: Comparison of the study with similar studies on the hypotheses testing Hypothesis What the Literature has investigated: How this study differs/contributes: H1. PU is positively related BI A longitudinal investigation of how employee The study uses a combination of TAM and H2.PEU is positively related to BI moods during computer technology training TRA. It is targeted specifically to health influence motivation (behavioral intention) to use apps for professionals and measures how that technology and how the latter is affected by user behavior intention (downloads) is perceived technology usefulness (answers from affected by perceived usefulness (as a survey) and perceived ease of use (required effort), construct-sum of app tasks) and perceived

137

based on TAM and tested in a regression model ease of use (as a construct-sum of app Venkatesh et al., 2002 features) H3.PT is positively related to BI They used survey data from Australian health The study is based on a regression model for practitioners to test whether one of the barriers to prediction based on data gathered from apps app prescription was the lack of trust towards and not on a description of survey results mHealth apps (Byambasuren et al. 2019) from users H4.SI is positively related to BI Empirical analysis of any type of apps with data The analysis is concentrated on mHealth collected from Google Play and Apple App stores apps targeted to health professionals with the determined through a regression model that a combination of a market and social greater number of downloads was significantly dimension incorporating also factors of positively associated with, user rating, number of reliability and usefulness as major individual reviews and in-app purchases amongst other and variables IV negatively with in-app ads(Ghose & Han, 2014) Empirical market analysis that investigated with the The analysis is concentrated on free apps use of regression the relationship between app targeted to health professionals with the ratings and app sales (paid apps) in several combination of a market and social countries (Kübler et al 2018) dimension including downloads as the dependent variable (DV) Empirical analysis of all apps provided in the The analysis is concentrated on free apps Apple’s App store that showed through a regression available in the Android app store targeted model that the willingness of consumers to pay for specifically to health professionals and an app is greater for high rated apps. The DV of the emphasizes the social dimension as well as model was bestseller rank (which incorporated the the market dimension. The DV is downloads element of downloads). The study also included and the are numerous different IVs data on age and size of apps but not on downloads (Carare, 2012) Empirical analysis of mHealth apps specifically for The study takes a similar approach but Psychiatry with data collected from Google Play includes all free mHealth apps for Store determined through a regression model that a professionals, for all medical specialties, greater number of downloads was significantly incorporating also factors of reliability and associated with apps with higher user rating, usefulness as major IV cheaper apps and with available in-app purchases but not with number of reviews, app size, number of screenshots, length of description, availability in the Apple App Store and new published versions. (Pinheiro et al. 2019) Empirical analysis of mHealth apps specifically for The study takes a similar approach but Maternal and Infant Health with data collected from includes all free mHealth apps for Google Play and Apple App stores determined professionals, for all medical specialties, through a regression model that a greater number of incorporating also factors of reliability and downloads was significantly positively associated usefulness as major IV with user rating, in-app purchases and in-app advertisements and negatively associated with app price and period since last update (Biviji et al. 2020) Empirical analysis of mHealth apps specifically for The study takes a similar approach but diabetes with data collected from Google Play Store includes all free mHealth apps for determined through a regression model that a professionals, for all medical specialties, greater number of downloads was significantly incorporating also factors of reliability and positively associated with apps with higher user usefulness as major IV rating, free apps, app age, developed in US but negatively with months since last update (Krishnan and Selvam 2019) There are no identified studies that investigate the The study incorporates the app quality score H5.AB is positively related to BI effect of app quality (measured in other ways than (measured by MARS) as an independent user rating) to app downloads. variable for determining downloads for mHealth apps for professionals

4.3c. Regression model variables

The identification of variables that were used as predictors of downloads were based on the literature ( Ghose, & Han, 2014; Lee, & Raghu, 2014; Pereira-Azevedo et al. 2016; Krishnan & Selvam 2019). Each app was reviewed for its characteristics based on the information provided by its developer and its user comments (Hoeppner et al.

138

2016). The collected descriptive information of the apps (e.g. number of downloads, star ratings, etc.) together with their usefulness, ease of use and trust aspects are all presented in Tables 15a-c.

A. Depended variable

Behavior Intention: Downloads (BI)

The dependent variable derives from the number of downloads of each app provided by the app store. The particular variable has been chosen following the approach of numerous studies that selected the particular measure to determine which apps are preferred for use (Ghose and Han 2014; Krishnan and Selvam 2019; Biviji et al. 2020; Pinheiro et al. 2019; Pereira-Azevedo et al. 2016; Engström & Forsell 2018). In this case the purpose is to measure the intention to try and use the innovation intervention offered by these mHealth apps (Krishnan and Selvam, 2019). However, in order to satisfy the linearity assumption of the regression model, instead of the actual number of downloads, the downloads have been measured as the log number of downloads of every app (0 < “1” < 1,000, 1,000≤ “2” <10,000, 10,000≤ “3” <100,000, 100,000≤ “4” <1,000,000, “5” ≥1,000,000).One app has been downloaded over 5,000,000 times and around 30% of the apps have been downloaded between 10,000 and 100,000 times (Table 19). In order to crosscheck the results for accuracy on download information, an external mobile app aggregator has been used, the App market platform 42matters (Krishnan & Selvam, 2019) without finding any descrepancies in the provided values.

B. Independent Variables Perceived Usefulness (PU) - Health app capabilities Perceived usefulness was measured based on the identified mHealth app capabilities of each app according to the health professionals’ needs. Following the conceptual model and the identified TAM variables, the ‘mHealth app capabilities” represent the “driving motivation of a health professional to adopt the innovation” or else what expected capabilities this innovation will bring to the professionals in every day clinical routine. Based on the literature (Mosa et al. 2012; Ventola 2014) and after screening the content of all selected apps, eight capabilities have been indentified that

139 would be very helpful to health professionals during practicing medicine, which were handled as binary variables based on whether they existed or not in each app. These capabilities are basically related to the purpose of the app, its content and its methods/processes. Table 15a presents the 8 capabilities and their percentage of their overall presence in the 168 apps. For the regression model the construct “perceived usefulness” (PU) has been created, which is measured as the sum of the capabilities offered by the app, assuming that medical professionals/students would consider more useful the apps that offer more capabilities.

Table 15a: mHealth app capabilities (PU) based on app content and features Capabilities Description N % 1. Bug fixes App upgrades for solving operational and 80 47.6 calculus/content issues 2.Software Clinical decision support systems that provide 78 46.4 applications assisting treatment guidelines, differential diagnosis aids, in decision -making medical calculators, laboratory test interpretation 3. Medical Training Knowledge assessment tests, board exam preparation, 76 45.2 case studies, eLearning and teaching, surgical simulation, skill assessment tests 4.Informational Connection to medical literature databases, textbooks, 63 37.5 resources journals, literature search portals, drug reference guides, medical news 5.Artificial Provision of image and speech recognition, customized 17 10.1 Intelligence results, cognitive computing, automatic analysis, machine learning, designed to accept different kind of personal data 6.Health information Keeping and/or accessing electronic medical records, 13 7.7 management images and scans, electronic prescribing, coding and billing 7.Communication Features designed to facilitate video conferencing, text, 10 5.9 capabilities with and e-mail, social networking doctors and patients 8.Patient Ability to schedule appointments & meetings 6 3.6 management

Perceived Ease of Use (PEU) Perceived usability (PEU) was measured based on the presence of certain app features (Table 15b). Ease of use describes the ability of users to use an application for a certain purpose without spending much time and effort to comprehend the functioning of the system. This is achieved through app features that offer a suitable environment for the user to learn (explainability) (Zapata et al. 2015). Therefore, perceived ease of use is important in the adoption of these applications, that in many cases are used by stakeholders who are not familiar with technology and mobile devices (Zapata et al. 2015). For this category four features have been used to measure an app’s usability.

140

First, the “Guidance” feature includes the navigation instructions provided by the app store on how to interact with the app and how data can be interpreted.“Social sharing” is the tab that provides customers the ability to share on social media (twitter, Facebook, etc.) or via email the app itself or the app info/results. This dimension has been classified as a metric that could provide ease of use because it provides a quick and easy way for professionals to communicate results and cases to their patients, admin teams or referral doctors. “Offline use” refers to the ability of users to use the app without internet connection. For the “Tutorial” feature has been considered a tutorial video for explaining the app functions at the app store description or app launch. For the regression model, it has been examined for each app the presence of these features and counted the number of available features.

Table 15b: Perceived Ease of Use (PEU) features PEU Features N % 1. Guidance 101 60.1 2. Social sharing 67 39.9 3. Offline use 51 30.2 4. Tutorial 29 19.3

Perceived Trustworthiness (PT) Four relevant features have been used to explain trustworthiness of an app including privacy, security and reliability elements, features that describe trust and have been considered in the relevant literature as significant (Llorens-Vernet & Miro 2020; Nicholas et al. 2015). The “Privacy Policy” feature, which describes the policies that the company follows for the interpretation of users’ data and its practices towards users’ personal information and security measures(e.g. the sale of users’ personal information to third parties for marketing purposes, use of firewalls), known as ‘data subject’ in the General Data Protection Regulation (GDPR). Out of 168 health apps, 87.5% had a declaration of privacy policy. Second, the feature “Authorization” has been used to describe user’s consent to the Privacy Policy declaration and for the app accessing phone features and personal data. 66 (39.28%) apps required user’s consent of terms and conditions otherwise app downloading was not possible. For the element of reliability, another two app features have been measured, with the purpose of gaining users’ trust and engagement. Under the feature “Credible source” are been classified the apps, which ensured the user that their content and info have derived from credible

141 organizations such as the FDA (U.S Food and Drug Administration), Elsevier, etc. guaranteeing information accuracy. In the sample around 40% of the apps have been developed and guided by national institutions and universities or based their content or advice on international peer-reviewed literature (JACC Journals, Lancet, etc.). The “Feedback Contact” is the app feature that allow users to communicate with the developer or other health professionals for queries about the app’s info/results and relevant medical issues. It indicates the trust relationship between the app user and the application for the resolution of questions on the basis of feedback provision through the analysis of user data. For the regression analysis model, it has been measured the sum of the trustworthiness features of each specific app.

Table 15c: Perceived Trust (PT) features Trustworthiness Features N % (PT) 1. Privacy policy 147 87.5 2. Authorization 66 39.2 3. Credible source 66 39.3 4. Feedback Contact 53 31.5

Subjective Norm - Social Influence: Star rating (SN) Based on the literature, positive user evaluation is a measurable characteristic of mobile application success (Alnsour et al. 2016) and there is relevance between the star rating of an app and its number of downloads (Wisniewski et al. 2019). In this research, the social influence of mHealth apps for professionals has been measured as the star ratings which reflect the positive or negative opinion of a user who has downloaded the health app. Taking into consideration that the nature of the investigated apps is targeted to professionals, their ratings are valid and their opinion is quite reliable and realistic. This variable as the average number of star ratings (1-5) given by its users for each app as described in the app store. The mean star rating of all apps was 3.9 at the time of the investigation, as shown in Table 18. Around 10% of the apps (16 apps) have not been rated at all and almost 34% have been rated with more than 4.6 stars (57 apps).

Attitude towards behavior - Quality evaluation: MARS (AB) Descriptive characteristics and online feedback, such as consumer ratings, is considered an aggregation of subjective evaluations (Grover et al. 2006) and therefore

142 alone may not be an adequate measure for app evaluation. Taking this into account, at this stage of the analysis, MARS (Mobile App Rating Scale) validated tool (Salazar et al. 2018) has been used to measure the quality of each app. As mentioned in the second analysis of the research, the MARS evaluation scale is a reliable and objective instrument that measures the degree with which mHealth apps satisfy quality criteria, it is easy to understand and use with minimal training (Stoyanov et al. 2016; Stoyanov et al. 2015). It is a common health app rating evaluation scale for apps about weight loss (Bardus et al. 2016) and smoking cessation (Patel 2015). As already mentioned the MARS 23-item rating scale is organized into 4 clusters, the engagement, functionality, aesthetics and information quality categories. The average scores on each cluster for all 168 apps are shown in Table 9. The 23 individual MARS items are measured against a 5-point Likert scale (1-Inadequate, 2- Poor, 3-Acceptable, 4-Good, 5-Excellent). In the regression model the quality evaluation variable (AB) was measured for each app by the average MARS score (Table 9).

Control Variables Control variables are related to dependent variables, influence the outcome, but mainly arise from the experimental design and are not of crucial interest variables to the different regression models. According to the literature, there is evidence that the following control variables can affect apps’ consumer intention (Ghose, & Han, S. P. 2014; Lee, & Raghu, 2014; Pereira-Azevedo et al. 2016; Krishnan and Selvam 2019). First is the variable “SIZE”, which is the size of every app in megabytes. Second is the variable “REVIEWS” presented as the number of users that rated the apps. The estimation of the length of description in the app store has been counted as the sum of the words’ letters of the description text in the app store (LENGTH). The next two control variables are “PURCH” and “ADS” which are dummy variables, taking the value 1 if the app includes in-app purchases or advertisements respectively and 0 otherwise. The “YEARS” are measured as the number of years from app’s first release since the date this research was conducted. The last control variable is multiplatform (MltPLATF), a dummy variable which identifies whether the app is offered in both android and its platforms or only android. Lastly, we added a control variable to capture developer’s experience in designing such apps (DEV) and we gave the value of 0 if the

143 app has a unique developer in our dataset and the value of 1 if the developer has designed more apps in our dataset.

Table 16: Description of regression analysis variables

Depended Variables Description 1 BI Log number of downloads (measurement scale 1-5) Independent Description Variables 1 PU The sum of capabilities offered by apps as explained in Table 3a (measurement scale 0-8) 2 PEU The sum of ease of use features as described in Table 3b (measurement scale 0-4) 3 PT The sum of trust features as described in Table 3c (measurement scale 0-4) 4 SN Average number of star rating (measurement scale 1-5) 5 AB Average MARS score (measurement scale 1-5)

Control Variables Description 1 SIZE Average size of apps (in MB) 2 REVIEWS Number of users who rated the app 3 LENGTH The sum of letters in Play Store’s description without spaces 4 PURCHAS Dummy value 1 if there are app purchases ,0 otherwise 5 ADS Dummy value 1 if app displays advertisements, 0 otherwise

6 YEARS Number of years since first release 7 MltPLATF Dummy value 1 if app displays in other App Stοres, 0 otherwise 8 DEV Dummy value 1 if app has been created from the same developer to at least one other app in the sample, 0 otherwise

Table 16 presents the intercollerations, and their level of significance, between any two variables that take part in the research framework including the mHealth apps’ individual features, which in the model form constructs. The highest positive significant correlations of around 0.5 are observed between log number of downloads and: a) star rating, b) number of reviews and c) year since first release as well as the quality score (MARS) with a more moderate effect. Moderate correlations are also observed between MARS and feedback, MARS and offline use, and reviews with size of app. The feature credible source has a somehow weak but accurate positive correlation with most variables and app features. Ads and offline use features are negatively correlated with many of the other variables. Size of apps is correlated only with number of reviews.

144

Table 17: Correlation coefficients of individual variables

4.3d. Regression analysis results

A. Descriptive statistics of variables

In Table 18 there are the accumulated descriptive characteristics of the sample of the 168 mHealth apps targeted to medical professionals. The mean star rating of all 168 apps is 3.9 out of 5 and that on average 1504 users rated each app. The average MARS quality score is 3.2, lower than the star rating score. The mean size of the apps is 28Mb (megabytes) and most of the apps have between 10,000 and 100,000 downloads (log scale 3). On average, the apps have been available in the app store for 3.6 years. Almost 50% of them are available in more than one platform and include in-app ads, and almost 40% of them incorporate in-app purchases.

Table 18: Descriptive statistics and frequencies of model variables Variables min max mean sd LgDOWNS 0.0 6.0 3.3 0.9 1 PU 1.0 7.0 2.0 0.8 2 PEU 0.0 4.0 2.0 0.9 3 PT 0.0 4.0 2.1 1.0 4 SN 0.0 5.0 3.9 1.3 5 AB 1.0 4.8 3.2 0.7 6 SIZE 0.001 1658.88 27.93 128.54

145

7 REVIEWS 0.0 35646.0 1503.9 4186.0 8 LENGTH 4.0 818.0 253.3 162.0 9 PURCHAS 0.0 1.0 0.38 0.49 10 ADS 0.0 1.0 0.48 0.50 11 YEARS 0.0 9.8 3.6 2.3 12 MltPLATF 0.0 1.0 0.48 0.5 13 DEV 0.0 1.0 0.21 0.41

B. Regression analysis results

Multiple linear regression analysis was performed to study the association between the mHealth apps downloads and the independent and control variables (Table 18) to validate the research hypotheses H1 to H5.The validity of the regression assumptions was first tested for the normal distribution of errors/residues with the graphical representation of the normal probability. The absence of multicollinearity was also confirmed. The table 19 presents the Pearson correlation matrix. The intercorrelations among the various constructs are all within an acceptable range. Taking into account the evidence of this evaluation, it can be stated that none of the model assumptions were violated.

Table 19: Pearson Correlations of regression model variables

146

The regression model has the following structure:

= + α1PU+ α2PEU+ α3PT + α4SN+ α5AB+ α6REVIEWS +α7SIZE + α8LENGTH +α9PURCHAS +α10ADS+ α11YEARS +α12MltPLATF+α13DEV+ u

The results of the linear regression, Table 19, provide support for accepting hypotheses H1, H4 and H5 by showing that there is a positive correlation between the independent variables of perceived usefulness (PU), users’ star rating (SN) and app quality evaluation (AB) and the dependent variable of the log number of downloads (BI). It further indicates that downloads are positively related to the number of reviews (REVIEWS), length of description in app store (LENGTH), in- app advertisements (ADS) and years since app’s first release (YEARS).

Table 20: Hypothesis analysis results Hypothesis Coefficients Tvalue p-value Conclusion Independed Variables H1 (BIis affected by PU) 0.121 2.008 0.046* supported H2 (BI is affected by PEU) -0.022 -0.391 0.696 Not supported H3 (BI is affected by PT) -0.026 -0.485 0.629 Not supported H4 (BI is affected by SN) 0.306 8.031 0.000** supported H5 (BIis affected by AB) 0.181 2.290 0.023* supported Control variables REVIEWS 6,709E-5 4.942 0.000** SIZE 0.000 -0.638 0.525 LENGTH 0.001 2.082 0.039* PURCHAS 0.119 1.106 0.270 ADS 0.220 2.071 0.040* YEARS 0.102 4.335 0.000** MltPLATF 0.096 0.911 0.364 DEV -0.166 -2.278 0.203 **p < 0.001, * p < 0.05 R-sq: = 0.615Prob> F = 18.945 Number of obs = 168

147

CHAPTER B. 5

5. Agenda with valuable insights about mHealth apps for professionals

5.1. mHealth apps: What exists

From the survey findings, the majority of the apps (66%) concerned a wide range of medical specialties and were intended to fulfill educational needs (63%) which appeals mostly to medical students, with “Handbooks/manuals” being the most populated app category (28%). “Decision making on diagnosis” was the next most useful group of apps (30%) with the “Differential diagnosis” category being the most popular of them (21%).Based on certain features are being explained three tiers that mHealth users based their trustworthiness on when experiencing trialability in order to further use an app in an everyday clinical routine. Features that describe the level of “security” of users’ personal data are addressed by the majority of designers (87.5%) by including an official declaration. However, only 39% of these apps were asking for users’ consent. If the users deny agreement, they are not allowed to continue with the app. Another analysis of the 600 most used mHealth applications, as of May 2013, showed that, on that date, only 30% of the apps had a privacy policy (Benjumea 2019). Considering this statistic it can be assumed that nowadays, and after the wide discussion of regulations upon the security of personal data, the vast majority of health apps choose to inform the potential users about their privacy policy and some of them require acceptance before allowing navigation to the app itself. However, this take it-or-leave-it business model makes users powerless to negotiate with the app provider and are somehow forced to accept the rules without adequate information. Reliability, as a measure of app trustworthiness, is given by the features of Credible source (present in 39% of apps) and Feedback contact (29%). Both features have low presence if we consider that these apps are made for professional use and therefore should make explicit the fact that their info and results are 100% reliable. If the app developers wish to expand their pool of users, these features should be taken into serious consideration as they would be responsible for shaping part of health professionals’ education, diagnosis and management capabilities. Usability is the third

148 measure of building app trustworthiness. The feature of Guidance is present in only 60% of the apps, which is considered low, and if we assume that new users would expect to get navigation instructions or a Tutorial (17%) for the use of the app to be able to use it and receive the maximum benefit of the app’s capabilities, but do not receive explanations on how to use the app, users may not want to keep it. Connection with doctors is the health app category that gathers the most features relevant to trustworthiness (with a mean of 5.00 out of 7 features) and Scientific news/ libraries follows with a mean of 3.85 features. Both are categories that involve a lot of personal data and credibility expectations from users. App quality was evaluated with the use of the MARS tool. The average score of the apps were 3.2/5 which leaves room for improvement. The aesthetics category scored lower (2.92) and the information quality higher (3.47), but the range of scores is limited without any cluster exhibiting a satisfactory result. The Patient management category scored higher in quality (3.87)without, however, gathering high scores in the trustworthiness tiers (features).All mHealth apps in this study are offered free of charge which means that their revenue source is the online advertisements showed to professionals as they use the app and the potential income from satisfied users who are willing to pay an extra fee for unlocking the maximum capabilities of some of these apps. The feature of “ads inclusion” is present in 60% of these apps and “in-app purchases” in 42%. These features usually cause reverse usability to the targeted users. The majority of the apps (77%) have received a positive rating by their users (>4*) and less than 10% were not rated at all. The overall mean rating was 3.9/5 and the mean downloads were above 115,000 meaning that a good number of professionals were at least interested in looking for a helpful app to assist them in the diagnosis or learning process. Concluding, it is obvious that the integration of mHealth in medicine differentiates according to the category of the health apps. Professionals and students mostly download the applications with the most features that include trustworthy tiers (Connection with doctors &Scientific news/libraries).On the other hand, users’ evaluation rating is in line with the quality of the apps, since Patient Management, Games/simulators and Calculators gather the highest scores in rating stars and the MARS quality evaluation system. Based on that, it seems that mHealth innovation equally concerns all app groups (Education & Training, Decision making in diagnosis & Management), but users prefer to use certain app categories within these broadergroups.

149

Their criterion for downloading may be trustworthiness and their criterion for a good rating may be the app’s quality.

5.2 Suggestions for new mHealth apps with increased usefulness

As the search in the App Store revealed, a relevantly small sample of AI mHealth applications currently available to professionals and due to the belief that these apps will keep emerging ( Wang et al. 2018) following the advances of medical informatics the study has focused on the identification of such capabilities that could become useful and marketable mHealth apps. Based on the use of AI and ML to investigate the future agenda of such apps as there is evidence that their capabilities in the health sector are limitless (Panch et al. 2018). Surveying the relevant literature and web developers’ sites about the use of AI in mobile apps techniques the identified capabilities could be integrated in mHealth apps for becoming even more useful to professionals. Company giants such as IBM and Google have focused their efforts in this area and much of their funding goes to this direction (Dwivedi et al. 2019). Computer scientists and medical scientists must work together to develop ML methods, which can bring benefits to healthcare. These AI algorithms must then become actionable inside a mobile experience in order to obtain in the near future these useful mHealth apps. There seems to be a massive untapped market waiting to be released with the right technology. The AI capabilities of image recognition that can be embedded in a mobile app, such as Google lens, can become extremely useful in medicine. For example, a trained algorithm can perform image matching and automatic diagnoses of skin cancer, of an MRI (Zakhem et al. 2018),or drug effectiveness from a petri dish image (Agarwal et al. 2019). ML capabilities of Natural language processing (NLP) help systems learn how to interact with and understand human language after training the algorithm to match text and voice. Adapted smart reply technologies and chatbots like Alexa and Siri can be used by medical professionals and their clinics as patient support systems and communication tools for automated messages (reminders for appointments, bookings of new appointments or even sending lab test results and drug prescriptions through an authentication process. The latter requires a much more robust safeguard approach, one

150 of them being a facial recognition ML algorithm to identify one’s face from a repository of the individual faces of patients held in the clinic. Such systems are commonplace in security, surveillance, and law enforcement. Doctors can also use a chatbox app to monitor their patients’ health by scheduling questions related to their health, and through text classification and trained algorithms, the chatbox could provide appropriate responses. Such an app enhanced with analytics of social and behavioral data received by the smartphone (such as phone activity, step counter, sleep and heart rate monitor) and video call capabilities could be used for doctors to monitor and manage their patients and intervene with telemedicine when judged necessary. Again, privacy and security issues are raised for developing such apps. Another potential mHealth app that can greatly help expanding doctor’s experiences to special disease cases and evolve the practice of medicine is one named “patients like mine” (Gombar et al. 2019). Such an app could help clinicians collect reliable insights out of stored patients Electronic Health Records (EHR). Clinicians could pose questions and, through a search engine for indexing patient timelines, the system could build cohorts matching a clinical phenotype, incorporating a propensity score matching for patients similarity, survival analysis, causal inference, and a clinical interpretation of the results and their limitations (Schuler et al. 2018). Such an app could be further enhanced with AI capabilities of real-time translation in order to increase the matching EHRs, incorporating EHRs from other countries written in the local language. This can only become possible if countries share EHRs or consent to create a shared, real-time, or frequently updated, repository of anonymized EHRs. The real-time translation ML capability (such as Google Translate) can also be an additional useful tool to the “Library” type of apps to increase availability of articles and other important clinical notes written in different languages. All of these apps and many more could shape the future of mHealth for medical practitioners. Nonetheless, there are many issues related to privacy, security, and data governance that must be resolved before such apps are released in the market.

5.3 Mitigation of Privacy and Security Challenges

151

From the 168 apps, the most popular app features were “Privacy Policy Declaration” (87.5%) and “Brand Recognition” (72%), which might mean that such features are either obligatory or designers believe they are more attractive to consumers than “Registration” of users (25%), which might be considered as a drawback to install the app. Overall, “Connection with doctors”, “Patient management tool”, followed by “Games” are the mHealth types that gather the most features related to ethical concerns (privacy and reliability) indicating that users who need these kinds of apps in clinical practice have major concerns about privacy, security, and trust issues. Moreover, the artificial intelligence apps score higher in all privacy and security features. The research indicated that these apps have the potential to interpret sensitive information, not only of the users, but of their patients too and therefore users need more safeguards for privacy and reliable sources. It is also interesting that “Handbooks” and “Dictionaries”, the most popular in terms of number of apps available in App Stores, neither include enough features to ensure privacy and security, which could be justified from the nature of the apps since they are mostly used for self-training in exams, therefore they could have a supporting and temporary character, nor do they excel in reliability, which one should expect. The latter is also true for “Diagnosis assisting tools” and “Calculators. Based on certain features are explained the challenges that mHealth faces nowadays. Ethical issues that include privacy of users’ personal data are addressed by the majority of designers by including an official declaration at the bottom of the app description in the App Store, or as a tab after downloading the app, or information on the designer’s website. In the sample 87.5% of the apps included a declaration, but only 39% of these were asking for users’ consent to access their data (e.g. “Patients records and Appointments for Doctors” app) or access to users’ device storage (e.g. “Lancet” app) or agreement to both privacy policy and terms and conditions of content usage (e.g. “Common Differential Diagnosis” app). If users deny agreement, they are not allowed to continue with the app. A study of the 600 most used mHealth applications as of May 2013 has shown that only 30% of the applications had a privacy policy on that date (Benjumea 2019). Considering this statistic, nowadays, and after the wide discussion of regulations on the security of personal data, the vast majority of health apps choose or are obliged to inform the potential users about their privacy policy and some of them require consent before allowing navigation to the app itself.

152

The main challenge of privacy, security, and data governance when processing sensitive, personal health data has become increasingly important because of the rapid development of new forms of data, and the ease of transferring and sharing data. Future research is directed towards the development of systems that will standardize and secure the process of extracting private healthcare datasets for further aggregated use (Galetsi et al. 2019). On the other hand, a service provider publishes a narrative privacy policy document made by lawyers and expects that a service is either accessed as is or not used (take it-or-leave-it model). As people have limited or no power to negotiate with the service provider, they are forced to accept service provider’s rules without sufficient and reliable information. However, users of mHealth apps would most probably want to control how and by whom their personal information is used and disclosed. One solution to this problem is the deployment of multiple, formal, and therefore computer understandable policies as defined by ISO 22600 (Blobel 2017). Unfortunately, only a few service providers support personal polices, and current laws do not force them to accept user’s privacy policies (Marsden 2017). Responsible AI devices focus on designing and implementing solutions allowing stakeholders to fully understand how applications process their data and provide information that may lead them to specific decisions (Wang et al. 2020). This can be addressed following practices to provide ethical, transparent, and accountable solutions that help maintain individual trust and minimize privacy invasion ( Wang et al. 2020). It might also be possible in the next decades that privacy is not considered a personal right anymore, but as a statistical risk (Hong et al. 2004) or commodity that can be sold to data companies (Davies 1997). There is no doubt that many governments, organizations, and service providers frequently interpret privacy as relative and think that the privacy level offered can be balanced with other interests such as business gain or national security (Litt 2013). To make meaningful privacy decisions, a person should understand the impact of selected policy rules and security attributes for the reliable use of services (Ruotsalainen 2014). However, detection of malicious applications is difficult due to limited resources available within smartphone devices and an efficient and accurate mechanism is needed to detect malware so that security can be guaranteed (Mehtab et al. 2019). As health care data has become an increasingly popular target for hackers (Filkins 2016), significant measures are required to ensure security and privacy of data

153 using traditional privacy techniques. These require developing new security protocols such as device-based authentication schemes, an example being the PMSec where no data related to the Internet-of-Medical-Things (IoMT) devices are stored in server memory (Yanambaka 2019), or a new dimension of information flow for smart health data to the cloud (Puthal 2019), such as Lattice model (Jayaraman et al. 2019). Lastly, efficient detection/prevention systems in mHealth applications should employ a strong multi-factor authentication, which should offer to the designer the least privilege with the necessary permission to accomplish a specific task (Yaacoub et al. 2020). Thus, privacy and confidentiality should be core concepts in such apps. They need to guarantee the security of data that are exchanged and that the whole interaction between the system and the doctor/patient is confidential. For all mentioned applications, data processing operation must be compliant with the different data protection regulations of all countries. Developers must provide a detailed description of how user personal data is used within the service, incorporate a consent agreement for data processing upon registration for every sort of data or system used/accessed (camera, images, text, phone logs, voice, etc.), request new authorizations if the app is not used for a couple of weeks, have a limited retention of data (for a couple of days) from deleted accounts after termination(Van Der Sype & Maalej, 2014), implement cryptographic protocols [e.g. Transport Layer Security (TLS)] to provide communications security over a computer network with Advanced Encryption Standard for data transit (Sun & Upadhyaya, 2015),and implement computer vision algorithms for face recognition that run entirely on the phone avoiding sending, for example, face 3D scans (FaceID) to cloud servers (a neural network can take the FaceID scan embed it into a vector, store it locally and compare future scans to that) (Jeon et al. 2019). The protection of mHealth users’ data requires legal and policy attention. mHealth apps’ development and usage will further grow if countries share global ICT standards and architecture and health structures and public health systems cooperate to identify and use best practices for enterprise architecture that will allow both data protection while also ensuring that anonymized data are safe and in the right format to be used for research purposes. Physicians must become more educated in informatics (Fernando & Lindley, 2018) in order to more thoroughly understand AI methods, ML algorithms deployment and the datasets that they are built on. Black boxes in the process of mHealth should not be allowed.

154

On the other hand, there are cases were mHealth apps may enhance issues of confidentiality and ethics. This is true for chatbots that reduce the sharing of information with human intermediaries such as, clinic’s assistants and physicians and therefore restrain confidentiality issues between patients and professionals, as long as the chatbot conversations remain secure. Moreover, the medical center’s chatbot that helps patients answer common medical questions could increase the population’s (virtual) access to healthcare and offer new solutions that are less dependent on expensive in-patient care (Schulman and Richman 2019).

5.4 Mitigation of Reliability Challenges

While privacy and security are interconnected concepts, before beginning to use mHealth services, users also need credible and reliable information to enable them determine the trustworthiness level of services and avoid the use of blind trust (Ruotsalainen 2014). Trust could be built by empowering the engagement of all stakeholders such as patients, medical professionals, healthcare organizations, policymakers, and governments and offer them training to embrace and understand the latest technology (Horgan et al. 2019). To gain the most of the outcomes of AI devices, reassurance should be provided for the right interaction between algorithm and data as well as the way results are calculated and explained to users (Horgan et al. 2019). In this study the apps’ developer “Brand Recognition” was quite high amongst the 168 apps (72%), although the “Help function” with developers’ contact details (30%) and the “Feedback” option for asking questions to the developer and partners about the results and the calculation process (31%), had limited presence. Moreover, credible sources were identifiable in only 39% of these apps. A surprising finding was that one of the most popular apps in terms of rating and downloads of the “Handbook/manual” type [Internal Organs in 3D (Anatomy)] does not provide information sources.

The reliability/credibility of the source and transparency of the ML algorithm are also very important challenges. Amongst the capabilities a system must possess in order to support clinical decisions are the relevance of the answer and solid scientific footing (Shortliffe & Sepúlveda, 2018). There are ethical and safety concerns for automated

155 healthcare consultation (Char et al. 2018) such as “Patients like mine” which may be offered based on incomplete and biased EHRs (Zulman et al. 2016), mostly collected for the Western world, that can result in inadvertent preconception and even racial mismatching (consider the effect of skin color on the skin cancer diagnosis app). The main characteristics of an ML algorithm used for diagnosis and consultation requires transparency so users can understand the scientific foundations of the recommendations. The ML algorithm should have solid, peer-reviewed, scientific evidence establishing its validity, reliability, usability, and reproducibility (Shortliffe & Sepúlveda, 2018). This requires a systematic process for identifying predictable errors and acceptable solutions, a monitoring system with intermediate outputs to control the calculations step by step, evidence of the system’s ease of workflow integration, and the use of linguistics variables to determine computational trust. The whole process should be designed to be fail-safe and the output to be as accurate as possible while doing no harm (Shortliffe &Sepúlveda, 2018). For more complex apps, such as “Patients like mine”, in order to generate safe and valid advice it would require an “expert in the loop” to contextualize results for clinical decision-making (Schuler et al. 2018). The expert in the loop may be a whole team comprised of an app developer, a clinical informatics trained physician for interfacing with the requesting provider and giving clinical context when interpreting findings, an EHR data specialist to create patient cohorts, and a data scientist to perform statistical analyses. This setup is very different from the “popular paradigm of self-serve AI-enabled tools that undertake data processing behind the scenes and directly present the results to a physician for interpretation”(Gombar et al. 2019). The suggested system/team would require at least 48 hours to provide an answer (Schuler et al. 2018). To keep the “expert in the loop” it also requires a per query charge as opposed to a fee for service charge (paying for downloading the app) that users are more acquainted with or a SaaS (Software as a Service) model. This causes extra challenges with regards to the commercialization and sustainability of the app. Therefore, it is important for such apps to be triggered and provided with the signature of international organizations, such as the World Health Origination (WHO) or the European Centre for Disease Prevention and Control (EDCD). This will alleviate many concerns surrounding the credibility of the apps’ sources, as these organizations will most probably take the necessary measures to ensure validity and reliability of the app. They can also undertake some of the financial burden in the attempt to promote health and life-saving processes in everyday medical practices. They also have the

156 authority and power to direct countries’ collaboration on a given purpose for providing ready access to high-quality clinical data via the creation of an open, accessible, patient- centered data architecture (contrary to the proprietary technology currently in use by many health structures and physicians) (Schulman and Richman 2019). The next argument is another reason for international organizations to take part in mHealth apps development. Given the increasing importance of quality metrics reflected on physician ratings and reviews to promote their career or on clinic evaluations that determine reimbursement rates, there may be a temptation from private, not regulated, designers who create apps for clinical use to teach ML algorithms to guide patients toward clinical actions (unnecessary interventions that physicians are experts at, recommendation of tests, drugs, devices in which they hold a stake or referral patterns alteration to please patients) that could improve these metrics and profits as well as, but not necessarily, improve individual’s health. Although ethical codes have been developed by professional organizations (e.g. the ACM Ethical Code for software Engineers and IMIA’s Ethical Code for Health Information professionals), these function as inspirational drivers to computing professionals rather than regulatory guides. Such mis-happenings are very difficult to identify since the correct diagnosis in a particular case and what constitutes best practice can be controversial (Char et al. 2018). Another issue related to the liability of the ML algorithms, that are used for apps predicting certain metrics (such as the “calculators” and these that predict patient’s possible mortality rate and other metrics), is that the ML algorithms are based on existing data (time series of numerical metrics, EHRs) that reflect what happened in the past, but not necessarily depict the future and moreover may also mirror human biases in decision-making (Rajkomar et al. 2019). Greater amounts of different data such as socio-economic factors, genomics, real-time data from sensors and smartphones must be incorporated in regression models to increase the predictability of the ML algorithms (Char et al. 2018).

CHAPTER B. 6

157

6. mHealth apps: What is needed

Technological progress is critical to the global economy. Studies on new applications of information and communications technologies lead to identification of promising avenues for future research (Sheng et al. 2019). In this section it will be an attempt to answer the current and future needs of appealing mHealth apps for professionals, starting with the app type usefulness attribute and we will provide a quick view of the prospects of such apps and their benefits, not only for medical practitioners but for the society as a whole. Then we will conclude with some solutions of how mHealth apps can increase their trustworthiness and quality evaluation which will hopefully lead to greater health professionals’ engagement and app popularity. The recommendations that follow are a synthesis of research ideas and findings from the recent relevant literature as well as from online sites of mobile app developers.

6.1 Building Useful and Trustworthy, Quality mHealth apps

The research findings have shown that decision-making and patient management apps score higher in quality. Therefore, further improvement of these apps will greatly appeal to health professionals. Differential diagnosis apps can be enhanced with image recognition technologies which will be increasingly useful for e.g. dermatologists, ophthalmologists and microbiologists who can take a picture of the patients’ skin, eye or petri dish and the app will match this image with the built-in library of images to provide a diagnosis based on the similarity percentage found between the taken image and the diagnosed images. This capability could prove life-saving to patients around the world where there is a lack of qualified health professionals and equipment, such as poor countries and war zones (Hu et al. 2019). Another important suggestion, which can greatly help expanding doctors’ experiences to specific disease cases so that they can provide more targeted therapeutic schemes to patients and also evolve the practice of medicine altogether, is the potential differential diagnosis app “patients like mine” (Gombar et al. 2019). Such an app could help clinicians collect reliable insights out of stored patients Electronic Health Records (EHR) and acquire medical domain expertise to contextualize results for clinical decision making at the point of care.

158

This app could also incorporate real-time translation for expanding the pool of similar EHRs from worldwide, including EHRs that are written in other languages. The real-time translation capability can also be a useful additional tool to the “Education & Training” group of apps to increase availability of articles, handbooks and other important clinical notes written in different languages and make them available to a wider population. Except virtual treatment and tracking of incidents, literature has reported many other outcomes of professional mHealth applications in the case of a general lockdown such as education, training and counselling using webinars and online lectures and “tertiary referral” to avoid in person visits (Iyengar et al. 2020). More importantly, using applications that are able to analyze big data, store vast amounts of patient information globally is very much suitable to track and control the worldwide spread of a pandemic, such as the disease COVID 19 (Javaid et al. 2020). Moreover, reporting of infectious diseases by doctors, clinics and laboratories to governmental agencies could lead to faster recognition of cases of infectious disease. More direct access to such data could enable surveillance epidemiologists to detect potential public health threats such as early-level warnings for epidemics (Velasco et al. 2014). With regards to the “Management” of patients app category, a convenient app for medical specialists and their clinics could be an app that uses smart reply technologies to enhance patient support and communication (replies to standard queries, such as clinic’s working hours but also replies to more advanced queries such as reminders for appointments, results’ collection or even sending online drug prescriptions with some authentication process and first aid answers to specific cases, e.g. snake bite). Like this, smart technologies can take over the majority of routine conversations and tasks to manage patients and patients’ records and could also improve the population’s (virtual) access to healthcare and provide new resolutions less reliant on expensive care (Schulman and Richman 2019).Doctors can also use smartphone AI applications (Vaishya et al. 2020) to track heart-rate and temperature to accurate model pandemic trends (Kapoor et al. 2020). These apps and many more are the future of mHealth for health professionals offering a helping hand to the whole society in terms of provision of healthcare. The challenge of integrating ethical decision making into technical development settings can advance the dialogue between work practices and ethical discussions across healthcare mobile development and encourage developers to prioritize privacy practices and

159 features in software design (Shilton and Greene 2019). Therefore, many problems with regards to trustworthiness of such apps must be resolved before they can be launched in the market. For example, security, reliability and usability issues are raised for automated healthcare consultations (Char et al. 2018) offered by differential diagnosis apps. With regards to security, app results may be calculated based on biased and incomplete EHRs (Zulman et al. 2016), available mostly from the developed countries where EHR information technology is implemented. Such results may lead to inadvertent preconception and perhaps racial mismatching or non-usable apps (e.g. the skin color validity on the skin cancer differential diagnosis app). Moreover, a fundamental prerequisite for the adoption of such technologies in diagnosis is reliability and transparency, as also shown in the findings of the study, which means to make known and comprehensible the scientific foundations of the solutions to the users. In increased complexity apps, such as “Patients like mine”, for the app to provide safe recommendation it would need a “health informatics specialist in the loop” to digest the automatically generated results and feed the app with valid clinical decision-making options (Schuler et al. 2018).Thus, such apps are best if they are provided by international or national health organizations (e.g. World Health Origination, American Society of Clinical Oncology), which can offer increased credibility about the app’s info and results, as they have the resources and expertise to ensure mHealth app validity and the financials to budget it for offering advanced care services in the daily clinical practice and democratizing innovation (Bergvall‐Kåreborn & Howcroft, 2014). Moreover, apps that are used for patient management and automated communication with patients, which may also involve official document exchange (as proposed above),would require an increased level of security measures, such as facial and other recognition systems, that are commonly used in electronic surveillance. In all mHealth apps, IT security features must be increased in order to develop a trusted environment for its users with regards to their personal data exploitation and the authentication process. Moreover, the regulators should enforce by law essential rules for warranting the fidelity of the mHealth apps. On the other hand, the responsibility is also on the users’ side who should be extra cautious when using such apps because of the associated security and privacy risks created by the generation of new information (Angst 2009).Education on the matter will make things better and will increase the demanding voices for trustworthy mHealth apps.

160

Small improvements in usability can be identified in all app categories, for example, the Connection with doctors’ apps could offer increased features’ customization with regards to user profile, everyday clinical requirements and prognosis needs. Nevertheless, it is also very important that mHealth apps should avoid the business model of free or freemium apps with income generation from ads and promotional material, and focus on the engagement of national health organizations, associations and academic institutions to finance or provide themselves the app and its essential updates, or else move to a business model which offers a free trial period and then charges a fee for its further use. In the latter case the period should be long enough (e.g. 1 month) for the health professionals to understand the usefulness and usability of the app and get involved enough to be persuaded to pay a small fee for keeping the app functional. This model will also trigger app providers to offer more appealing and quality apps in order to engage health professionals. With this addition, the MARS scores could be improved across all functions. The goal is to increase mHealth apps’ usefulness with the use of innovative technological capabilities to cover identified needs, starting from areas with limited health resources and limited access to expensive biotechnology (developing world, inaccessible geographic areas) but also in everyday professionals’ routine as a supporting tool so that the app’s recommendations are judged by experts and not taking for granted automated decisions. Simultaneously, innovation in healthcare technology must outline the four ethical principles for healthcare managers: respect for others by maintaining confidentiality, beneficiality by providing aid to patients, non maleficence by not putting at risk patients during treatment, and justice by equally providing treatment to all patients without taking under consideration gender, ethnicity or social status (Oddo 2001). In this Thesis there is a presentation and discussion of the benefits and challenges of mHealth applications. Frameworks and taxonomies were presented to ensure a better understanding of the adoption process of such an innovation technology. mHealth apps should maintain a high level of security, reliability and usability by incorporating features that ensure these elements. Moreover, the development of new mHealth apps, which will increasingly utilize the capabilities of new technologies, will be of great help to medical professionals, the health sector and the society overall. Some of these apps, such as “Patients like mine”, under a certain spectrum may be capable of

161 evolving medical science as a whole. It is the next challenge for IT developers to create such apps in the most responsible way with the support of the health world.

6.2 Managerial Implications

The research also carries managerial implications for app developers, health professionals and the health system. The increasing in size mobile app industry can find valuable implications for the future direction of health e–commerce (Ghose, & Han, 2014) and in particular the features that customer value the most in the mHealth app market .This study revealed that features relevant to usefulness (app content/capabilities), in contrast to features relevant to reliability and usability, affect mHealth apps’ downloads. Health professionals’ intentions to use mobile health apps during their daily clinical activities follow certain dimensions and app developers should consider that consumers’ behavior is related to their need to gain assistance while performing their work. Therefore, this study shows that health professionals are mostly attracted by the sought result/info which derives from the app in order to increase their job performance. The comments and rating of other colleagues that have already used the application and the quality of the app, including engagement, functionality, aesthetics and information quality are also important factors in their decision to download the app. The length of the description is also crucial and includes the needed explainability that a heath professional expects in order to use an app.

The 21st century is a technology era, information age, network trade era and artificial intelligence era. In this rapidly evolving environment, management and administration practices need to be quickly adjusted and innovations must adopted in a continuous mode (Chong & Sheng-bin 2006).mHealth is considered one of the most transformative drivers for healthcare delivery innovations and has brought to the clinical research environment revolutionary insights with new challenges and opportunities (Cleary 2018). Based on a leading mobile analytics company, 90% of the time that we spend on mobile devices is spent on mobile apps where statistics show a constant increase in global app downloads per year. For the years 2016-2019 this increase was at the level of 7%-27% (© Statista 2020). This is encouraging news for app developers. In the market of mHealth apps for medical professionals, The research has shown that there is still a lot of space for improvement. A range of apps still remain untapped relevant to medical professionals that would be of use and could promote their medical practice and medicine as a whole. Such future apps may utilize artificial intelligence at a greater level.

But in order to be successful in this space, two things need to happen. First, the designers of health apps need to make improvements in the credibility of their app’s information and the visibility of this credibility, as well as IT security features must increase its ability to create a trustworthy environment for its users. Second, the regulators should also take legislative action to enforce minimum rules for ensuring the trustworthiness of the mHealth apps’ market. Moreover, our

162 society demands responsible consumers of IT systems. Education should play an important role in this area. Towards this direction tools have been created for users to assess the credibility of an app, such as the CRAAP Assessment of Health Promotion Mobile Applications (Apps) (McNiel and McArthur 2016).

Finally, governments must consider the large-scale implementation of successful and useful mHealth innovations. For example, “Tools for patient management’ and “Connection with doctors” are the two types of apps that could have a tremendous effect in the healthcare at busy health centers in the first instance and of geographically disperse areas in the other. The World Health Organization and other stakeholders will need to issue guidance to help prioritize and accelerate governmental mHealth adoption (Labrique et al. 2013).

Therefore, technology firms can become wiser about customer engagement, download intention behavior and brand loyalty by performing customer analytics with regards to the app capabilities they use (e.g. demographic characteristic of health professionals who download apps with specific purpose/capabilities). Based on this information they can optimize app content for specific user segments and length of description for better navigation, since this feature was also found to be important to users. This reveals that increased app downloads depend on the app publishers’ behavior since developers can manipulate app features and app descriptions. Moreover, a similar approach can be followed with regards to star rating and reviews. Since rating and reviews of other users are crucial for app downloading, app developers should use customer analytics (e.g. text analytics to user reviews) to understand customer expectations, likes and dislikes. Based on this information, they can improve app content, features, etc., for offering a more appealing app and offer the necessary bug fixes as identified in the user reviews. Developers can also increase the number of reviews by responding to each review which a thank you note, or an explanation note in case of complaints as the study has shown the number of reviews highly related to the number of downloads.

Nonetheless, developers must be cautious with the finding that more capabilities (usefulness) lead to more downloads and should find a balance between increased capabilities and low complexity, as other research has shown that complex functionalities tend to confuse users in decision-making (Alnsour et al. 2016) and decrease the engagement of health professionals to the new technology (Venkatesh et al. 2002; Zapata et al. 2015; P. Llorens-Vernet, & Miró, J. 2020). Meaning that although the app is downloaded it is used only for a limited time period and then it is uninstalled. .

CHAPTER B. 7

163

7. Discussion and Conclusions

These days, smartphones are more than accessible and provide unique processing powers that allow for complex apps to run, such as those providing health services (Balapour et al 2019). Medicine has been profoundly affected by the availability of mobile devices (smartphones and tablets) in clinical practice and education due to fast access to information and better communication and information resources at the point of care. Given the increased adoption of smartphones by healthcare professionals (Cleary 2018; Mosa 2012) the second part of the PhD Thesis provided three separated analysis based on a unique sample of 168 health applications addressed to health professionals and students. In the first analysis there is a description of the apps of the sample and an explanation of their features, especially the ones that can be explicable in terms of ethical concerns in privacy, security and reliability. An expanded version of the Communication Privacy Management theory has been followed to explain privacy and reliability features that developers should incorporate in the apps, so that the boundaries (criteria) of data owners (health professionals) to trust an mHealth app are not shaken. Moreover, based on the content of each app, there is also a classification of the apps into certain categories denoting their usefulness, spotted the apps that included artificial intelligence and emphasized their outcomes. Technology studies emphasize the way technology acts on the world as well as that its deployment creates socio-ethical implications that reshapes professional relationships, creates moral consequences about issues such as privacy and trust, reinforces or undercuts ethical principles and enables or diminishes stakeholder rights and dignity (Martin et al. 2019).The contextualization of technology and the way is used in the health sector reveals that the information systems innovators will need to consider the conflicts and challenges it creates (Flick et al. 2020).Technological innovations in medicine have allowed the use of mobile devices in clinical practice and education due to fast access to information and better communication at the point of care. Following the Normalization Process Theory (May, 2013), the second analysis of the study, explained step by step the path to the adoption of an innovative health application in an everyday clinical routine measured every stage of this path by using relevant scales. Namely, there has been used mHealth apps’ categories (education, decision-making, patient management) at the implementation

164 stage, the app features organized in three tiers (security, reliability, usability)for the embedding stage, and finally the MARS quality score, stars rating and number of downloads as the dimensions of integration of the NPT model. Observing the significant associations among the research framework elements, it is obvious that the integration dimensions of mHealth app downloads, their star rating and their quality score are all highly intercorrelated with each other. Furthermore, from the integration dimensions of the mHealth apps, the quality score is correlated to the reliability tier of the embedding stage of the conceptual model that incorporate the features of credible source and the feedback communication feature that users can benefit from while using the app. The social sharing feature is positively associated with the existence of the feedback mechanism and ads, as marketeers most probably find this a useful tactic to increase ads exposure. Privacy is positively correlated to the in-app purchases and ads indicating that developers include a privacy policy statement when they try to earn from the release of their app by acknowledging the promotional material. There is also evidence that the more features in an app the better its quality score, especially when credible source, feedback opportunities and guidance on how to use the app are present. Apps belonging to the categories of decision-making for prognosis and especially patient management score higher in quality and incorporate more features overall compared to the educational category apps. Finally, the last part of the second research of the PhD Thesis, attempted to understand the relative importance of the determinants of the professionals’ consumer preference for the health mobile apps. Based on the literature review, it has been developed a research model applying the combination of two theoretical models related to the acceptance of innovation, namely the Technology Acceptance Model and the Theory of Reasoned Action in order to explain the integration intention of the innovation of health apps in clinical routine. This study reviewed a number (n = 168) of health apps addressed to professionals available in the Play Store and is the first study to have quantitatively modeled the health mobile apps’ use for professionals. The main goal was to understand consumer behavior about digital health and to specify the variables that attract health stakeholders in downloading related apps. There is a wide literature with regards to the adoption of mobile apps in health using questionnaires to targeted populations (Jahn et al. 2019; Holdener et al. 2020; Byambasuren et al. 2019)or statistical analysis (Krishnan and Selvam 2019) to investigate and predict the factors that influence app downloads in the healthcare field.

165

Literature has also investigated the importance of mobile apps in specific clinical routines by health professionals (Ventola 2014) or the effectiveness of mobile health (mHealth) technologies to train healthcare professionals (O'Donovan et al. 2015) conducting systematic reviews (Gagnon et al. 2016) or surveys (Hofer & Haluza, 2019). All the above-mentioned literature has indicated the usefulness, usability and trust, among others, as the most important moderators affecting users’ intention to download and use mobile applications. However, this analysis gathered all possible determinants that proved to be influential for app downloading and statistically tested whether these are important at a job relevant population such as health professionals using features that are distinctive and easily recognizable in the search for a useful app. This study has also revealed some interesting results controversial to past evidence. For example the statistical analysis of the sample, using the moderators based on the TAM model of perceived usefulness, perceived ease of use and trust, has clearly indicated that health professionals intend to download apps that fulfil their temporary or long term need to complete a work task (usefulness) without taking under consideration the ease of use and trust features. Therefore, app features that define the perceived ease of use (usability) and perceived trust do not significantly affect the health professionals’ decision to download the app. However, Gagnon et al (2016) in their systematic review found evidence in ten published surveys that ease of use is an important factor of using mobile technology in the working environment. Moreover, in Austria, a survey on 151 doctors recognized the usability elements are among the most important features in order to use a new app (Hofer & Haluza, 2019). Nonetheless, although health professionals, when participating in surveys, state that usability is important, when it comes to download the app, this study has shown, that ease of use is not the factor that will influence their decision. Similarly, the same conclusion can be made about trust. Privacy, security and brand recognition have been indicated as very important elements of behavior intention to engage with mobile technology ( Lee, & Raghu, 2014; Fuller et al. 2017; Goldhahn 2018), nonetheless, this study has proved that app downloads are not related to trust, although a positive relationship would have been expected especially with regards to the trust of mHealth app content. On the other hand, with regard to the perceived usefulness of TAM ( Davis 1989) and influence moderators of TRA that determine the subjective norms of an individual’s behavior intentions (Moore & Benbasat, 1996),the study has identified that

166 app’s quality, social influence and app’s specific features which promote usefulness directly influence consumer behavior. In particular, the research model identified that app downloads, which reflects the dimension of consumer preference, are positively significant to the app quality, the users’ evaluation and the features that relate to increased operational capabilities for medical professionals. Moreover, it indicated that downloads are positively related to the number of reviews, the length of the app description, the existence of in-app ads and the years since app first release. Therefore, based on the results of this study, mHealth apps which receive a higher rating from the users, lead to a greater number of downloads. Similarly, the quality of apps, which in our case is represented by an evaluation rating scale called MARS, is also positively significant to the mHealth apps downloads from professionals and therefore the quality of an app could motivate professionals to use it. These findings are aligned with the findings of many other relevant studies about health mobile apps (Chatterjee et al. 2009; A. Ghose, & Han, 2014; Pereira-Azevedo et al. 2016; Salazar et al. 2018; Krishnan and Selvam 2019). The results also indicate that app downloads are positively related to the number of reviews from other users, consistent to the results of Kübler et al (2018) and Krishnan & Selvam (2019). The time since first app release has also a positive significant coefficient indicating that apps with longer presence in the market have more downloads, which is generally reasonable. Moreover, the length of the app description also has a positive effect which indicates that the more explanatory the description of an app in the app store is, the more downloads is expected to have. All the above results has also been indicated from previous literature ( Ghose, & Han, 2014). Obscure is, however, the result that in-app ads are related to downloads although one would expect that in-app ads would deter the professionals from using these apps, as it has been proved in Ghose & Han (2016). However, it is true that professionals would come to know about the existence of in-app ads only after they download it, although some app descriptions in the Play Store do mention it. During the empirical analysis, and from testing the relationship of any two conceptual framework variables and app features, a few additional insights have been gained. These are mostly related with the following: the existence of the feedback feature and the offline use feature increase the perceived quality of the app, and apps of bigger size (in MBs) gather more reviews. No other research has identified these relationships.

167

Conclusively, the explosion of mHealth and the wide adoption of mobile apps in medicine practice and education have created an innovative and vibrant mobile ecosystem. As health professionals tend to use mobile devices in their daily clinical routine, it becomes important for app developers to understand the factors that influence user demand for mobile apps (Ghose, & Han, S. P. 2014). This three stage analysis has also attempted to understand how consumers, and more specifically health professionals interact during the decision of the adoption of a mobile app for assisting them in practicing medicine, giving insights to developers for the success of their brand strategy but nonetheless for the launch of useful mHealth apps that can promote healthcare

CHAPTER B. 8

8. Limitations and Future Research

This research is not without limitations. Due to the fact that a mobile app can be regarded as an innovative product recently introduced in the market, it is likely that personal innovativeness would influence the intention to adopt a healthcare device (Lee & Lee 2018). In the present study there is an examination of the concept of the integration of technology innovation in a clinical context and the associated dynamics, providing a framework for mapping important elements (Murray et al. 2010). However, there is no prediction of the engagement rate of health professionals to the use these apps in their everyday routine for a long period, meaning that an app can be downloaded, used only for a limited period and then been uninstalled, if proved not useful/acceptable. Therefore, future research can investigate through a concept centric survey targeted to health professionals, with panel data, whether the identified features’ correlations can influence the long-term adoption of mHealth apps in clinical diagnosis and increase the users’ engagement rate, also taking into consideration behavior intention. The findings of this research form the basis for an extended future investigation of this hot topic with the use of various methodologies.

Future research could also investigate through surveys and questionnaires whether app features that promote trust are positive related to mHealth app adoption by professionals. With the assumption that professional apps will probably require a particular trust level for their final adoption, many issues concerning the trustworthiness must be resolved before their introduction in the market and especially with regards to their credible source which is currently present in less

168 than 40% of the examined mHealth apps. For instance, security, reliability and usability issues are raised for automated healthcare consultation (Char et al. 2018) offered by differential diagnosis apps. For example, app results may be calculated based on biased information or incomplete Electronic Health Records (Zulman et al. 2016) because of developer’s ignorance or profit intention. Such results may lead to unintentional prejudice or non-usable apps and future research should examine this assumption. Moreover, all mHealth apps IT security measures should be enforced to create a reliable environment for treating users’ personal information. Important here is the role of the legislation, as well as the education of health professionals, for demanding and using only trustworthy mHealth apps.

Advertisements on the other hand, and their existence while downloading or using the mobile is a debated matter because most of the times are offered instead of a fee for accessing the app. Therefore, surprisingly this study, does not indicate negative relationship between downloads and ads, probably because health professionals understand that this is the developer’s funding source who offers it to them for free or they see the ads as a source of information about new products on their field or they are not aware of their existence before the download. A suggestion on this subject could be that instead of the business model of free or freemium apps with revenue collection from ads and promotional material, which is currently present in 40-50% of the examined apps, to promote the engagement of national health organizations, associations and academic institutions to finance or create the app themselves and its essential updates. Else developers can move to a structure which offers a free trial period for the professional to appreciate the usefulness and usability of the app and then charge a fee for its further use. This model will also prompt app providers to offer more attractive and quality apps, so as to engage the health professionals. Following this, the MARS score, which is currently below satisfactory level (average score of examined apps at 3.9/5), would also increase for all clusters. Future research could focus on these matters to test whether they hold true.

Moreover, integration of mHealth apps into the healthcare system may slowly evolve over time. mHealth apps can be prescribed by doctors and recommended by hospitals or health websites (Research 2 Guidance 2016).All stakeholders must cooperate in finding the right ways to increase mHealth apps trustworthiness and quality which will hopefully lead to a greater health professionals’ engagement and app popularity for actual use in their everyday medical practice. Future research is required to evaluate these assumptions.

CONCLUSIONS OF THE TWO PARTS OF THE Ph.D THESIS

Summarizing the scope and the findings of the Ph.D Thesis, this research consists two parts: the first part, in which there is a systematic review of the literature in the field of BDA and the second part where there is an investigation of technology

169 innovation in healthcare concentrating in mobile applications addressed to health professionals. For that reason, a sampling literature review in the domain has been conducted and 804 papers published for the years 2000-2016 have been identified and systematically reviewed using content analysis to provide an in depth description of BDA elements in health concerning the innovative decision-support systems in the field. With reference to the resource-based view theory this Doctoral Thesis has focused on how big data resources are utilised to create organizational and social values and capabilities, discussing the classification of big data types related to healthcare, the associate analysis techniques, the created values for the stakeholders and society, the platforms and tools for handling big health data and future aspects in the field. It also reveals that one of the main values created is the development of analytical techniques which provides personalized health services to users and supports human decision- making using automated algorithms, challenging the power issues in the doctor-patient relationship and creating new working conditions. A main challenge to data analytics is data management and security when processing large volumes of sensitive, personal health data. Future research is directed towards the development of systems that will standardize and secure the process of extracting private healthcare datasets from relevant organizations. The results of the study have been presented under a number of pragmatic examples to show how the advances in healthcare were made possible. The second Part of the Doctoral Thesis reviewed available mHealth apps addressed to medical professionals and students for the diagnosis process and explore ethical challenges related to their data governance and reliability. It also investigated the specific apps’ market which is addressed to medical professionals and students and explained how the specific consumers’ behaviour is affected by certain app characteristics and attributes. Content analysis has been conducted to the descriptions, functions and user’s reviews for 168 smartphone apps which were classified based on their type for diagnosis and features with a special discussion on artificial intelligence applications. Moderate levels of trustworthiness and quality are observed for the existing mHealth apps revealing margins of potential improvement. Development of apps from credible sources can increase their trustworthiness and further technological advances embedded in these apps can increase their usefulness in order to improve mHealth apps’ quality for the benefit of health professionals in their everyday practice as well as for improved provision of healthcare for society. An empirical analysis has

170 also been conducted to test the existing relationships between consumer intention and quality evaluation of health mobile applications. Therefore, based on the results of this study, mHealth apps which receive a higher rating from the users, higher quality evaluation and more reviews than others lead to a greater number of downloads staying in the line with the findings of many other relevant studies. The more explanatory the description of an app in the app store is, the more downloads are expected to have. Finally, the findings of both reviews are stimulating and provide valuable information to practitioners, policy makers and researchers while presenting them with certain paths for future research and implications. They also prove that Health analytics and mHealth provide a unique opportunity for advancing health information research and medical decision-making. New tools in problem-solving are offering new avenues in prognosis and diagnosis of diseases and can benefit health professionals, managers and researchers for the forthcoming years that medicine and humanity are facing new challenges that require fast solutions.

REFERENCES

Abbas, A., Ali, M., Khan, M. U. S., & Khan, S. U. (2016). Personalized healthcare cloud services for disease risk assessment and wellness management using social media. Pervasive and Mobile Computing, 28, 81-99.

Abbas, A., Bilal, K., Zhang, L., & Khan, S. U. (2015). A cloud based health insurance plan recommendation system: A user centered approach. Future Generation Computer Systems, 43, 99-109.

AbdelRahman, S. E., Zhang, M., Bray, B. E., & Kawamoto, K. (2014). A three-step approach for the derivation and validation of high-performing predictive models using an operational dataset: congestive heart failure readmission case study. BMC medical informatics and decision making, 14(1), 41.

Agag, G. (2019). E-commerce ethics and its impact on buyer repurchase intentions and loyalty: An empirical study of small and medium Egyptian businesses. Journal of Business Ethics, 154(2), 389-410.

Agarwal, R. & Dhar, V. (2014). Editorial—Big Data, Data Science, and Analytics: The Opportunity and Challenge for IS Research. Information Systems Research 25(3):443- 448

Agarwal, I., Kolakaluri, R., Dorin, M., & Chong, M. (2019) Tensor Flow for Doctors. In Annual International Symposium on Information Management and Big Data. pp. 76- 88, Springer

171

Aghaei Chadegani, A., Salehi, H., Yunus, M., Farhadi, H., Fooladi, M., Farhadi, M., & Ale Ebrahim, N. (2013). A comparison between two main academic literature collections: Web of Science and Scopus databases. Asian Social Science, 9(5), 18-26.

Ajorlou, S., Shams, I., & Yang, K. (2015). An analytics approach to designing patient centered medical homes. Health care management science, 18(1), 3-18.

Akter, S., & Wamba, S. F. (2016). Big data analytics in E-commerce: a systematic review and agenda for future research. Electronic Markets, 26(2), 173-194.

Al-Shaqi, R., Mourshed, M. & Rezgui Y. (2016). Progress in ambient assisted systems for independent living by the elderly. Springerplus, 5: 624.

Ali, M. A., Ahsan, Z., Amin, M., Latif, S., Ayyaz, A., & Ayyaz, M. N. (2016). ID- Viewer: a visual analytics architecture for infectious diseases surveillance and response management in Pakistan. Public health, 134, 72-85.

Ali, O., Shrestha, A., Soar, J., & Wamba, S. F. (2018). Cloud computing-enabled healthcare opportunities, issues, and applications: A systematic review. International Journal of Information Management, 43, 146-158.

Allen, G. I., Amoroso, N., Anghel, C., Balagurusamy, V., Bare, C. J., Beaton, D., ... & Caberlotto, L. (2016). Crowdsourced estimation of cognitive decline and resilience in Alzheimer's disease. Alzheimer's & Dementia, 12(6), 645-653.

Alnsour, Y., Hazarika, B., & Khuntia, J. (2016). Health Apps’ Functionalities, Effectiveness, and Evaluation. Paper presented at the Workshop on E-Business.

Alotaibi, S. R. (2020). Applications of Artificial Intelligence and Big Data Analytics in m-Health: A Healthcare System Perspective. Journal of Healthcare Engineering, 2020.

Althebyan, Q., Yaseen, Q., Jararweh, Y., & Al-Ayyoub, M. (2016). Cloud support for large scale e-healthcare systems. Annals of Telecommunications, 71(9-10), 503-515.

Alyass, A., Turcotte, M., & Meyre, D. (2015). From big data analysis to personalized medicine for all: challenges and opportunities. BMC medical genomics, 8(1), 33.

Amankwah-Amoah, J. (2016). Emerging economies, emerging challenges: Mobilising and capturing value from big data. Technological Forecasting and Social Change, 110, 167-174.

Ameel, M., Leino, H., Kontio, R., van Achterberg, T., & Junttila, K. (2020). Using the Nursing Interventions Classification to identify nursing interventions in free‐text nursing documentation in adult psychiatric outpatient care setting. Journal of clinical nursing, 29 (17-18), 3435

Andreu-Perez, J., Poon, C. C., Merrifield, R. D., Wong, S. T., & Yang, G. Z. (2015). Big data for health. IEEE J Biomed Health Inform, 19(4), 1193-1208.

172

Angelelli, P., Oeltze, S., Turkay, C., Haasz, J., Hodneland, E., Lundervold, A., ... & Hauser, H. (2014). Interactive visual analysis of heterogeneous cohort study data. IEEE computer graphics and applications, (1), 1-1.

Angst, C. M. (2009). Protect my privacy or support the common-good? Ethical questions about electronic health information exchanges. Journal of Business Ethics, 90(2), 169-178.

Angulo, D. A., Schneider, C., Oliver, J. H., Charpak, N., & Hernandez, J. T. (2016). A multi-facetted visual analytics tool for exploratory analysis of human brain and function datasets. Frontiers in Neuroinformatics, 10, 36.

Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., et al. (2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion, 58, 82-115.- 3444

Asghar, I., Cang, S., & Yu, H. (2017). Assistive technology for people with dementia: an overview and bibliometric study. Health Information & Libraries Journal, 34(1), 5- 19.

Azadmanjir, Z., Torabi, M., Safdari, R., Bayat, M., & Golmahi, F. (2015). A map for clinical laboratories management indicators in the intelligent dashboard. Acta Informatica Medica, 23(4), 210.

Azmak, O. et al., (2015).Using Big Data to Understand the Human Condition: The Kavli HUMAN Project. Big Data, 3(3), 173-188. Backman, E., Granlund, M., & Karlsson, A. K. (2020). Documentation of everyday life and health care following gastrostomy tube placement in children: a content analysis of medical records. Disability and rehabilitation, 42(19), 2747-2757.

Balapour, A., Reychav, I., Sabherwal, R., & Azuri, J. (2019). Mobile technology identity and self-efficacy: Implications for the adoption of clinically supported mobile health apps. International Journal of Information Management, 49, 58-68).

Banaee, H., & Loutfi, A. (2015). Data-Driven Rule Mining and Representation of Temporal Patterns in Physiological Sensor Data. IEEE J. Biomedical and Health Informatics, 19(5), 1557-1566.

Banos, O., Amin, M. B., Khan, W. A., Afzal, M., Hussain, M., Kang, B. H., & Lee, S. (2016). The Mining Minds digital health and wellness framework. Biomedical Engineering Online, 15(1), 76.

Bardhan, I., Oh, J. Zheng, Z. & Kirksey, K. (2015). Predictive Analytics for Readmission of Patients with Congestive Heart Failure, Information Systems Research, 26 (1), 19-39.

Bardus, M., van Beurden, S. B., Smith, J. R., & Abraham, C. (2016). A review and content analysis of engagement, functionality, aesthetics, information quality, and

173 change techniques in the most popular commercial apps for weight management. Int J Behav Nutr Phys Act, 13, 35. doi:10.1186/s12966-016-0359-9

Barkley, R. Greenapple, R. & J. Whang. (2013). Actionable Data Analytics in Oncology: Are We There Yet? Cancer Center Business Development Group, 93-96. Barney, J. B. (1991). Firm Resources and Sustained Competitive Advantage, Journal of Management, 17, 99-120.

Baro, E., Degoul, S., Beuscart, R., & Chazard, E. (2015). Toward a literature-driven definition of big data in healthcare. BioMed research international, 2015.

Barrett, J. S., M. Narayan, M., K. Vijayakumar, K. & S. Vijayakumar, S. (2008). Integration of modeling and simulation into hospital-based decision support systems guiding pediatric pharmacotherapy. BMC Med Information Decision Making, 8:6.

Basole, R. C., Braunstein, M. L., Kumar, V., Park, H., Kahng, M., Chau, D. H., ... & Lesnick, B. (2015). Understanding variations in pediatric asthma care processes in the emergency department using visual analytics. Journal of the American Medical Informatics Association, 22(2), 318-323.

Batarseh, F. A., & Latif, E. A. (2016). Assessing the quality of service using Big Data analytics: With application to healthcare. Big Data Research, 4, 13-24.

Bates, D. W., Saria, S., Ohno-Machado, L., Shah, A., & Escobar, G. (2014). Big data in health care: using analytics to identify and manage high-risk and high-cost patients. Health Affairs, 33(7), 1123-1131. Baum, D. (2010). An intelligent patient focus. Cambridge Memorial Hospital is increasing efficiency and improving patient care with a new emergency room tracking board and business-intelligence system. Health management technology, 31(4), 12-4.

Belle, A., Thiagarajan, R., Soroushmehr, S. M., Navidi, F., Beard, D. A., & Najarian, K. (2015). Big data analytics in healthcare. BioMed research international, 2015.

Bello-Orgaz, G., Hernandez-Castro, J., & Camacho, D. (2017). Detecting discussion communities on vaccination in twitter. Future Generation Computer Systems, 66, 125- 136.

Ben-Ari Fuchs, S., Lieder, I., Stelzer, G., Mazor, Y., Buzhor, E., Kaplan, S., ... & Kohn, A. (2016). GeneAnalytics: an integrative gene set analysis tool for next generation sequencing, RNAseq and microarray data. Omics: a Journal of Integrative Biology, 20(3), 139-151.

Benharref, A., Serhani, M. A. & Al Ramzana, N. (2014). Closing the loop from continuous M-health monitoring to fuzzy logic-based optimized recommendations, IEEE (2014), 2698- 2701.

Benjumea, J., Dorronzoro, E., Ropero, J., Rivera-Romero, O., & Carrasco, A. (2019). Privacy in Mobile Health Applications for Breast Cancer Patients. In 2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS) IEEE, pp. 634- 639.

174

Beratarrechea, A., Diez-Canseco, F., Irazola, V., Miranda, J., Ramirez-Zea, M., & Rubinstein, A. (2016). Use of m-health technology for preventive interventions to tackle cardiometabolic conditions and other non-communicable diseases in Latin America- challenges and opportunities. Progress in cardiovascular diseases, 58(6), 661-673.

Berger, M. L., & Doban, V. (2014). Big data, advanced analytics and the future of comparative effectiveness research. Journal of Comparative Effectiveness Research, 3(2), 167-176.

Bergvall‐Kåreborn, B., & Howcroft, D. (2014). Persistent problems and practices in information systems development: a study of mobile applications development and distribution. Information Systems Journal, 24(5), 425-444.

Bertsimas, D., O’Hair, A., Relyea, S., & Silberholz, J. (2016). An analytics approach to designing combination chemotherapy regimens for cancer. Management Science, 62(5), 1511-1531.

Beyan, T. & Yeşim, A. S. (2014). Emerging Technologies in Health Information Systems: Genomics Driven Wellness Tracking and Management System (GO-WELL), 546:315-339

Bharadwaj, A. S. (2000). A Resource-based view Perspective on information Technology capability and firm performance: An empirical investigation, MIS Quarterly, 24(1), 169-196.

Bhattacharya, I., Ramachandran, A., & Jha, B. K. (2012). Healthcare Data Analytics on the Cloud. Online Journal of Health and Allied Sciences, 11(1 (1)).

Bhuvaneshwar, K., Belouali, A., Singh, V., Johnson, R. M., Song, L., Alaoui, A., ... & Madhavan, S. (2016). G-DOC Plus–an integrative bioinformatics platform for precision medicine. BMC bioinformatics, 17(1), 193.

Biviji, R., Vest, J. R., Dixon, B. E., Cullen, T., &Harle, C. A. (2020). Factors related to user ratings and user downloads of mobile apps for maternal and infant health: Cross- sectional study. JMIR mHealth and uHealth, 8(1), e15663.

Bjarnadóttir, M. V., Malik, S., Onukwugha, E., Gooden, T., & Plaisant, C. (2016). Understanding adherence and prescription patterns using large-scale claims data. PharmacoEconomics, 34(2), 169-179.

Blake, H. (2008). Innovation in practice: mobile phone technology in patient care. British journal of community nursing, 13(4), 160-165.

Blakely, T., Atkinson, J., Kvizhinadze, G., Nghiem, N., McLeod, H., Davies, A., & Wilson, N. (2015). Updated New Zealand health system cost estimates from health events by sex, age and proximity to death: further improvements in the age of ‘big data’. NZ Med J, 128(1422), 13-23.

175

Blobel, B., Lopez, D. M. & Gonzalez, C. (2016). Patient privacy and security concerns on big data for personalized medicine, Health and Technology, 6(1), 75-81.

Blobel, B. (2017). Standardization for Mastering Healthcare Transformation– Challenges and Solutions. , 13(1), 9-15. European Journal for Biomedical Informatics, 13(1), 9-15.

Bodenreider, O., & Burgun, A. (2005). Biomedical ontologies. In Medical Informatics (pp. 211-236). Springer, Boston, MA.

Bose, A., & Das, S. (2012). Trial analytics-a tool for clinical trial management. Acta Poloniae Pharmaceutica-Drug Research, 69(3), 523-33.

Boulos, M. N. K., Sanfilippo, A. P., Corley, C. D., & Wheeler, S. (2010). Social Web mining and exploitation for serious applications: Technosocial Predictive Analytics and related technologies for public health, environmental and national security surveillance. Computer methods and programs in biomedicine, 100(1), 16-23

Brouard, B., Bardo, P., Bonnet, C., Mounier, N., Vignot, M., & Vignot, S. (2016). Mobile applications in oncology: is it possible for patients and healthcare professionals to easily identify relevant tools? Ann Med, 48(7), 509-515, doi:10.1080/07853890.2016.1195010.

Boudreaux, E. D., Waring, M. E., Hayes, R. B., Sadasivam, R. S., Mullen, S., & Pagoto, S. (2014). Evaluating and selecting mobile health apps: strategies for healthcare providers and healthcare organizations. Translational behavioral medicine, 4(4), 363- 371.

Boytcheva, S., Angelova, G., Angelov, Z., & Tcharaktchiev, D. (2015). Text mining and big data analytics for retrospective analysis of clinical texts from outpatient care. Cybernetics and Information Technologies, 15(4), 58-77.

Bradley, P. & J., Kaplan, (2010). Turning hospital data into dollars. Healthcare financial management, 64-68.

Bram, J. T., Warwick-Clark, B., Obeysekare, E., & Mehta, K. (2015). Utilization and monetization of healthcare data in developing countries. Big data, 3(2), 59-66.

Broughman, J. R., & Chen, R. C. (2016). Using big data for quality assessment in oncology. Journal of comparative effectiveness research, 5(3), 309-319. .

Brandeau, M. L., Sainfort, F., & Pierskalla, W. P. (Eds.). (2004). Operations research and health care: a handbook of methods and applications (Vol. 70). Springer Science & Business Media.

176

Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.

Brooks, P., El-Gayar, O., & Sarnikar, S. (2015). A framework for developing a domain specific business intelligence maturity model: Application to healthcare. International Journal of Information Management, 35(3), 337-345.

Broughman, J. R., & Chen, R. C. (2016). Using big data for quality assessment in oncology. Journal of comparative effectiveness research, 5(3), 309-319.

Brown, T., Gutman, S. A., Ho, Y. S., & Fong, K. N. (2018). A bibliometric analysis of occupational therapy publications. Scandinavian journal of occupational therapy, 25(1), 1-14.

Burke, J. (2013). Health analytics: gaining the insights to transform health care (Vol. 71). John Wiley & Sons.

Burnham, J. F. (2006). Scopus database: a review. Biomedical digital libraries, 3(1), 1

Byambasuren, O., Beller, E., & Glasziou, P. (2019). Current Knowledge and Adoption of Mobile Health Apps Among Australian General Practitioners: Survey Study. JMIR Mhealth Uhealth, 7(6), e13199. doi:10.2196/13199

Cabitza, F., Rasoini, R., & Gensini, G. F. (2017). Unintended consequences of machine learning in medicine. Jama, 318(6), 517-518.

Calabrese, N., Minkoff, B. & Rawlings, K. (2014). Pharmacosynchrony: Road Map to Transformation in Pharmacy Benefit Management. The American Journal of Pharmacy Benefits.

Calyam, P., Mishra, A., Antequera, R. B., Chemodanov, D., Berryman, A., Zhu, K., ... & Skubic, M. (2016). Synchronous big data analytics for personalized and remote physical therapy. Pervasive and Mobile Computing, 28, 3-20.

Carare, O. (2012). The Impact Of Bestseller Rank On Demand: Evidence From The App Market. International Economic Review, 53(3), 717-742. Char, D. S., Shah, N. H., & Magnus, D. (2018). Implementing machine learning in health care—addressing ethical challenges. The New England journal of medicine, 378(11), 981. Carroll, L. N., Au, A. P., Detwiler, L. T., Fu, T. C., Painter, I. S., & Abernethy, N. F. (2014). Visualization and analytics tools for infectious disease epidemiology: a systematic review. Journal of Biomedical Informatics, 51, 287-298.

Carvalho, J. V., Rocha, Á., Vasconcelos, J., & Abreu, A. (2019). A health data analytics maturity model for hospitals information systems. International Journal of Information Management, 46, 278-285

Castellanos, F. X., Di Martino, A., Craddock, R. C., Mehta, A. D., & Milham, M. P. (2013). Clinical applications of the functional connectome. Neuroimage, 80, 527-540.

177

Catlin, A. C., Malloy, W. X., Arthur, K. J., Gaston, C., Young, J., Fernando, S., & Fernando, R. (2015). Comparative analytics of infusion pump data across multiple hospital systems. American Journal of Health-System Pharmacy, 72(4), 317-324.

Cegarra-Sánchez, J., Cegarra-Navarro, J.-G., Chinnaswamy, A. K., & Wensley, A. (2020). Exploitation and exploration of knowledge: An ambidextrous context for the successful adoption of telemedicine technologies. Technological Forecasting and Social Change, 157, 120089. Celler, B. G., Sparks, R., Nepal, S., Alem, L., Varnfield, M., Li, J., ... & Jayasena, R. (2014). Design of a multi-site multi-state clinical trial of home monitoring of chronic disease in the community in Australia. BMC Public Health, 14(1), 1270.

Chae, B. & Olson, D. L. (2013). Business analytics for supply chain: A dynamic- capabilities framework. International Journal of Information Technology & Decision Making, 12(1), 9-26.

Chalmers, E., Hill, D., Zhao, V., & Lou, E. (2015). Prescriptive analytics applied to brace treatment for AIS: a pilot demonstration. Scoliosis, 10(2), S13.

Char, D. S., Shah, N. H., & Magnus, D. (2018). Implementing machine learning in health care—addressing ethical challenges. The New England journal of medicine, 378(11), 981.

Chatterjee, S., Chakraborty, S., Sarker, S., Sarker, S., & Lau, F. Y. (2009). Examining the success factors for mobile work in healthcare: a deductive study. Decision Support Systems, 46(3), 620-633.

Chawla, N. V., & Davis, D. A. (2013). Bringing big data to personalized healthcare: a patient-centered framework. Journal of general internal medicine, 28(3), 660-665.

Chen, H., Chen, W., Liu, C., Zhang, L., Su, J., & Zhou, X. (2016). Relational Network for Knowledge Discovery through Heterogeneous Biomedical and Clinical Features. Scientific reports, 6, 29915.

Chen, H., Chiang, R., H., L. & Storey, V. C. (2012). Business intelligence and analytics: From big data to big impact. MIS Quarterly, 36(4), 1165-1188. Chen, C. P., & Zhang, C. Y. (2014). Data-intensive applications, challenges, techniques and technologies: A survey on Big Data. Information Sciences, 275, 314-347.

Chen, M., Hao, Y., Hwang, K., Wang, L., & Wang, L. (2017). Disease prediction by machine learning over big data from healthcare communities. IEEE Access, 5, 8869- 8879.

Chen, H., & Fu, Z. (2015). Hadoop-based healthcare information system design and wireless security communication implementation. Mobile Information Systems, 2015.

Chen, W., Zhang, Q., Jin, M., & Yang, J. (2019). Research on online consumer behavior and psychology under the background of big data. Concurrency and Computation: Practice and Experience, 31(10), e4852.

178

Chen, T. J., & Kotecha, N. (2014). Cytobank: providing an analytics platform for community cytometry data analysis and collaboration. In High-Dimensional Single Cell Analysis (pp. 127-157). Springer, Berlin, Heidelberg.

Chong, X., & Sheng-bin, Z. The Process Model of Organization Innovation. In 2006 International Conference on Management Science and Engineering, 2006 (pp. 1302- 1306): IEEE

Choucair, B., Bhatt, J., & Mansour, R. (2015). A bright future: innovation transforming public health in Chicago. Journal of Public Health Management and Practice, 21(Suppl 1), S49.

Chuah, M. H. & Wong, K.L. (2011). A review of business intelligence and its maturitymodels. African Journal of Business Management, 5(9), 3424-3428.

Chute, C. G., Beck, S. A., Fisk, T. B., & Mohr, D. N. (2010). The Enterprise Data Trust at Mayo Clinic: a semantically integrated warehouse of biomedical data. Journal of the American Medical Informatics Association, 17(2), 131-135.

Cleary, M. (2018). How mHealth technology is revolutionizing clinical research. Value & Outcomes Spotlight, 4(5), 20-23

Clift, K., Scott, L., Johnson, M., & Gonzalez, C. (2014). Leveraging geographic information systems in an integrated health care delivery organization. The Permanente Journal, 18(2), 71.

Cook, T. S., & Nagy, P. (2014). Business intelligence for the radiologist: making your data work for you. Journal of the American College of Radiology, 11(12), 1238-1240.

Cooley, R., Mobasher, B., & Srivastava, J. (1997, November). Web mining: Information and pattern discovery on the world wide web. In Tools with Artificial Intelligence, 1997. Proceedings., Ninth IEEE International Conference on (pp. 558- 567). IEEE.

Cowan, L. T., Van Wagenen, S. A., Brown, B. A., Hedin, R. J., Seino-Stephan, Y., Hall, P. C., et al. (2013). Apps of steel: are exercise apps providing consumers with realistic expectations? A content analysis of exercise apps for presence of behavior change theory. Health Education & Behavior, 40(2), 133-139.

Cui, L., Tao, S., & Zhang, G. Q. (2016). Biomedical ontology quality assurance using a big data approach. ACM Transactions on Knowledge Discovery from Data (TKDD), 10(4), 41

Cutillo, C. M., Sharma, K. R., Foschini, L., Kundu, S., Mackintosh, M., Mandl, K. D., et al. (2020). Machine intelligence in healthcare-perspectives on trustworthiness, explainability, usability, and transparency. NPJ Digit Med, 3, 47, doi:10.1038/s41746- 020-0254-2.

179

Cuquet, M., & Fensel, A. (2018). The societal impact of big data: A research roadmap for Europe. Technology in Society, 54, 74-86.

Curcin, V., Woodcock, T., Poots, A. J., Majeed, A., & Bell, D. (2014). Model-driven approach to data collection and reporting for quality improvement. Journal of biomedical informatics, 52, 151-162.

Davenport, T. H. (2006). Competing on analytics, Harvard Business Review, 84(11): 98.

Davenport, T., & Kalakota, R. (2019). The potential for artificial intelligence in healthcare. Future healthcare journal, 6(2), 94.

Davies, S. G. (1997). Re-engineering the right to privacy: how privacy has been transformed from a right to a commodity. Technology and privacy: The new landscape, 143, 144.

Davidson, M. W., Haim, D. A., & Radin, J. M. (2015). Using networks to combine “big data” and traditional surveillance to improve influenza predictions. Scientific reports, 5, 8154.

Davis, F. D. (1989). Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly, 13(3), 319-340.

Dawson, R. M., Felder, T. M., Donevant, S. B., McDonnell, K. K., Card III, E. B., King, C. C., et al. (2020). What makes a good health ‘app’? Identifying the strengths and limitations of existing mobile application evaluation tools. Nursing Inquiry, 27(2), e12333. de Camargo, I. (2012). Screen to Script: The doctor's digital path to treatment: Google/Manhattan Research.

Dean, J., & Ghemawat, S. (2008). MapReduce: simplified data processing on large clusters. Communications of the ACM, 51(1), 107-113.

Delen, D. (2009). Analysis of cancer data: a data mining approach. Expert Systems, 26(1), 100-112

Demir, E. (2014). A decision support tool for predicting patients at risk of readmission: A comparison of classification trees, logistic regression, generalized additive models, and multivariate adaptive regression splines. Decision Sciences, 45(5), 849-880.

Deo, R. C. (2015). Machine Learning in Medicine. Circulation, 132(20), 1920-1930

Devinsky, O., Dilley, C., Ozery-Flato, M., Aharonov, R., Goldschmidt, Y. A., Rosen- Zvi, M., ... & Fritz, P. (2016). Changing the approach to treatment choice in epilepsy using big data. Epilepsy & Behavior, 56, 32-37.

180

De Camargo Fiorini, P., Seles, B. M. R. P., Jabbour, C. J. C., Mariano, E. B., & de Sousa Jabbour, A. B. L. (2018). Management theory and big data literature: From a review to a research agenda. International Journal of Information Management, 43, 112-129.

De Silva, D., Burstein, F. & Jelinek, H. (2015). Addressing the Complexities of Big Data Analytics in Healthcare: The Diabetes Screening Case. Australasian Journal of Information Systems, 19, 99-115.

Dilsizian, S. E. & Siegel, E. L. (2014). Artificial intelligence in medicine and cardiac imaging: harnessing big data and advanced computing to provide personalized medical diagnosis and treatment. Curr Current Cardiology Reports, 16(1), 441.

Dimitriadis, S. I., Laskaris, N. A., Bitzidou, M. P., Tarnanas, I., & Tsolaki, M. N. (2015). A novel biomarker of amnestic MCI based on dynamic cross-frequency coupling patterns during cognitive brain responses. Frontiers in Neuroscience, 9, 350.

Dinov, I. D. (2016). Methodological challenges and analytic opportunities for modeling and interpreting Big Healthcare Data, Gigascience, 5, 12.

Donker, T., Petrie, K., Proudfoot, J., Clarke, J., Birch, M. R., & Christensen, H. (2013). Smartphones for smarter delivery of mental health programs: a systematic review. J Med Internet Res, 15(11), e247, doi:10.2196/jmir.2791.

Doumpos, M., & Zopounidis, C. (2016). Editorial to the special issue “business analytics”. Omega, (59), 1-3.

Duan, L. & Xiong, Y. (2015). Big data analytics and business analytics, Journal of Management Analytics, 2(1), 1-21.

Dubey, R., Gunasekaran, A., & Papadopoulos, T. (2017). Green supply chain management: theoretical framework and further research directions. Benchmarking: An International Journal, 24(1), 184-218.

Dubey, R., Gunasekaran, A., Childe, S. J., Fosso Wamba, S., Roubaud, D., & Forupon, C. (2019a). Empirical Investigation of Data Analytics Capability and Organizational Flexibility as Complements to Supply Chain Resilience. International Journal of Production Research,1-19.

Dubey,R., Gunasekaran, A., Childe,S., Roubaud,D., Fosso Wamba, S., Giannakis,M., & Forupon, C. (2019b). Big Data Analytics and Organizational Culture as Complements to Swift Trust and Collaborative Performance in the Humanitarian Supply Chain. International Journal of Production Economics, 210, 120-136.

Dugan, T. M., Mukhopadhyay, S., Carroll, A., & Downs, S. (2015). Machine learning techniques for prediction of early childhood obesity. Applied clinical informatics, 6(03), 506-520.

Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., et al. (2019). Artificial Intelligence (AI): Multidisciplinary perspectives on emerging

181 challenges, opportunities, and agenda for research, practice and policy. International Journal of Information Management, 101994.

Ehrenhard, M., Wijnhoven, F., van den Broek, T., & Stagno, M. Z. (2017). Unlocking how start-ups create business value with mobile applications: Development of an App- enabled Business Innovation Cycle. Technological Forecasting and Social Change, 115, 26-36.

Ellaway, R. H., Fink, P., Graves, L., & Campbell, A. (2014). Left to their own devices: medical learners’ use of mobile technologies. Medical teacher, 36(2), 130-138.

Engström, P., &Forsell, E. (2018). Demand effects of consumers’ stated and revealed preferences. Journal of Economic Behavior & Organization, 150, 43-61.

European Commission (2018). Privacy Code of Conduct on mobile health apps. Available at: https://ec.europa.eu/digital-single-market/en/privacy-code-conduct- mobile-health-apps.

Evangelatos, N., Reumann, M., Lehrach, H., & Brand, A. (2016). Clinical Trial Data as Public Goods: Fair Trade and the Virtual Knowledge Bank as a Solution to the Free Rider Problem-A Framework for the Promotion of Innovation by Facilitation of Clinical Trial Data Sharing among Biopharmaceutical Companies in the Era of Omics and Big Data. Public health genomics, 19(4), 211-219.

Fabian, B., Ermakova, T. & Junghanns, P. (2015). Collaborative and secure sharing of healthcare data in multi-clouds. Information Systems, 48, 132-150.

Feldman, B., Martin, E. M., & Skotnes, T. (2016). Big data in healthcare hype and hope. Dr. Bonnie 360 (2012).

Fernando, J., & Lindley, J. (2018). Lessons learned from piloting mHealth informatics practice curriculum into a medical elective. Journal of the American Medical Informatics Association, 25(4), 380-384.

Filkins, B. L., Kim, J. Y., Roberts, B., Armstrong, W., Miller, M. A., Hultner, M. L., ... & Steinhubl, S. R. (2016). (2016). Privacy and security in the era of digital health: what should translational researchers know and do about it? American journal of translational research, 8(3), 1560.

Flechet, M., Grandas, F. G., & Meyfroidt, G. (2016). Informatics in neurocritical care: new ideas for Big Data. Current Opinion in Critical Care, 22(2), 87-93.

Ferguson, M. (2012). Architecting a big data platform for analytics. A Whitepaper prepared for IBM, 30.

Fleurence, R. L., Beal, A. C., Sheridan, S. E., Johnson, L. B., & Selby, J. V. (2014). Patient-powered research networks aim to improve patient care and health research. Health Affairs, 33(7), 1212-1219.

182

Flick, C., Zamani, E. D., Stahl, B. C., & Brem, A. (2020). The future of ICT for health and ageing: Unveiling ethical and social issues through horizon scanning foresight. Technological Forecasting and Social Change, 155, 119995.

Flores, C. D., Barros, P., Cazella, S., & Bez, M. R (2013). Leveraging the Learning Process in Health through Clinical Cases Simulator. IEEE 2nd International Conference on Serious Games and Applications for Health (SeGAH). Forsberg, J. A., Potter, B. K., Wagner, M. B., Vickers, A., Dente, C. J., Kirk, A. D., & Elster, E. A. (2015). Lessons of war: turning data into decisions. EBioMedicine, 2(9), 1235-1242.

Fox, G., & Connolly, R. (2018). Mobile health technology adoption across generations: Narrowing the digital divide. Information Systems Journal, 28(6), 995-1019.

Fox, G., & James, T. L. (2020). Toward an Understanding of the Antecedents to Health Information Privacy Concern: A Mixed Methods Study. Information Systems Frontiers, 1-26.

Free, C., Phillips, G., Watson, L., Galli, L., Felix, L., Edwards, P., et al. (2013). The effectiveness of mobile-health technologies to improve health care service delivery processes: a systematic review and meta-analysis. PLoS medicine, 10(1).

Fuller, D., Shareck, M., & Stanley, K. (2017). Ethical implications of location and accelerometer measurement in health research studies with mobile sensing devices. Soc Sci Med, 191, 84-88. doi:10.1016/j.socscimed.2017.08.043

Gagnon, M. P., Ngangue, P., Payne-Gagnon, J., & Desmartis, M. (2016). m-Health adoption by healthcare professionals: a systematic review. J Am Med Inform Assoc, 23(1), 212-220. doi:10.1093/jamia/ocv052

Gaitanou, P., Garoufallou, E. & Balatsoukas, P. (2014). The Effectiveness of Big Data in Health Care: Systematic Review, Metadata and Semantics Research, 141-153.

Gale, T. C., Chatterjee, A., Mellor, N. E., & Allan, R. J. (2016). Health worker focused distributed simulation for improving capability of health systems in Liberia. Simulation in Healthcare, 11(2), 75-81.

Galetsi, P., Katsaliaki, K., & Kumar, S. (2019). Values, challenges and future directions of big data analytics in healthcare: A systematic review. Social Science & Medicine, 112533.

Gamble, A. (2020). Artificial intelligence and mobile apps for mental healthcare: a social informatics perspective. Aslib Journal of Information Management.

Gandomi, A., & Haider, M. (2015). Beyond the hype: Big data concepts, methods, and analytics. International Journal of Information Management, 35(2), 137-144.

García, M. V., Blasco López, M. F., & Sastre Castillo, M. Á. (2019). Determinants of the acceptance of mobile learning as an element of human capital training in organisations. Technological Forecasting and Social Change, 149(C).

183

Ghasemaghaei, M., Hassanein, K., & Turel, O. (2017). Increasing firm agility through the use of data analytics: The role of fit. Decision Support Systems, 101, 95-105.

Ghose, A., & Han, S. P. (2014). Estimating Demand for Mobile Applications in the New Economy. Management Science, 60(6), 1470-1488.

Ginsberg, J., Mohebbi, M. H., Patel, R. S., Brammer, L., Smolinski, M. S., & Brilliant, L. (2009). Detecting influenza epidemics using search engine query data. Nature, 457(7232), 1012.

Gligorijević, V., Malod‐Dognin, N., & Pržulj, N. (2016). Integrative methods for analyzing big data in precision medicine. Proteomics, 16(5), 741-758.

Goh, W. P., Tao, X., Zhang, J., & Yong, J. (2016). Decision support systems for adoption in dental clinics: a survey. Knowledge-Based Systems, 104, 195-206.

Goldenholz, D. M., Moss, R., Scott, J., Auh, S., & Theodore, W. H. (2015). Confusing placebo effect with natural history in epilepsy: a big data approach. Annals of Neurology, 78(3), 329-336.

Goldhahn, J., Rampton, V., & Spinas, G. A. (2018). G. A. (2018). Could artificial intelligence make doctors obsolete? BMJ : British medical journal,, 363, 4563, doi:10.1136/bmj.k4563 10.5167/uzh-158375. Gombar, S., Callahan, A., Califf, R., Harrington, R., & Shah, N. H. (2019). It is time to learn from patients like mine. NPJ digital medicine, 2(1), 1-3. Gorman, M. F., & Klimberg, R. K. (2014). Benchmarking academic programs in business analytics. Interfaces, 44(3), 329-341.

Grimmer, J. (2015). We are all social scientists now: how big data, machine learning, and causal inference work together. PS: Political Science & Politics, 48(1), 80-83.

Grover, V., Lim, J., & Ayyagari, R. (2006). The dark side of information and market efficiency in e‐markets. Decision Sciences, 37(3), 297-324.

Groves, P., Kayyali, B., Knott, D., & Van Kuiken, S. (2013). The ‘big data’revolution in healthcare. McKinsey Quarterly, 2, 3.

Guimarães, A. J., Silva Araujo, V. J., de Campos Souza, P. V., Araujo, V. S., & Rezende, T. S. (2018). Using Fuzzy Neural Networks to the Prediction of Improvement in Expert Systems for Treatment of Immunotherapy. 11238, 229-240, doi:10.1007/978- 3-030-03928-8_19.

Gulshan V, Peng L, Coram M, et al. Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs.JAMA. 2016;316(22):2402-2410.

184

Gurtner, S., Reinhardt, R., & Soyez, K. (2014). Designing mobile business applications for different age groups. Technological Forecasting and Social Change, 88, 177-188.

Gruebner, O., Sykora, M. D., Lowe, S. R., Shankardass, K., Galea, S., & Subramanian, S. V. (2017). Big data opportunities for social behavioral and mental health research. Social Science and Medicine, 189, pp. 167-169

Gunasekaran, A., Papadopoulos, T., Dubey, R., Wamba, S. F., Childe, S. J., Hazen, B., & Akter, S. (2017). Big data and predictive analytics for supply chain and organizational performance. Journal of Business Research, 70, 308-317.

Guo, J., Liu, H., & Zheng, J. (2015). SynLethDB: synthetic lethality database toward discovery of selective and sensitive anticancer drug targets. Nucleic acids research, 44(D1), D1011-D1017.

Gupta, S., Kar, A. K., Baabdullah, A., & Al-Khowaiter, W. A. (2018). Big data with cognitive computing: a review for the future. International Journal of Information Management, 42, 78-89.

Hao, S., Jin, B. O., Shin, A. Y., Zhao, Y., Zhu, C., Li, Z., ... & Zhao, Y. (2014). Risk prediction of emergency department revisit 30 days post discharge: a prospective study. PloS one, 9(11), e112944.

Hazen, B. T., Skipper, J. B., Boone, C. A., & Hill, R. R. (2018). Back in business: Operations research in support of big data analytics for operations and supply chain management. Annals of Operations Research, 270(1-2), 201-211.

He, P., Wang, P., Gao, J., & Tang, B. (2015). City-wide smart healthcare appointment systems based on cloud data virtualization paas. International Journal of Multimedia and Ubiquitous Engineering, 10(2), 371-382.

Higgins, J. P. (2016). Smartphone applications for patients' health and fitness. The American journal of medicine, 129(1), 11-19.

Hilario, M., Kalousis, A., Pellegrini, C., & Mueller, M. (2006). Processing and classification of protein mass spectra. Mass spectrometry reviews, 25(3), 409-449.

Hilbert, M. (2014). Technological information inequality as an incessantly moving target: The redistribution of information and communication capacities between 1986 and 2010. Journal of the Association for Information Science and Technology, 65(4), 821-835.

Hindle, G. A., Vidgen, R. (2018) Developing a business analytics methodology: A case study in the foodbank sector. European Journal of Operational Research 268, 836–851.

Hoeppner, B. B., Hoeppner, S. S., Seaboyer, L., Schick, M. R., Wu, G. W., Bergman, B. G., et al. (2016). How smart are smartphone apps for smoking cessation? A content analysis. Nicotine & Tobacco Research, 18(5), 1025-1031.

185

Hofer, F., & Haluza, D. (2019). Are Austrian practitioners ready to use medical apps? Results of a validation study. BMC Med Inform Decis Mak, 19(1), 88. doi:10.1186/s12911-019-0811-2

Holdener, M., Gut, A., & Angerer, A. (2020). Applicability of the User Engagement Scale to Mobile Health: A Survey-Based Quantitative Study. JMIR Mhealth Uhealth, 8(1), e13244. doi:10.2196/13244

Holzinger, A., & Jurisica, I. (2014). Knowledge discovery and data mining in biomedical informatics: The future is in integrative, interactive machine learning solutions. In Interactive knowledge discovery and data mining in biomedical informatics (pp. 1-18). Springer, Berlin, Heidelberg.

Holzinger, A., Schantl, J., Schroettner, M., Seifert, C., & Verspoor, K. (2014). Biomedical text mining: state-of-the-art, open problems and future challenges. In Interactive knowledge discovery and data mining in biomedical informatics (pp. 271- 300). Springer, Berlin, Heidelberg.

Hong, J. I., Ng, J. D., Lederer, S., & Landay, J. A. Privacy risk models for designing privacy-sensitive ubiquitous computing systems. In Proceedings of the 5th conference on Designing interactive systems: processes, practices, methods, and techniques, 2004 (pp. 91-100)

Horgan, D., Romao, M., Morre, S. A., & Kalra, D. (2019). Artificial Intelligence: Power for Civilisation - and for Better Healthcare. Public Health Genomics, 22(5-6), 145-161, doi:10.1159/000504785.

Hsu, C.-L., & Lin, J. C.-C. (2016). Effect of perceived value and social influences on mobile app stickiness and in-app purchase intention. Technological Forecasting and Social Change, 108, 42-53.

Hsu, C.-L., & Lin, J. C.-C. (2015). What drives purchase intention for paid mobile apps?–An expectation confirmation model with perceived value. Electronic Commerce Research and Applications, 14(1), 46-57.

Hu, H., Langford, J., Caruana, R., Mukherjee, S., Horvitz, E. J., & Dey, D. Efficient forward architecture search. In Advances in Neural Information Processing Systems, 2019 (pp. 10122-10131)

Huang, M., Nichols, T., Huang, C., Yu, Y., Lu, Z., Knickmeyer, R. C., ... & Alzheimer's Disease Neuroimaging Initiative. (2015). FVGWAS: Fast voxelwise genome wide association analysis of large-scale imaging genetic data. Neuroimage, 118, 613-627.

Huang, M., Han, H., Wang, H., Li, L., Zhang, Y., & Bhatti, U. A. (2018). A clinical decision support framework for heterogeneous data sources. IEEE journal of biomedical and health informatics, 22(6), 1824-1833.

Hughes, A., Landers, D., Arkenau, H. T., Shah, S., Stephens, R., Mahal, A., ... & Royle, J. (2016). Development and evaluation of a new technological way of engaging patients

186 and enhancing understanding of drug tolerability in early clinical development: PROACT. Advances in Therapy, 33(6), 1012-1024.

Ienca, M., Vayena, E., & Blasimme, A. (2018). Big Data and Dementia: charting the route Ahead for research, ethics, and Policy. Frontiers in medicine, 5, 13.

Iqbal, U., Hsu, C. K., Nguyen, P. A. A., Clinciu, D. L., Lu, R., Syed-Abdul, S., ... & Chang, Y. C. (2016). Cancer-disease associations: A visualization and animation through medical big data. Computer methods and programs in biomedicine, 127, 44-51.

Ismagilova, E., Hughes, L., Rana, N. P., & Dwivedi, Y. K. (2020). Security, Privacy and Risks Within Smart Cities: Literature Review and Development of a Smart City Interaction Framework. Information Systems Frontiers, 1-22.

Istepanian, R. S., & Al-Anzi, T. (2018). m-Health 2.0: new perspectives on mobile health, machine learning and big data analytics. Methods, 151, 34-40.

Istephan, S., & Siadat, M. R. (2016). Unstructured medical image query using big data– an epilepsy case study. Journal of biomedical informatics, 59, 218-226.

Ivan, M., & Velicanu, M. (2015). Healthcare industry improvement with business intelligence. Informatica Economica, 19(2), 81.

Iyengar, K., Upadhyaya, G. K., Vaishya, R., & Jain, V. (2020). COVID-19 and applications of smartphone technology in the current pandemic. Diabetes & Metabolic Syndrome: Clinical Research & Reviews.

Jahangirian, M., Naseer, A., Stergioulas, L., Young, T., Eldabi, T., Brailsford, S., ... & Harper, P. (2012). Simulation in health-care: lessons from other sectors. Operational Research, 12(1), 45-55.

Jahn, H. K., Jahn, I. H., Roland, D., Lyttle, M. D., Behringer, W., & Peruki. (2019). Mobile device and app use in paediatric emergency care: a survey of departmental practice in the UK and Ireland. Arch Dis Child, 104(12), 1203-1207. doi:10.1136/archdischild-2019-316872

Jaklič, J., Grublješič, T., & Popovič, A. (2018). The role of compatibility in predicting business intelligence and analytics use intentions. International Journal of Information Management, 43, 305-318.

Javaid, M., Haleem, A., Vaishya, R., Bahl, S., Suman, R., & Vaish, A. (2020). Industry 4.0 technologies and their applications in fighting COVID-19 pandemic. Diabetes & Metabolic Syndrome: Clinical Research & Reviews.

Jayaraman, P. P., Forkan, A. R. M., Morshed, A., Haghighi, P. D., & Kang, Y. B. (2019). Healthcare 4.0: A review of frontiers in digital health. WIREs Data Mining and Knowledge Discovery, 10(2), doi:10.1002/widm.1350.

187

Jensen, P. B., Jensen, L. J., & Brunak, S. (2012). Mining electronic health records: towards better research applications and clinical care. Nature Reviews Genetics, 13(6), 395.

Jeon, B., Jeong, B., Jee, S., Huang, Y., Kim, Y., Park, G. H., et al. (2019). A facial recognition mobile app for patient safety and biometric identification: Design, development, and validation. JMIR mHealth and uHealth, 7(4), e11472.

Jones, S., Irani, Z., Sivarajah, U., & Love, P. E. (2017). Risks and rewards of cloud computing in the UK public sector: A reflection on three Organisational case studies. Information Systems Frontiers, 1-24.

Jun, J. B., Jacobson, S. H., & Swisher, J. R. (1999). Application of discrete-event simulation in health care clinics: A survey. Journal of the operational research society, 50(2), 109-123.

Jung, Y. (2014). What a smartphone is to me: understanding user values in using smartphones. Information Systems Journal, 24(4), 299-321.

Junglas, I., Goel, L., Ives, B., & Harris, J. (2019). Innovation at work: The relative advantage of using consumer IT in the workplace. Information Systems Journal, 29(2), 317-339.

Jutel, A., & Lupton, D. (2015). Digitizing diagnosis: a review of mobile applications in the diagnostic process. Diagnosis, 2(2), 89-96.

Kambatla, K., Kollias, G., Kumar, V., & Grama, A. (2014). Trends in big data analytics. Journal of Parallel and Distributed Computing, 74(7), 2561-2573

Kapoor, A., Guha, S., Das, M. K., Goswami, K. C., & Yadav, R. (2020). Digital healthcare: The only solution for better healthcare during COVID-19 pandemic? Indian Heart Journal.

Katsaliaki, K., & Mustafee, N. (2011). Applications of simulation within the healthcare context. Journal of the Operational Research Society, 62(8), 1431-1451.

Katsaliaki, K., Mustafee, N., & Kumar, S. (2014). A game-based approach towards facilitating decision making for perishable products: An example of blood supply chain. Expert Systems with Applications, 41(9), 4043-4059.

Katircioglu, K., Gooby, R., Helander, M., Drissi, Y., Chowdhary, P., Johnson, M., & Yonezawa, T. (2014). Supply chain scenario modeler: A holistic executive decision support solution. Interfaces, 44(1), 85-104.

Kay, M., Santos, J., & Takane, M. (2011). mHealth: New horizons for health through mobile technologies. World Health Organization, 64(7), 66-71.

Keith, M. J., Babb, J. S., Lowry, P. B., Furner, C. P., & Abdullat, A. (2015). The role of mobile‐computing self‐efficacy in consumer information disclosure. Information Systems Journal, 25(6), 637-667.

188

Kenner, A. (2016). Asthma on the move: how mobile apps remediate risk for disease management. Health, Risk & Society, 17(7-8), 510-529.

Kernebeck, S., Busse, T. S., Böttcher, M. D., Weitz, J., Ehlers, J., & Bork, U. (2020). Impact of mobile health and medical applications on clinical practice in gastroenterology. World journal of gastroenterology, 26(29), 4182.

Khalaf, M., Hussain, A. J., Keight, R., Al-Jumeily, D., Fergus, P., Keenan, R., & Tso, P. (2017). Machine learning approaches to the application of disease modifying therapy for sickle cell using classification models. Neurocomputing, 228, 154-164.

Khan, N., Yaqoob, I., Hashem, I. A. T., Inayat, Z., Ali, M., Kamaleldin, W., ... & Gani, A. (2014). Big data: survey, technologies, opportunities, and challenges. The Scientific World Journal, 2014.

Kim, H., Lee, C. H., Kim, S. H., & Kim, Y. D. (2018). Epidemiology of complex regional pain syndrome in Korea: An electronic population health data study. PloS one, 13(6), e0198147.

Kim, B. Y., Sharafoddini, A., Tran, N., Wen, E. Y., & Lee, J. (2018). Consumer Mobile Apps for Potential Drug-Drug Interaction Check: Systematic Review and Content Analysis Using the Mobile App Rating Scale (MARS). JMIR Mhealth Uhealth, 6(3), e74. doi:10.2196/mhealth.8613

Kite, B. J., Tangasi, W., Kelley, M., Bower, J. K., & Foraker, R. E. (2015). Electronic medical records and their use in health promotion and population research of cardiovascular disease. Current Cardiovascular Risk Reports, 9(1), 422.

Klein, S. M. (2015). Generating Real‐Time, Actionable Outcome Measures at Blythedale Children's Hospital. Global Business and Organizational Excellence, 34(4), 6-17.

Knitza, J., Tascilar, K., Messner, E. M., Meyer, M., Vossen, D., Pulla, A., ... & Mucke, J. (2019). German mobile Apps in rheumatology: review and analysis using the mobile application rating scale (MARs). JMIR mHealth and uHealth, 7(8), e14991.

Kodratoff, Y. (2014). Research in machine learning: Recent progress, classification of methods, and future directions. Y. Kodratoff, & RS Michalski. Machine learning: an artificial intelligence approach, 3-30.

Koumpouros, Y., & Georgoulas, A. (2020). A systematic review of mHealth funded R&D activities in EU: Trends, technologies and obstacles. Informatics for Health and Social Care, 45(2), 168-187

Kosse, R. C., Murray, E., Bouvy, M. L., de Vries, T. W., Stevenson, F., & Koster, E. S. (2020). Potential normalization of an asthma mHealth intervention in community pharmacies: Applying a theory-based framework. Res Social Adm Pharm, 16(2), 195- 201, doi:10.1016/j.sapharm.2019.05.004.

189

Krishnan, G., & Selvam, G. (2019). Factors influencing the download of mobile health apps: Content review-led regression analysis. Health Policy and Technology, 8(4), 356- 364, doi:10.1016/j.hlpt.2019.09.001

Krumholz, H. M. (2014). Big data and new knowledge in medicine: the thinking, training, and tools needed for a learning health system. Health Affairs, 33(7), 1163- 1170.

Kübler, R., Pauwels, K., Yildirim, G., & Fandrich, T. (2018). App Popularity: Where in the World are Consumers Most Sensitive to Price and User Ratings? Journal of Marketing, 82(5), 20-44. doi:10.1509/jm.16.0140 Kuiler, E. W. (2014). From Big Data to Knowledge: An Ontological Approach to Big Data Analytics. Review of Policy Research, 31(4), 311-318.

Kulkarni, P., Smith, L. D., & Woeltje, K. F. (2016). Assessing risk of hospital readmissions for improving medical practice. Health care management science, 19(3), 291-299.

Kumar, M., Singh, J. B., Chandwani, R., & Gupta, A. (2020). “Context” in healthcare information technology resistance: A systematic review of extant literature and agenda for future research. International Journal of Information Management, 51, 102044.

Kumar, S., Abowd, G. D., Abraham, W. T., al’Absi, M., Gayle Beck, J., Chau, D. H., ... & Ganesan, D. (2015). Center of excellence for mobile sensor data-to-knowledge (MD2K). Journal of the American Medical Informatics Association, 22(6), 1137-1142.

Labrique, A. B., Vasudevan, L., Kochi, E., Fabricant, R., & Mehl, G. (2013). mHealth innovations as health system strengthening tools: 12 common applications and a visual framework. Global health: science and practice, 1(2), 160-171.

Larson, D., & Chang, V. (2016). A review and future direction of agile, business intelligence, analytics and data science. International Journal of Information Management, 36(5), 700-710.

Lary, D. J., Woolf, S., Faruque, F., & LePage, J. P. (2014). Holistics 3.0 for Health. ISPRS International Journal of Geo-Information, 3(3), 1023-1038. Lashitew, A. A., Bals, L., & van Tulder, R. (2020). Inclusive business at the base of the pyramid: the role of embeddedness for enabling social innovations. Journal of Business Ethics, 162(2), 421-448.

Lazarou, I., Karakostas, A., Stavropoulos, T. G., Tsompanidis, T., Meditskos, G., Kompatsiaris, I., & Tsolaki, M. (2016). A novel and intelligent home monitoring system for care support of elders with cognitive impairment. Journal of Alzheimer's Disease, 54(4), 1561-1591.

190

Lazer, D., Kennedy, R., King, G., & Vespignani, A. (2014). The parable of Google Flu: traps in big data analysis. Science, 343(6176), 1203-1205.

Lee, S. Y., & Lee, K. (2018). Factors that influence an individual's intention to adopt a wearable healthcare device: The case of a wearable fitness tracker. Technological Forecasting and Social Change, 129, 154-163.

Lee, G., & Raghu, T. S. (2014). Determinants of Mobile Apps Success: Evidence from App Store Market. Journal of Management Information Systems, 3(12), 133-170.

Li, Y., & Guo, Y. (2016). Wiki-health: from quantified self to self-understanding. Future Generation Computer Systems, 56, 333-359.

Li, Y., Zhang, X., Guo, X., & Wang, L. (2019). Underlying Emotional Mechanisms of Routine m-Health Use in Chronically Ill Patients. IEEE Transactions on Engineering Management.

Lim, C., Kim, K. H., Kim, M. J., Heo, J. Y., Kim, K. J., & Maglio, P. P. (2018). From data to value: A nine-factor framework for data-based value creation in information- intensive services. International Journal of Information Management, 39, 121-135.

Lister, C., West, J. H., Cannon, B., Sax, T., & Brodegard, D. (2014). Just a fad? Gamification in health and fitness apps. JMIR serious games, 2(2), e9.

Litt, R. S. (2013). Privacy, Technology and National Security: An Overview of Intelligence Collection. Speech at The Brookings Institution, Washington, DC, July, 19.

Liu, P., & Wu, S. (2016). An agent-based simulation model to study accountable care organizations. Health care management science, 19(1), 89-101.

Llorens-Vernet, P., & Miro, J. (2020). Standards for Mobile Health-Related Apps: Systematic Review and Development of a Guide. JMIR Mhealth Uhealth, 8(3), doi:10.2196/13057.

Loebbecke, C., & Picot, A. (2015). Reflections on societal and business model transformation arising from digitization and big data analytics: A research agenda. The Journal of Strategic Information Systems, 24(3), 149-157. Loi, M., Heitz, C., Ferrario, A., Schmid, A., & Christen, M. (2019). Towards an Ethical Code for Data-Based Business. 6-12, doi:10.1109/sds.2019.00-15.

Lomotey, R. K., & Deters, R. (2016). Middleware for mobile medical data management with minimal latency. Information Systems Frontiers, 20(6), 1281-1296, doi:10.1007/s10796-016-9729-8.

191

López-Martínez, F., Schwarcz, A., Núñez-Valdez, E. R., & García-Díaz, V. (2018). Machine Learning Classification Analysis for a Hypertensive Population as a Function of Several Risk Factors. Expert Systems with Applications.

Lowton, K., Hiley, C., & Higgs, P. (2017). Constructing embodied identity in a ‘new’ ageing population: A qualitative study of the pioneer cohort of childhood liver transplant recipients in the UK. Social Science & Medicine, 172, 1-9.

Lupton, D. (2014). The commodification of patient opinion: the digital patient experience economy in the age of big data. Sociology of health & illness, 36(6), 856- 869.

Lupton, D., & Jutel, A. (2015). ‘It's like having a physician in your pocket!’A critical analysis of self-diagnosis smartphone apps. Social Science & Medicine, 133, 128-135.

Luo, J., Wu, M., Gopukumar, D., & Zhao, Y. (2016). Big data application in biomedical research and health care: a literature review. Biomedical Informatics Insights, 8, BII- S31559.

Luo, L., Li, L., Hu, J., Wang, X., Hou, B., Zhang, T., & Zhao, L. P. (2016). A hybrid solution for extracting structured medical information from unstructured data in medical records via a double-reading/entry system. BMC medical informatics and decision making, 16(1), 114. Lyles, C. R., Harris, L. T., Le, T., Flowers, J., Tufano, J., Britt, D., et al. (2011). Qualitative evaluation of a mobile phone and web-based collaborative care intervention for patients with type 2 diabetes. Diabetes technology & therapeutics, 13(5), 563-569.

Ma, J., Stahl, L., & Knotts, E. (2018). Emerging roles of health information professionals for library and information science curriculum development: a scoping review. Journal of the Medical Library Association: JMLA, 106(4), 432

Maccione, A., Gandolfo, M., Zordan, S., Amin, H., Di Marco, S., Nieus, T., ... & Berdondini, L. (2015). Microelectronics, bioinformatics and neurocomputation for massive neuronal recordings in brain circuits with large scale multielectrode array probes. Brain research bulletin, 119, 118-126.

Madanian, S., Parry, D., Airehrour, D., & Cherrington, M. (2019). mHealth and big- data integration: promises for healthcare system in India. BMJ Health & Informatics, 26(1).

Mamary, G. (2013). Making the leap to real-time analytics: Hunterdon Healthcare implements a new decision-support system. Health management technology, 34(2), 16.

192

Mamonov, S. & Triantoro, T. M. (2018). The strategic value of data resources in emergent industries, International Journal of Information Management 39, 146–155.

Manchanda, R., & Jacobs, I. (2016). Genetic screening for gynecological cancer: where are we heading?. Future Oncology, 12(2), 207-220.

Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., & Byers, A. H. (2011). Big data: The next frontier for innovation, competition, and productivity, McKinsey & Company

Marés, J., Shamardin, L., Weiler, G., Anguita, A., Sfakianakis, S., Neri, E., ... & Coveney, P. V. (2014). p-medicine: A medical informatics platform for integrated large scale heterogeneous patient data. In AMIA Annual Symposium Proceedings (Vol. 2014, p. 872). American Medical Informatics Association.

Marsden, C. (2017). How law and computer science can work together to improve the information society. Communications of the ACM, 61(1), 29-31

Martin, K., Shilton, K., & Smith, J. (2019). Business and the Ethical Implications of Technology: Introduction to the Symposium. Journal of Business Ethics(160), 307–317

Martínez-Pérez, B., de la Torre-Díez, I., López-Coronado, M., Sainz-De-Abajo, B., Robles, M., & García-Gómez, J. M. (2014). Mobile clinical decision support systems and applications: a literature and commercial review. Journal of medical systems, 38(1), 4.

Massey, A. P., Khatri, V., & Montoya‐Weiss, M. M. (2007). Usability of online services: The role of technology readiness and context. Decision Sciences, 38(2), 277- 308.

May, C. (2013). Towards a general theory of implementation. Implementation Science, 8(18), 14.

May, C., & Finch, T. (2009). Implementing, embedding, and integrating practices: an outline of normalization process theory. Sociology, 43(3), 535-554.

McClay, W. A., Yadav, N., Ozbek, Y., Haas, A., Attias, H. T., & Nagarajan, S. S. (2015). A real-time magnetoencephalography brain-computer interface using interactive 3D visualization and the Hadoop ecosystem. Brain sciences, 5(4), 419-440.

McGirt, M. J., Sivaganesan, A., Asher, A. L., & Devin, C. J. (2015). Prediction model for outcome after low-back surgery: individualized likelihood of complication, hospital readmission, return to work, and 12-month improvement in functional disability. Neurosurgical focus, 39(6), E13.

McKay, F. H., Cheng, C., Wright, A., Shill, J., Stephens, H., & Uccellini, M. (2018). Evaluating mobile phone applications for health behaviour change: a systematic review. Journal of telemedicine and telecare, 24(1), 22-30.

193

McKie, E. C., Ferguson, M. E., Galbreth, M. R., & Venkataraman, S. (2018). How Do Consumers Choose between Multiple Product Generations and Conditions? An Empirical Study of iPad Sales on eBay. Production and Operations Management, 27(8), 1574-1594. doi:10.1111/poms.12884

McNiel, P., & McArthur, E. C. (2016). Evaluating health mobile apps: information literacy in undergraduate and graduate nursing courses. Journal of Nursing Education, 55(8), 480-480.

Mehtab, A., Shahid, W. B., Yaqoob, T., Amjad, M. F., Abbas, H., Afzal, H., et al. (2019). AdDroid: Rule-Based Machine Learning Framework for Android Malware Analysis. Mobile Networks and Applications, 25(1), 180-192, doi:10.1007/s11036-019- 01248-0.

Mendiola, M. F., Kalnicki, M., & Lindenauer, S. (2015). Valuable features in mobile health apps for patients and consumers: content analysis of apps and user ratings. JMIR Mhealth Uhealth, 3(2), e40, doi:10.2196/mhealth.4283.

Miriovsky, B. J., Shulman, L. N., & Abernethy, A. P. (2012). Importance of health information technology, electronic health records, and continuously aggregating data to comparative effectiveness research and learning health care. Journal of Clinical Oncology, 30(34), 4243-4248.

Mittelstadt, B. D., & Floridi, L. (2016). The ethics of big data: current and foreseeable issues in biomedical contexts. Science and engineering ethics, 22(2), 303-341.

Mohan M, Vigneshwaran B., Vineeth Raj G. & Harlin Jesuva Prince S. (2016). Disease Diagnosis for Personalized Health Care Using Map Reduce Technique, I J C T A, 9(5), 2153-2164.

Montiel, I., Delgado-Ceballos, J., Ortiz-de-Mandojana, N., & Antolin-Lopez, R. (2020). New ways of teaching: using technology and mobile apps to educate on societal grand challenges. Journal of Business Ethics, 161(2), 243-251.

Mooney, S. J., & Pejaver, V. (2018). Big data in public health: terminology, machine learning, and privacy. Annual review of public health, 39, 95-112.

Moore, G. C., & Benbasat, I. (1996). Integrating diffusion of innovations and theory of reasoned action models to predict utilization of information technology by end-users Diffusion and adoption of information technology (pp. 132-146): Springer.

Mosa, A. S. M., Yoo, I., & Sheets, L. (2012). A systematic review of healthcare applications for smartphones. BMC medical informatics and decision making, 12(1), 67.

194

Murdoch, T. B., & Detsky, A. S. (2013). The inevitable application of big data to health care. Jama, 309(13), 1351-1352.

Murphy, S. N., Weber, G., Mendis, M., Gainer, V., Chueh, H. C., Churchill, S., & Kohane, I. (2010). Serving the enterprise and beyond with informatics for integrating biology and the bedside (i2b2). Journal of the American Medical Informatics Association, 17(2), 124-130.

Murray, E., Treweek, S., Pope, C., MacFarlane, A., Ballini, L., Dowrick, C., et al. (2010). Normalisation process theory: a framework for developing, evaluating and implementing complex interventions. BMC medicine, 8(1), 63.

Nerminathan, A., Harrison, A., Phelps, M., Scott, K. M., & Alexander, S. (2017). Doctors’ use of mobile devices in the clinical setting: A mixed methods study. Internal Medicine Journal, 47(3), 291-298.

Newel, Sl & Marabelli, M. (2015). Strategic opportunities (and challenges) of algorithmic decision – making: A call for action on the long term societal effects of “datification”, Journal of Strategic Information Systems, 24(1), 3-14.

Nie, P. & Li, B. (2011). A Cluster-based Data Aggregation Architecture in WSN for Structural Health Monitoring, IEEE, 546-552.

Ngwenya, N., Farquhar, M., & Ewing, G. (2016). Sharing bad news of a lung cancer diagnosis: understanding through communication privacy management theory. Psycho‐Oncology, 25(8), 913-918.

Nicholas, J., Larsen, M. E., Proudfoot, J., & Christensen, H. (2015). Mobile Apps for Bipolar Disorder: A Systematic Review of Features and Content Quality. J Med Internet Res, 17(8), e198, doi:10.2196/jmir.4581.

Noorbakhsh-Sabet, N., Zand, R., Zhang, Y., & Abedi, V. (2019). Artificial Intelligence Transforms the Future of Health Care. Am J Med, 132(7), 795-801, doi:10.1016/j.amjmed.2019.01.017.

Nouri, R., R Niakan Kalhori, S., Ghazisaeedi, M., Marchand, G., & Yasini, M. (2018). Criteria for assessing the quality of mHealth apps: a systematic review. Journal of the American Medical Informatics Association, 25(8), 1089-1098.

Oddo, A. R. (2001). Healthcare ethics: a patient-centered decision model. Journal of Business Ethics, 29(1-2), 125-134.

O’Connell, M. (2012). Big Data Analytics: Scaling Up and Out in the Event-Enabled Enterprise. Wall Street Technology Association, Ticker, (3).

195

O’Driscoll, A., Daugelaite, J., & Sleator, R. D. (2013). ‘Big data’, Hadoop and cloud computing in genomics. Journal of biomedical informatics, 46(5), 774-781.

O'Donovan, J., Bersin, A., & O'Donovan, C. (2015). The effectiveness of mobile health (mHealth) technologies to train healthcare professionals in developing countries: a review of the literature. BMJ Innovations, 1(1), 33-36.

O'Loughlin, K., Neary, M., Adkins, E. C., & Schueller, S. M. (2019). Reviewing the data security and privacy policies of mobile apps for depression. Internet interventions, 15, 110-115.

Obermeyer, Z. & Ezekiel, J.E. (2016). Predicting the Future - Big Data, Machine Learning, and Clinical Medicine , The New England Journal of Medicine, 375, 13.

Paglialonga, A., Lugo, A., & Santoro, E. (2018). An overview on the emerging area of identification, characterization, and assessment of health apps. J Biomed Inform, 83, 97- 102. doi:10.1016/j.jbi.2018.05.017

Panch, T., Szolovits, P., & Atun, R. (2018). Artificial intelligence, machine learning and health systems. Journal of global health, 8(2).

Patel, R., Sulzberger, L., Li, G., Mair, J., Morley, H., Shing, M. N., ... & Sharpe, C. . . (2015). Smartphone apps for weight loss and smoking cessation: Quality ranking of 120 apps. NZ Med J, 128(1421), 73-76.

Payne, H. E., Lister, C., West, J. H., & Bernhardt, J. M. (2015). Behavioral functionality of mobile apps in health interventions: a systematic review of the literature. JMIR mHealth and uHealth, 3(1), e20.

Peng, Y., Shi, J., Fantinato, M., & Chen, J. (2017). A study on the author collaboration network in big data. Information Systems Frontiers, 19(6), 1329-1342.

Perianes-Rodriguez, A., Waltman, L., & Van Eck, N. J. (2016). Constructing bibliometric networks: A comparison between full and fractional counting. Journal of Informetrics, 10(4), 1178-1195.

Pereira-Azevedo, N., Osorio, L., Cavadas, V., Fraga, A., Carrasquinho, E., Cardoso de Oliveira, E., . . . Roobol, M. J. (2016). Expert Involvement Predicts mHealth App Downloads: Multivariate Regression Analysis of Urology Apps. JMIR Mhealth Uhealth, 4(3), e86. doi:10.2196/mhealth.5738

Petronio, S., & Child, J. T. (2020). Conceptualization and operationalization: utility of communication privacy management theory. Curr Opin Psychol, 31, 76-82

196

Pinheiro, M., Serra, M., & Pereira-Azevedo, N. (2019). Predictors of the Number of Installs in Psychiatry Smartphone Apps: Systematic Search on App Stores and Content Analysis. JMIR mental health, 6(11), e15064.

Pujol, E., Hohlfeld, O., & Feldmann, A. (2015, October). Annoyed users: Ads and ad- block usage in the wild. In Proceedings of the 2015 Internet Measurement Conference (pp. 93-106). Puthal, D. (2019). Lattice-modeled information flow control of big sensing data streams for smart health application. IEEE Internet of Things Journal, 6(2), 1312-1320.

Quinn, C. C., Clough, S. S., Minor, J. M., Lender, D., Okafor, M. C., & Gruber-Baldini, A. (2008). WellDoc™ mobile diabetes management randomized controlled trial: change in clinical and behavioral outcomes and patient and physician satisfaction. Diabetes technology & therapeutics, 10(3), 160-168.

Raghupathi, W. & Raghupathi, V. (2013). An Overview of Health Analytics. Journal of Health Medical Information, 4(132), 2.

Raghupathi, W., & Raghupathi, V. (2014). Big data analytics in healthcare: promise and potential. Health information science and systems, 2(1), 3.

Raghupathi, V., & Raghupathi, W. (2014). An Unstructured Information Management Architecture Approach to Text Analytics of Cancer Blogs. International Journal of Healthcare Information Systems and Informatics (IJHISI), 9(2), 16-33.

Rajkomar, A., Dean, J., & Kohane, I. (2019). Machine learning in medicine. New England Journal of Medicine, 380(14), 1347-1358.

Regulation, G. D. P. (2016). Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46. Official Journal of the European Union (OJ), 59(1-88), 294.

Research 2 Guidance. (2016). mHealth Economics 2016 – Current Status and Trends of the mHealth App Market

Rich, E., & Miah, A. (2017). Mobile, wearable and ingestible health technologies: towards a critical research agenda. Health sociology review, 26(1), 84-97.

Riemer, K., Ciriello, R., Peter, S., & Schlagwein, D. (2020). Digital contact-tracing adoption in the COVID-19 pandemic: IT governance for collective action at the societal level. European Journal of Information Systems, 1-15

197

Robinson, E., Higgs, S., Daley, A. J., Jolly, K., Lycett, D., Lewis, A., et al. (2013). Development and feasibility testing of a smart phone based attentive eating intervention. BMC public health, 13(1), 639

Rodrigues Jr, J. F., Paulovich, F. V., de Oliveira, M. C., & de Oliveira Jr, O. N. (2016). On the convergence of nanotechnology and Big Data analysis for computer-aided diagnosis. Nanomedicine, 11(8), 959-982.

Royston, G. (2013). Operational Research for the Real World: big questions from a small island. Journal of the Operational Research Society, 64(6), 793-804.

Rumbold, J. M. (2017). The Effect of the General Data Protection Regulation on Medical Research. Journal of Medical Internet Research, 19(2). Ruotsalainen, P. (2014). Trust Information and Privacy Policies - Enablers for pHealth and Ubiquitous Health. . Studies in Health Technology and Informatics., 200, 133-139.

Sadegh, S. S., Saadat, P. K., Sepehri, M. M., & Assadi, V. (2018). A framework for m- health service development and success evaluation. International journal of medical informatics, 112, 123-130.

Sahoo, S. S., Wei, A., Valdez, J., Wang, L., Zonjy, B., Tatsuoka, C., ... & Lhatoo, S. D. (2016). NeuroPigPen: A Scalable Toolkit for Processing Electrophysiological Signal Data in Neuroscience Applications Using Apache Pig. Frontiers in neuroinformatics, 10, 18.

Salazar, A., de Sola, H., Failde, I., & Moral-Munoz, J. A. (2018). Measuring the quality of mobile apps for the management of pain: systematic search and evaluation using the mobile app rating scale. JMIR mHealth and uHealth, 6(10), e10718.

Sarker, I. H., Hoque, M. M., Uddin, M. K., & Alsanoosy, T. (2020). Mobile Data Science and Intelligent Apps: Concepts, AI-Based Modeling and Research Directions. Mobile Networks and Applications, 1-19.

Schnitzer, M. E., & Blais, L. (2018). Methods for the assessment of selection bias in drug safety during pregnancy studies using electronic medical data. Pharmacology research & perspectives, 6(5), e00426.

Scherer, R., Siddiq, F., & Tondeur, J. (2019). The technology acceptance model (TAM): A meta-analytic structural equationmodeling approach to explaining teachers’ adoption of digital technology in education. Computers & Education, 128, 13-35. doi:10.1016/

Schuler, A., Callahan, A., Jung, K., & Shah, N. H. (2018). Performing an informatics consult: methods and challenges. Journal of the American College of Radiology, 15(3), 563-568.

198

Schulman, K. A., & Richman, B. D. (2019). Toward an effective innovation agenda. The New England journal of medicine, 380(10), 900-901.

Sedrati, H., Nejjari, C., Chaqsare, S., & Ghazal, H. (2016). Mental and physical mobile health apps. Procedia Computer Science, 100, 900-906.

Shameer, K., Johnson, K. W., Glicksberg, B. S., Dudley, J. T., & Sengupta, P. P. (2018). Machine learning in cardiovascular medicine: are we there yet?. Heart, 104(14), 1156-1164.

Shankar, A., & Kumari, P. (2019). A Study of Factors Affecting Mobile Governance (mGov) Adoption Intention in India using an Extension of the Technology Acceptance Model (TAM). South Asian Journal of Management, 26(4), 71-94.

Sharma, R. S., Conrath, D. W., & Dilts, D. M. (1991). A socio-technical model for deploying expert systems. I. The general theory. IEEE Transactions on Engineering Management, 38(1), 14-23.

Sheng, J., Amankwah-Amoah, J., & Wang, X. (2019). Technology in the 21st century: New challenges and opportunities. Technological Forecasting and Social Change, 143, 321-335.

Shilton, K., & Greene, D. (2017). Linking Platforms, Practices, and Developer Ethics: Levers for Privacy Discourse in Mobile Application Development. Journal of Business Ethics, 155(1), 131-146, doi:10.1007/s10551-017-3504-8.

Shortliffe, E. H., & Sepúlveda, M. J. (2018). Clinical decision support in the era of artificial intelligence. Jama, 320(21), 2199-2200.

Sir, M. Y., Dundar, B., Steege, L. M. B., & Pasupathy, K. S. (2015). Nurse–patient assignment models considering patient acuity metrics and nurses’ perceived workload. Journal of biomedical informatics, 55, 237-248

Sivarajah, U., Irani, Z., & Weerakkody, V. (2015). Evaluating the use and impact of Web 2.0 technologies in local government. Government Information Quarterly, 32(4), 473-487.

Siuly, S., & Zhang, X. (2020). Guest Editorial: Special issue on "Application of artificial intelligence in health research". Health Inf Sci Syst, 8(1), 1, doi:10.1007/s13755-019-0089-x.

Soltanisehat, L., Alizadeh, R., Hao, H., & Choo, K.-K. R. (2020). Technical, Temporal, and Spatial Research Challenges and Opportunities in Blockchain-Based Healthcare: A Systematic Literature Review. IEEE Transactions on Engineering Management.

199

Srinivasan, R., & Swink, M. (2018). An investigation of visibility and flexibility as complements to supply chain analytics: An organizational information processing theory perspective. Production and Operations Management, 27(10), 1849-1867.

Stoyanov, S. R., Hides, L., Kavanagh, D. J., & Wilson, H. (2016). Development and Validation of the User Version of the Mobile Application Rating Scale (uMARS). JMIR Mhealth Uhealth, 4(2), e72. doi:10.2196/mhealth.5849

Stoyanov, S. R., Hides, L., Kavanagh, D. J., Zelenko, O., Tjondronegoro, D., & Mani, M. (2015). Mobile app rating scale: a new tool for assessing the quality of health mobile apps. JMIR mHealth and uHealth, 3(1), e27.

Strauß, S. (2015). Datafication and the Seductive Power of Uncertainty—A Critical Exploration of Big Data Enthusiasm. Information, 6(4), 836-847.

Sumarsono, M. Anshari and M. N. Almunawar, "Big Data in Healthcare for Personalization & Customization of Healthcare Services," 2019 International Conference on Information Management and Technology (ICIMTech), Jakarta/Bali, Indonesia, 2019, pp. 73-77

Sun, Y., & Upadhyaya, S. (2015). Secure and privacy preserving data processing support for active authentication. Information Systems Frontiers, 17(5), 1007-1015.

Szlezák, N., Evers, M., Wang, J., & Pérez, L. (2014). The role of big data and advanced analytics in drug discovery, development, and commercialization. Clinical Pharmacology & Therapeutics, 95(5), 492-495.

Thouin, M. F., Hoffman, J. J., & Ford, E. W. (2008). The effect of information technology investment on firm-level performance in the health care industry. Health Care Management Review, 33(1), 60-68.

Toerper, M. F., Flanagan, E., Siddiqui, S., Appelbaum, J., Kasper, E. K., & Levin, S. (2015). Cardiac catheterization laboratory inpatient forecast tool: a prospective evaluation. Journal of the American Medical Informatics Association, 23(e1), e49-e57.

Toews, M., Wachinger, C., Estepar, R. S. J., & Wells, W. M. (2015, June). A feature- based approach to big data analysis of medical images. In International Conference on Information Processing in Medical Imaging (pp. 339-350). Springer, Cham.

Turaga, D. S. (2018). Introduction to the Interfaces Special Issue: Applications of Analytics and Operations Research in Big Data Analysis. Interfaces, 48(2), 93-93.

200 ur Rehman, M. H., Chang, V., Batool, A., & Wah, T. Y. (2016). Big data reduction framework for value creation in sustainable enterprises. International Journal of Information Management, 36(6), 917-928.

Valletti, T., & Wu, J. (2020). Consumer Profiling with Data Requirements: Structure and Policy Implications. Production and Operations Management, 29(2), 309-329.

Vaishya, R., Javaid, M., Khan, I. H., & Haleem, A. (2020). Artificial Intelligence (AI) applications for COVID-19 pandemic. Diabetes & Metabolic Syndrome: Clinical Research & Reviews. van Der Sype, Y. S., & Maalej, W. On lawful disclosure of personal user data: What should app developers do? In 2014 IEEE 7th International Workshop on Requirements Engineering and Law (RELAW), 2014 (pp. 25-34): IEEE van Eck, N. J., & Waltman, L. (2017). Citation-based clustering of publications using CitNetExplorer and VOSviewer. Scientometrics, 111(2), 1053-1070. van Haasteren, A., Gille, F., Fadda, M., & Vayena, E. (2019). Development of the mHealth App Trustworthiness checklist. Digit Health, 5 van Poucke, S., Zhang, Z., Schmitz, M., Vukicevic, M., Vander Laenen, M., Celi, L. A., & De Deyne, C. (2016). Scalable predictive analysis in critically ill patients using a visual open data analysis platform. PloS one, 11(1), e0145791

Velasco, E., Agheneza, T., Denecke, K., Kirchner, G., & Eckmanns, T. (2014). Social media and internet‐based data in global systems for public health surveillance: a systematic review. The Milbank Quarterly, 92(1), 7-33.

Velsko, S. & Bates, T. (2016). A Conceptual Architecture for National Biosurveillance: Moving Beyond Situational Awareness to Enable Digital Detection of Emerging Threats. Health Security, 14(3), 189-201.

Ventola, C. L. (2014). Mobile devices and apps for health care professionals: uses and benefits. Pharmacy and Therapeutics, 39(5), 356.

Venkatesh, V., Speier, C., & Morris, M. G. (2002). User acceptance enablers in individual decision making about technology: Toward an integrated model. Decision Sciences, 33(2), 297-316 Vesselkov, A., Hämmäinen, H., & Töyli, J. (2018). Technology and value network evolution in telehealth. Technological Forecasting and Social Change, 134, 207-222.

Vidgen, M. Shaw, S. & Grant, D. B. (2017) Management challenges in creating value from business analytics. European Journal of Operational Research, 261(2), 626–639.

201

Voisin, S., Pinto, F., Morin Ducote, G., Hudson, K. B., & Tourassi, G. D. (2013). Predicting diagnostic error in radiology via eye tracking and image analytics: Preliminary investigation in mammography. Medical Physics, 40(10).

Wahl, B., Cossy-Gantner, A., Germann, S., & Schwalbe, N. R. (2018). Artificial intelligence (AI) and global health: how can AI contribute to health in resource-poor settings? BMJ Glob Health, 3(4), e000798, doi:10.1136/bmjgh-2018-000798.

Waller, M. A. & Fawcett, S. E. (2013). Data Science, Predictive Analytics, and Big Data: A Revolution That Will Transform Supply Chain Design and Management. Journal of Business Logistics, 34(2), 77–84.

Walsh, G., Shiu, E., Hassan, L., Hille, P., & Takahashi, I. (2019). Fear of online consumer identity theft: Cross-country application and short scale development. Information Systems Frontiers, 21(6), 1251-1264.

Wamba, S. F., Anand, A., & Carter, L. (2013). A literature review of RFID-enabled healthcare applications and issues. International Journal of Information Management, 33(5), 875-891.

Wamba, S. F., Akter, S., Edwards, A., Chopin, G., & Gnanzou, D. (2015). How ‘big data’can make big impact: Findings from a systematic review and a longitudinal case study. International Journal of Production Economics, 165, 234-246.

Wamba, S. F., Gunasekaran, A., Akter, S., Ren, S. J. F., Dubey, R., & Childe, S. J. (2017). Big data analytics and firm performance: Effects of dynamic capabilities. Journal of Business Research, 70, 356-365.

Wang, G., Gunasekaran, A., Ngai, E. W., & Papadopoulos, T. (2016). Big data analytics in logistics and supply chain management: Certain investigations for research and applications. International Journal of Production Economics, 176, 98-110.

Wang, Y., Kung, L., Wang, W. Y. C., & Cegielski, C. G. (2018). An integrated big data analytics-enabled transformation model: Application to health care. Information & Management, 55(1), 64-79.

Wang, J., Cao, B., Yu, P., Sun, L., Bao, W., & Zhu, X. (2018) Deep learning towards mobile applications. In 2018 IEEE 38th International Conference on Distributed Computing Systems (ICDCS), pp. 1385-1393): IEEE

Wang, J., Wang, Y., Wei, C., Yao, N., Yuan, A., Shan, Y., et al. (2014). Smartphone interventions for long-term health management of chronic diseases: an integrative review. Telemedicine and e-Health, 20(6), 570-583.

202

Wang, Y., Xiong, M., & Olya, H. (2020) Toward an Understanding of Responsible Artificial Intelligence Practices. In Proceedings of the 53rd Hawaii International Conference on System Sciences

Ward, M. J., Marsolo, K. A., & Froehle, C. M. (2014). Applications of business analytics in healthcare. Business horizons, 57(5), 571-582.

Warrington, L., Absolom, K., & Velikova, G. (2015). Integrated care pathways for cancer survivors–a role for patient-reported outcome measures and health informatics. Acta Oncologica, 54(5), 600-608.

West, V. L., Borland, D., & Hammond, W. E. (2014). Innovative information visualization of electronic health record data: a systematic review. Journal of the American Medical Informatics Association, 22(2), 330-339.

West, J. H., Hall, P. C., Hanson, C. L., Barnes, M. D., Giraud-Carrier, C., & Barrett, J. (2012). There’s an app for that: content analysis of paid health and fitness apps. Journal of medical Internet research, 14(3), e72.

Wilbanks, B. A. & Langford, P. A. (2014). A review of dashboards for data analytics in nursing. Computers, Informatics, Nursing: CIN, 32(11), 545-9.

Wills, M. J. (2014). Decisions through data: Analytics in healthcare. Journal of Healthcare Management, 59(4), 254-262.

Wisniewski, H., Liu, G., Henson, P., Vaidyam, A., Hajratalli, N. K., Onnela, J.-P., et al. (2019). Understanding the quality, effectiveness and attributes of top-rated smartphone health apps. Evidence-based mental health, 22(1), 4-9.

Woods, M., Paulus, T., Atkins, D. P., & Macklin, R. (2016). Advancing qualitative research using qualitative data analysis software (QDAS)? Reviewing potential versus practice in published studies using ATLAS. ti and NVivo, 1994–2013. Social Science Computer Review, 34(5), 597-617.

Wu, W., Dasgupta, S., Ramirez, E. E., Peterson, C., & Norman, G. J. (2012). Classification accuracies of physical activities using smartphone motion sensors. Journal of medical Internet research, 14(5), e130.

Yaacoub, J.-P. A., Noura, M., Noura, H. N., Salman, O., Yaacoub, E., Couturier, R., et al. (2020). Securing internet of medical things systems: Limitations, issues and recommendations. Future Generation Computer Systems, 105, 581-606.

203

Yanambaka, V. P., Mohanty, S. P., Kougianos, E., & Puthal, D. (2019). ( Pmsec: Physical unclonable function-based robust and lightweight authentication in the internet of medical things. IEEE Transactions on Consumer Electronics, 65(3), 388-397.

Yao, Q. A., Zheng, H., Xu, Z. Y., Wu, Q., Li, Z. W., & Lifen, Y. (2014). Massive medical images retrieval system based on Hadoop. Journal of Multimedia, 9(2), 216.

Yaqoob, I., Hashem, I. A. T., Gani, A., Mokhtar, S., Ahmed, E., Anuar, N. B., & Vasilakos, A. V. (2016). Big data: From beginning to future. International Journal of Information Management, 36(6), 1231-1247.

Ye, T., Xue, J., He, M., Gu, J., Lin, H., Xu, B., & Cheng, Y. (2019). Psychosocial factors affecting artificial intelligence adoption in health care in China: Cross-sectional study. Journal of medical Internet research, 21(10), e14316.

Yildirim, P., Majnarić, L., Ekmekci, O. I., & Holzinger, A. (2014). Knowledge discovery of drug data on the example of adverse reaction prediction. BMC bioinformatics, 15(6), S7.

Yin, S.& Kaynak, O. (2015). Big Data for Modern Industry: Challenges and Trends, Proceedings of the IEEE, 103(2), 143-146.

You, Q., Shiaofen, F. & Y. C. Jake, Y. C. (2008). Gene Terrain: Visual exploration of differential gene expression profiles organized in native biomolecular interaction networks, Information Visualization, 9(1), 1-12.

Zakhem, G. A., Motosko, C. C., & Ho, R. S. (2018). How should artificial intelligence screen for skin cancer and deliver diagnostic predictions to patients? JAMA dermatology, 154(12), 1383-1384.

Zapata, B. C., Fernández-Alemán, J. L., Idri, A., & Toval, A. (2015). Empirical studies on usability of mHealth apps: a systematic literature review. Journal of medical systems, 39(2), 1.

Zarate, O. A., Brody, J. G., Brown, P., Ramirez‐Andreotta, M. D., Perovich, L., & Matz, J. (2016). Balancing benefits and risks of immortal data: Participants’ views of open consent in the Personal Genome Project. Hastings Center Report, 46(1), 36-45.

Zhang, Y., Guo, S. L., Han, L. N., & Li, T. L. (2016). Application and exploration of big data mining in clinical medicine. Chinese Medical Journal, 129(6), 731.

Zhang, X., Yang, L. T., Liu, C., & Chen, J. (2014). A scalable two-phase top-down specialization approach for data anonymization using mapreduce on cloud. IEEE Transactions on Parallel and Distributed Systems, 25(2), 363-373.

204

Zhang, Y., Sun, Y., & Xie, B. (2015). Quality of health information for consumers on the web: a systematic review of indicators, criteria, tools, and evaluation results. Journal of the Association for Information Science and Technology, 66(10), 2071-2084.

Zhang, Y., & Li, X. (2017). Uses of information and communication technologies in HIV self-management: A systematic review of global literature. International Journal of Information Management, 37(2), 75-83.

Zhu, K., Lou, Z., Zhou, J., Ballester, N., Kong, N., & Parikh, P. (2015). Predicting 30- day Hospital Readmission with Publicly Available Administrative Database. Methods of Information in Medicine, 54(06), 560-567.

Zimmer, M., Kumar, P., Vitak, J., Liao, Y., & Chamberlain Kritikos, K. (2020). ‘There’s nothing really they can do with this information’: Unpacking how users manage privacy boundaries for personal fitness information. Information, Communication & Society, 23(7), 1020-1037.

Ziuzianski, P., Furmankiewicz, M. & Soltysik-Piorunkiewicz, A.(2014). E-health artificial intelligence system implementation: Case study of knowledge management dashboard of epidemiological data in Poland. International Journal of Biology and Biomedical Engineering, 8, 164-171.

Zhao, J., Freeman, B., & Li, M. (2016). Can mobile phone apps influence people’s health behavior change? An evidence review. Journal of medical Internet research, 18(11), e287.

Zulman, D. M., Shah, N. H., & Verghese, A. (2016). Evolutionary pressures on the electronic health record: caring for complexity. Jama, 316(9), 923-924.

APPENDIX

List of 804 articles 1. Abbas, A., Ali, M., Khan, M. U. S., & Khan, S. U. (2016). Personalized healthcare cloud services for disease risk assessment and wellness management using social media. Pervasive and Mobile Computing, 28, 81-99. 2. Abbas, A., Bilal, K., Zhang, L., & Khan, S. U. (2015). A cloud based health insurance plan recommendation system: A user centered approach. Future Generation Computer Systems, 43, 99-109. 3. AbdelRahman, S. E., Zhang, M., Bray, B. E., & Kawamoto, K. (2014). A three-step approach for the derivation and validation of high-performing predictive models using an

205

operational dataset: congestive heart failure readmission case study. BMC medical informatics and decision making, 14(1), 41. 4. Adamusiak, T., Shimoyama, N., & Shimoyama, M. (2014). Next generation phenotyping using the unified medical language system. JMIR Med Inform, 2(1), e5. 5. Adler, M., & Spengler, M. (2009). Novel strategies and tools for enhanced sensitivity in routine biomolecule analytics. Current Pharmaceutical Analysis, 5(4), 390-407. 6. Agneeswaran, V. S., Mukherjee, J., Gupta, A., Tonpay, P., Tiwari, J., & Agarwal, N. (2013). Real-Time Analytics for the Healthcare Industry: Arrhythmia Detection. Big Data, 1(3), 176-182. 7. Agnoletti, V., Buccioli, M., Padovani, E., Corso, R. M., Perger, P., Piraccini, E., ...& Vicini, C. (2013). Operating room data management: improving efficiency and safety in a surgical block. BMC surgery, 13(1), 7. 8. Ahmad, M., Amin, M. B., Hussain, S., Kang, B. H., Cheong, T., & Lee, S. (2016). Health Fog: a novel framework for health and wellness applications. The Journal of Supercomputing, 72(10), 3677-3695. 9. Ajorlou, S., Shams, I., & Yang, K. (2015). An analytics approach to designing patient centered medical homes. Health care management science, 18(1), 3-18. 10. Al Kazzi, E. S., & Hutfless, S. (2015). Better big data. Expert Rev Pharmacoecon Outcomes Res, 15(6), 873-876. 11. Al Mallah, M. H., Keteyian, S. J., Brawner, C. A., Whelton, S., & Blaha, M. J. (2014). Rationale and design of the Henry Ford Exercise Testing Project (the FIT project). Clinical cardiology, 37(8), 456-461. 12. Albrecht, H. H. (2016). Can Big Data Analyses Help Speed Up the Clinical Development of Mucoactive Drugs for Symptomatic RTIs?.Lung, 194(1), 31-34. 13. Alemayehu, D., & Berger, M. L. (2016). Big Data: transforming drug development and health policy decision making. Health Services and Outcomes Research Methodology, 16(3), 92-102 14. Ali, M. A., Ahsan, Z., Amin, M., Latif, S., Ayyaz, A., & Ayyaz, M. N. (2016). ID-Viewer: a visual analytics architecture for infectious diseases surveillance and response management in Pakistan. Public health, 134, 72-85. 15. Allen, G. I., Amoroso, N., Anghel, C., Balagurusamy, V., Bare, C. J., Beaton, D., ...& Caberlotto, L. (2016). Crowdsourced estimation of cognitive decline and resilience in Alzheimer's disease. Alzheimer's & Dementia, 12(6), 645-653. 16. Almashaqbeh, G., Hayajneh, T., Vasilakos, A. V., & Mohd, B. J. (2014). QoS-aware health monitoring system using cloud-based WBANs. Journal of medical systems, 38(10), 121. 17. Al-Shaqi, R., Mourshed, M., & Rezgui, Y. (2016). Progress in ambient assisted systems for independent living by the elderly. SpringerPlus, 5(1), 624. 18. Althari, S., & Gloyn, A. L. (2015). When is it MODY? Challenges in the Interpretation of Sequence Variants in MODY Genes. The review of diabetic studies: RDS, 12(3-4), 330. 19. Althebyan, Q., Yaseen, Q., Jararweh, Y., & Al-Ayyoub, M. (2016). Cloud support for large scale e-healthcare systems. Annals of Telecommunications, 71(9-10), 503-515. 20. Alyass, A., Turcotte, M., & Meyre, D. (2015). From big data analysis to personalized medicine for all: challenges and opportunities. BMC medical genomics, 8(1), 33. 21. Amankwah-Amoah, J. (2016). Emerging economies, emerging challenges: Mobilising and capturing value from big data. Technological Forecasting and Social Change, 110, 167- 174. 22. Amarasingham, R., Patzer, R. E., Huesch, M., Nguyen, N. Q., & Xie, B. (2014). Implementing electronic health care predictive analytics: considerations and challenges. Health Affairs, 33(7), 1148-1154. 23. Anderson, J. E., & Chang, D. C. (2015). Using electronic health records for surgical quality improvement in the era of big data. JAMA surgery, 150(1), 24-29. 24. Andreu-Perez, J., Poon, C. C., Merrifield, R. D., Wong, S. T., & Yang, G. Z. (2015). Big data for health. IEEE J Biomed Health Inform, 19(4), 1193-1208.

206

25. Androwich, I. M. (2013). Nursing as a learning discipline: A call to action. Nursing science quarterly, 26(1), 37-41. 26. Angelelli, P., Oeltze, S., Turkay, C., Haasz, J., Hodneland, E., Lundervold, A., ...& Hauser, H. (2014). Interactive visual analysis of heterogeneous cohort study data. IEEE computer graphics and applications, (1), 1-1. 27. Angulo, D. A., Schneider, C., Oliver, J. H., Charpak, N., & Hernandez, J. T. (2016). A multi-facetted visual analytics tool for exploratory analysis of human brain and function datasets. Frontiers in neuroinformatics, 10, 36. 28. Anneke Fitzgerald, J., & Dadich, A. (2009). Using visual analytics to improve hospital scheduling and patient flow. Journal of theoretical and applied electronic commerce research, 4(2), 20-30. 29. Anoushiravani, A. A., Patton, J., Sayeed, Z., El-Othmani, M. M., & Saleh, K. J. (2016). Big data, big research: implementing population health-based research models and integrating care to reduce cost and improve outcomes. Orthopedic Clinics, 47(4), 717-724. 30. Antman, E. M., Benjamin, E. J., Harrington, R. A., Houser, S. R., Peterson, E. D., Bauman, M. A., ... & Daugherty, A. (2015). Acquisition, analysis, and sharing of data in 2015 and beyond: a survey of the landscape: a conference report from the American Heart Association Data Summit 2015. Journal of the American Heart Association, 4(11), e002810. 31. Arsiwalla, X. D., Zucca, R., Betella, A., Martinez, E., Dalmazzo, D., Omedas, P., ...& Verschure, P. F. (2015). Network dynamics with BrainX3: a large-scale simulation of the human brain network with real-time interaction. Frontiers in neuroinformatics, 9, 2. 32. Ashrafi, N., Kelleher, L., & Kuilboer, J. P. (2014). The impact of business intelligence on healthcare delivery in the USA. Interdisciplinary Journal of Information, Knowledge, and Management, 9, 117-130. 33. Auffray, C., Balling, R., Barroso, I., Bencze, L., Benson, M., Bergeron, J., ...& Del Signore, S. (2016). Making sense of big data in health research: towards an EU action plan. Genome medicine, 8(1), 71. 34. Austin, C., & Kusumoto, F. (2016). The application of Big Data in medicine: current implications and future directions. J Interv Card Electrophysiol, 47(1), 51-59. doi: 10.1007/s10840-016-0104-y 35. Azadmanjir, Z., Torabi, M., Safdari, R., Bayat, M., & Golmahi, F. (2015). A map for clinical laboratories management indicators in the intelligent dashboard. Acta Informatica Medica, 23(4), 210. 36. Azmak, O., Bayer, H., Caplin, A., Chun, M., Glimcher, P., Koonin, S., & Patrinos, A. (2015). Using Big data to understand the human condition: the kavli HUMAN project. Big data, 3(3), 173-188. 37. Badgeley, M. A., Shameer, K., Glicksberg, B. S., Tomlinson, M. S., Levin, M. A., McCormick, P. J.,& Dudley, J. T. (2016). EHDViz: clinical dashboard development using open-source technologies. BMJ open, 6(3), e010579. 38. Baldwin, J. N., Bootman, J. L., Carter, R. A., Crabtree, B. L., Piascik, P., Ekoma, J. O., & Maine, L. L. (2015). Pharmacy practice, education, and research in the era of big data: 2014-15 Argus Commission Report. American Journal of Pharmaceutical Education, 79(10), S26. 39. Banaee, H., & Loutfi, A. (2015). Data-Driven Rule Mining and Representation of Temporal Patterns in Physiological Sensor Data. IEEE J. Biomedical and Health Informatics, 19(5), 1557-1566. 40. Banos, O., Amin, M. B., Khan, W. A., Afzal, M., Hussain, M., Kang, B. H., & Lee, S. (2016). The Mining Minds digital health and wellness framework. Biomedical engineering online, 15(1), 76. 41. Bardhan, I., Oh, J. H., Zheng, Z., & Kirksey, K. (2014). Predictive analytics for readmission of patients with congestive heart failure. Information Systems Research, 26(1), 19-39. 42. Barkhordari, M., & Niamanesh, M. (2015). ScaDiPaSi: an effective scalable and distributable MapReduce-Based method to find patient similarity on huge healthcare networks. Big Data Research, 2(1), 19-27.

207

43. Barkley, R., Greenapple, R., & Whang, J. (2014). Actionable data analytics in oncology: Are we there yet? Journal of oncology practice, 10(2), 93-96. 44. Barrett, J. S., Mondick, J. T., Narayan, M., Vijayakumar, K., & Vijayakumar, S. (2008). Integration of modeling and simulation into hospital-based decision support systems guiding pediatric pharmacotherapy. BMC medical informatics and decision making, 8(1), 6. 45. Barrett, M. A., Humblet, O., Hiatt, R. A., & Adler, N. E. (2013). Big data and disease prevention: from quantified self to quantified communities. Big data, 1(3), 168-175. 46. Basole, R. C., Braunstein, M. L., Kumar, V., Park, H., Kahng, M., Chau, D. H., ...& Lesnick, B. (2015). Understanding variations in pediatric asthma care processes in the emergency department using visual analytics. Journal of the American Medical Informatics Association, 22(2), 318-323. 47. Basu, R. K., Gist, K., & Wheeler, D. S. (2015). Improving acute kidney injury diagnostics using predictive analytics. Current opinion in critical care, 21(6), 473-478. 48. Batarseh, F. A., & Latif, E. A. (2016). Assessing the quality of service using Big Data analytics: With application to healthcare. Big Data Research, 4, 13-24. 49. Bates, D. W., Saria, S., Ohno-Machado, L., Shah, A., & Escobar, G. (2014). Big data in health care: using analytics to identify and manage high-risk and high-cost patients. Health Affairs, 33(7), 1123-1131. 50. Baum, D. (2015). An intelligent patient focus. Health Management Technology, 12-16. 51. Behadada, O., Trovati, M., Chikh, M. A., & Bessis, N. (2016). Big data‐based extraction of fuzzy partition rules for heart arrhythmia detection: a semi‐automated approach. Concurrency and Computation: Practice and Experience, 28(2), 360-373. 52. Beim, P. Y., Elashoff, M., & Hu-Seliger, T. T. (2013). Personalized reproductive medicine on the brink: progress, opportunities and challenges ahead. Reproductive biomedicine online, 27(6), 611-623. 53. Bell, E., Campbell, S., & Goldberg, L. R. (2015). Nursing identity and patient-centredness in scholarly health services research: a computational text analysis of PubMed abstracts 1986-2013. BMC Health Serv Res, 15, 3. 54. Belle, A., Thiagarajan, R., Soroushmehr, S. M., Navidi, F., Beard, D. A., & Najarian, K. (2015). Big data analytics in healthcare. BioMed research international, 2015. 55. Bello-Orgaz, G., Hernandez-Castro, J., & Camacho, D. (2017). Detecting discussion communities on vaccination in twitter. Future Generation Computer Systems, 66, 125-136. 56. Bello-Orgaz, G., Hernandez-Castro, J., & Camacho, D. (2015). A survey of social web mining applications for disease outbreak detection. In Intelligent Distributed Computing VIII (pp. 345-356). Springer, Cham. 57. Ben-Ari Fuchs, S., Lieder, I., Stelzer, G., Mazor, Y., Buzhor, E., Kaplan, S., ...& Kohn, A. (2016). GeneAnalytics: an integrative gene set analysis tool for next generation sequencing, RNAseq and microarray data. Omics: a journal of integrative biology, 20(3), 139-151. 58. Benharref, A., Serhani, M. A., & Nujum, A. R. (2014, August). Closing the loop from continuous M-health monitoring to fuzzy logic-based optimized recommendations. In Engineering in Medicine and Biology Society (EMBC), 2014 36th Annual International Conference of the IEEE (pp. 2698-2701). IEEE. 59. Bereman, M. S. (2015). Tools for monitoring system suitability in LC MS/MS centric proteomic experiments. Proteomics, 15(5-6), 891-902. 60. Berger, M. L., Curtis, M. D., Smith, G., Harnett, J., & Abernethy, A. P. (2016). Opportunities and challenges in leveraging electronic health record data in oncology. Future Oncology, 12(10), 1261-1274. 61. Berger, M. L., & Doban, V. (2014). Big data, advanced analytics and the future of comparative effectiveness research. Journal of Comparative Effectiveness Research, 3(2), 167-176. 62. Berger, M. L., Lipset, C., Gutteridge, A., Axelsen, K., Subedi, P., & Madigan, D. (2015). Optimizing the leveraging of real-world data to improve the development and use of medicines. Value Health, 18(1), 127-130. doi: 10.1016/j.jval.2014.10.009

208

63. Berler, A., Pavlopoulos, S., & Koutsouris, D. (2005). Using key performance indicators as knowledge-management tools at a regional health-care authority level. IEEE Transactions on Information Technology in Biomedicine, 9(2), 184-192. 64. Bernard, J., Sessler, D., May, T., Schlomm, T., Pehrke, D., & Kohlhammer, J. (2015). A visual-interactive system for prostate cancer cohort analysis. IEEE computer graphics and applications, 35(3), 44-55. 65. Bertsimas, D., O’Hair, A., Relyea, S., & Silberholz, J. (2016). An analytics approach to designing combination chemotherapy regimens for cancer. Management Science, 62(5), 1511-1531. 66. Bettencourt-Silva, J. H., Clark, J., Cooper, C. S., Mills, R., Rayward-Smith, V. J., & De La Iglesia, B. (2015). Building data-driven pathways from routinely collected hospital data: a case study on prostate cancer. JMIR medical informatics, 3(3). 67. Beyan, T., & Son, Y. A. (2014). Emerging Technologies in Health Information Systems: Genomics Driven Wellness Tracking and Management System (GO-WELL). In Big Data and Internet of Things: A Roadmap for Smart Environments (pp. 315-339). Springer, Cham. 68. Bhargava, R., & Madabhushi, A. (2016). Emerging themes in image informatics and molecular analysis for digital pathology. Annual review of biomedical engineering, 18, 387- 412. 69. Bhattacharya, I., Ramachandran, A., & Jha, B. K. (2012). Healthcare Data Analytics on the Cloud. Online Journal of Health and Allied Sciences, 11(1 (1)). 70. Bhavnani, S. K., Dang, B., Bellala, G., Divekar, R., Visweswaran, S., Brasier, A. R., & Kurosky, A. (2015). Unlocking proteomic heterogeneity in complex diseases through visual analytics. Proteomics, 15(8), 1405-1418. 71. Bhuvaneshwar, K., Belouali, A., Singh, V., Johnson, R. M., Song, L., Alaoui, A., ...& Madhavan, S. (2016). G-DOC Plus–an integrative bioinformatics platform for precision medicine. BMC bioinformatics, 17(1), 193. 72. Binder, H., & Blettner, M. (2015). Big Data in Medical Science—a Biostatistical View: Part 21 of a Series on Evaluation of Scientific Publications. Deutsches Ärzteblatt International, 112(9), 137. 73. Bjarnadóttir, M. V., Malik, S., Onukwugha, E., Gooden, T., & Plaisant, C. (2016). Understanding adherence and prescription patterns using large-scale claims data. Pharmacoeconomics, 34(2), 169-179. 74. Blakely, T., Atkinson, J., Kvizhinadze, G., Nghiem, N., McLeod, H., Davies, A., & Wilson, N. (2015). Updated New Zealand health system cost estimates from health events by sex, age and proximity to death: further improvements in the age of ‘big data’. NZ Med J, 128(1422), 13-23. 75. Blobel, B., Lopez, D. M., & Gonzalez, C. (2016). Patient privacy and security concerns on big data for personalized medicine. Health and Technology, 6(1), 75-81. 76. Bolouri, H., Zhao, L. P., & Holland, E. C. (2016). Big data visualization identifies the multidimensional molecular landscape of human gliomas. Proceedings of the national academy of sciences, 113(19), 5394-5399. 77. Bose, A., & Das, S. (2012). Trial analytics-a tool for clinical trial management. Acta Poloniae Pharmaceutica-Drug Research, 69(3), 523-33. 78. Boulos, M. N. K., Sanfilippo, A. P., Corley, C. D., & Wheeler, S. (2010). Social Web mining and exploitation for serious applications: Technosocial Predictive Analytics and related technologies for public health, environmental and national security surveillance. Computer methods and programs in biomedicine, 100(1), 16-23. 79. Boverman, G., & Genc, S. (2014, August). Prediction of mortality from respiratory distress among long-term mechanically ventilated patients. In Engineering in Medicine and Biology Society (EMBC), 2014 36th Annual International Conference of the IEEE (pp. 3464-3467). IEEE. 80. Boytcheva, S., Angelova, G., Angelov, Z., & Tcharaktchiev, D. (2015). Text mining and big data analytics for retrospective analysis of clinical texts from outpatient care. Cybernetics and Information Technologies, 15(4), 58-77.

209

81. Bradley, P. (2012). Predictive analytics can support the ACO model: using data to identify trends and patterns can help drive better outcomes. Healthcare Financial Management, 66(4), 102-107. 82. Bradley, P., & Kaplan, J. (2010). Turning hospital data into dollars: healthcare financial executives can use predictive analytics to enhance their ability to capture charges and identify underpayments. Healthcare Financial Management, 64(2), 64-69. 83. Bradley, P. S. (2013). Implications of big data analytics on population health management. Big data, 1(3), 152-159. 84. Bram, J. T., Warwick-Clark, B., Obeysekare, E., & Mehta, K. (2015). Utilization and monetization of healthcare data in developing countries. Big data, 3(2), 59-66. 85. Brandhorst, G., Luthe, H., Domke, I., Knoke, C., Rhode, K. H., Sauter, H., & Oellerich, M. (2009). Therapeutic drug monitoring and drugs of abuse testing on the cobas® 6000 analyzer series: analytical performance under routine-like conditions. Clinical chemistry and laboratory medicine, 47(7), 854-859. 86. Bravo, À., Piñero, J., Queralt-Rosinach, N., Rautschka, M., & Furlong, L. I. (2015). Extraction of relations between genes and diseases from text and large-scale data analysis: implications for translational research. BMC bioinformatics, 16(1), 55. 87. Brennan, N., Oelschlaeger, A., Cox, C., & Tavenner, M. (2014). Leveraging the big-data revolution: CMS is expanding capabilities to spur health system transformation. Health Affairs, 33(7), 1195-1202. 88. Brennan, P. F., & Bakken, S. (2015). Nursing needs big data and big data needs nursing. Journal of Nursing Scholarship, 47(5), 477-484. 89. Brinkmann, B. H., Bower, M. R., Stengel, K. A., Worrell, G. A., & Stead, M. (2009). Large-scale electrophysiology: acquisition, compression, encryption, and storage of big data. Journal of neuroscience methods, 180(1), 185-192. 90. Brooks, M. G. (2008). Getting a handle on analytics for strategic success: in the current economy, it is more crucial than ever for healthcare financial executives to identify and manage the metrics that determine strategic success. Healthcare Financial Management, 62(7), 100-105. 91. Brooks, P., El-Gayar, O., & Sarnikar, S. (2015). A framework for developing a domain specific business intelligence maturity model: Application to healthcare. International Journal of Information Management, 35(3), 337-345. 92. Broughman, J. R., & Chen, R. C. (2016). Using big data for quality assessment in oncology. Journal of comparative effectiveness research, 5(3), 309-319. 93. Brown, R. E., Buryanek, J., Tammisetti, V. S., McGuire, M. F., & Csencsits-Smith, K. (2016). Morphoproteomics and biomedical analytics confirm the mTORC2/Akt pathway as a resistance signature and activated ERK and STAT3 as concomitant prosurvival/antiapoptotic pathways in metastatic renal cell carcinoma (RCC) progressing on rapalogs: Pathogenesis and therapeutic options. Oncotarget, 7(27), 41612. 94. Brown, J. B., Nakatsui, M., & Okuno, Y. (2014). Constructing a Foundational Platform Driven by Japan’s K Supercomputer for Next‐Generation Drug Design. Molecular informatics, 33(11‐12), 732-741. 95. Brown, R. E., & McGuire, M. F. (2012). Oncogenesis recapitulates embryogenesis via the hypoxia pathway: morphoproteomics and biomedical analytics provide proof of concept and therapeutic options. Annals of Clinical & Laboratory Science, 42(3), 243-257. 96. Bruining, N., Caiani, E., Chronaki, C., Guzik, P., & van der Velde, E. (2014). Acquisition and analysis of cardiovascular signals on smartphones: potential, pitfalls and perspectives: by the Task Force of the e-Cardiology Working Group of European Society of Cardiology. European journal of preventive cardiology, 21(2_suppl), 4-13. 97. Bub, A., Kriebel, A., Dörr, C., Bandt, S., Rist, M., Roth, A., ...& Watzl, B. (2016). The Karlsruhe Metabolomics and Nutrition (KarMeN) study: protocol and methods of a cross- sectional study to characterize the metabolome of healthy men and women. JMIR research protocols, 5(3).

210

98. Buchem, I., Merceron, A., Kreutel, J., Haesner, M., & Steinert, A. (2014). Wearable enhanced learning for healthy ageing: conceptual framework and architecture of the ‘Fitness MOOC’. Interact Des Archit J, 24, 111-124. 99. Buell, D. (2013). Leveraging data and analytics to generate new revenue. Healthcare Financial Management, 67(4), 40-44. 100. Bunyavanich, S., & Schadt, E. E. (2015). Systems biology of asthma and allergic diseases: a multiscale approach. Journal of Allergy and Clinical Immunology, 135(1), 31- 42. 101. Bureš, V., Otčenášková, T., Čech, P., & Antoš, K. (2012). A proposal for a computer- based framework of support for public health in the management of biological incidents: the Czech Republic experience. Perspectives in Public Health, 132(6), 292-298. 102. Cai, G., Mahadevan, S. (2016). Big Data Analytics in Online Structural Health Monitoring. International Journal of Prognostics and Health Management. 103. Caie, P. D., Zhou, Y., Turnbull, A. K., Oniscu, A., & Harrison, D. J. (2016). Novel histopathologic feature identified through image analysis augments stage II colorectal cancer clinical reporting. Oncotarget, 7(28) 104. Calabrese, RPh, MHP; Neil B. Minkoff, MD; and Kristine Rawlings, PharmD. (2014). Pharmacosynchrony: Road Map to Transformation in Pharmacy Benefit Management. The American Journal of Pharmacy Benefits. 105. Callery-D’Amico, S., Sam, L. M., Grey, T. H., & Greenwood, D. J. (2016). TransCelerate’s Clinical Quality Management System: Issue Management. Therapeutic Innovation & Regulatory Science, 50(5), 530-535. 106. Calvert, M., Thwaites, R., Kyte, D., & Devlin, N. (2015). Putting patient-reported outcomes on the ‘Big Data Road Map’. Journal of the Royal Society of Medicine, 108(8), 299-303. 107. Calyam, P., Mishra, A., Antequera, R. B., Chemodanov, D., Berryman, A., Zhu, K., ...& Skubic, M. (2016). Synchronous big data analytics for personalized and remote physical therapy. Pervasive and Mobile Computing, 28, 3-20. 108. Canela-Xandri, O., Law, A., Gray, A., Woolliams, J. A., & Tenesa, A. (2015). A new tool called DISSECT for analysing large genomic data sets using a Big Data approach. Nature communications, 6, 10162. 109. Caron, F., Vanthienen, J., Vanhaecht, K., Van Limbergen, E., Deweerdt, J., & Baesens, B. (2014). A process mining-based investigation of adverse events in care processes. Health Information Management Journal, 43(1), 16-25. 110. Carroll, L. N., Au, A. P., Detwiler, L. T., Fu, T. C., Painter, I. S., & Abernethy, N. F. (2014). Visualization and analytics tools for infectious disease epidemiology: a systematic review. Journal of biomedical informatics, 51, 287-298. 111. Carter, T. C., & He, M. M. (2016). Challenges of identifying clinically actionable genetic variants for precision medicine. Journal of healthcare engineering, 2016. 112. Caspers, B. A., & Pickard, B. (2013). Value-based resource management: a model for best value nursing care. Nursing administration quarterly, 37(2), 95-104. 113. Castellani, G. C., Menichetti, G., Garagnani, P., Giulia Bacalini, M., Pirazzini, C., Franceschi, C., ...& Mosca, E. (2015). Systems medicine of inflammaging. Briefings in bioinformatics, 17(3), 527-540. 114. Castellanos, F. X., Di Martino, A., Craddock, R. C., Mehta, A. D., & Milham, M. P. (2013). Clinical applications of the functional connectome. Neuroimage, 80, 527-540. 115. Caster, O., Juhlin, K., Watson, S., & Norén, G. N. (2014). Improved statistical signal detection in pharmacovigilance by combining multiple strength-of-evidence aspects in vigiRank. Drug safety, 37(8), 617-628. 116. Catlin, A. C., Malloy, W. X., Arthur, K. J., Gaston, C., Young, J., Fernando, S., & Fernando, R. (2015). Comparative analytics of infusion pump data across multiple hospital systems. American Journal of Health-System Pharmacy, 72(4), 317-324. 117. Cato, K. D., Bockting, W., & Larson, E. (2016). Did I tell you that? ethical issues related to using computational methods to discover non-disclosed patient characteristics. Journal of Empirical Research on Human Research Ethics, 11(3), 214-219.

211

118. Celi, L. A., Zimolzak, A. J., & Stone, D. J. (2014). Dynamic clinical data mining: search engine-based decision support. JMIR medical informatics, 2(1). 119. Celler, B. G., Sparks, R., Nepal, S., Alem, L., Varnfield, M., Li, J., ...& Jayasena, R. (2014). Design of a multi-site multi-state clinical trial of home monitoring of chronic disease in the community in Australia. BMC public health, 14(1), 1270. 120. Celler, B. G., & Sparks, R. S. (2015). Home telemonitoring of vital signs—Technical challenges and future directions. IEEE journal of biomedical and health informatics, 19(1), 82-91. 121. Chalmers, E., Hill, D., Zhao, V., & Lou, E. (2015). Prescriptive analytics applied to brace treatment for AIS: a pilot demonstration. Scoliosis, 10(2), S13. 122. Chatellier, G., Varlet, V., Blachier-Poisson, C., Beslay, N., Behier, J. M., Braunstein, D., ...& Josseran, A. (2016). “Big data” and “open data”: What kind of access should researchers enjoy?.Therapie, 71(1), 107-114. 123. Chaussabel, D., & Pulendran, B. (2015). A vision and a prescription for big data– enabled medicine. Nature immunology, 16(5), 435. 124. Chawla, N. V., & Davis, D. A. (2013). Bringing big data to personalized healthcare: a patient-centered framework. Journal of general internal medicine, 28(3), 660-665. 125. Chen, B., & Butte, A. J. (2016). Leveraging big data to transform target selection and drug discovery. Clinical Pharmacology & Therapeutics, 99(3), 285-297. 126. Chen, H., Chen, W., Liu, C., Zhang, L., Su, J., & Zhou, X. (2016). Relational Network for Knowledge Discovery through Heterogeneous Biomedical and Clinical Features. Scientific reports, 6, 29915. 127. Chen, H., & Fu, Z. (2015). Hadoop-based healthcare information system design and wireless security communication implementation. Mobile Information Systems, 2015. 128. Chen, J. H., Podchiyska, T., & Altman, R. B. (2015). OrderRex: clinical order decision support and outcome predictions by data-mining electronic medical records. Journal of the American Medical Informatics Association, 23(2), 339-348. 129. Chen, M., Ma, Y., Song, J., Lai, C. F., & Hu, B. (2016). Smart clothing: Connecting human with clouds and big data for sustainable health monitoring. Mobile Networks and Applications, 21(5), 825-845. 130. Chen, P. H., Chen, Y. J., & Cook, T. S. (2015). Capricorn–A Web-Based Automatic Case Log and Volume Analytics for Diagnostic Radiology Residents. Academic radiology, 22(10), 1242-1251. 131. Chen, T. J., & Kotecha, N. (2014). Cytobank: providing an analytics platform for community cytometry data analysis and collaboration. In High-Dimensional Single Cell Analysis (pp. 127-157). Springer, Berlin, Heidelberg. 132. Chen, A. Y., Lu, T. Y., Ma, M. H. M., & Sun, W. Z. (2016). Demand Forecast Using Data Analytics for the Preallocation of Ambulances. IEEE journal of biomedical and health informatics, 20(4), 1178-1187. 133. Chen, Y., & Yang, H. (2014, August). Heterogeneous postsurgical data analytics for predictive modeling of mortality risks in intensive care units. In Engineering in Medicine and Biology Society (EMBC), 2014 36th annual international conference of the IEEE (pp. 4310-4314). IEEE. 134. Chen, Y., Argentinis, J. E., & Weber, G. (2016). IBM Watson: how cognitive computing can be applied to big data challenges in life sciences research. Clinical therapeutics, 38(4), 688-701. 135. Chen, Y., Kao, S. L., Tai, E. S., Wee, H. L., Khoo, E. Y. H., Ning, Y., ... & Tan, C. S. (2016). Utilizing distributional analytics and electronic records to assess timeliness of inpatient blood glucose monitoring in non-critical care wards. BMC medical research methodology, 16(1), 40. 136. Choi, G., Lee, K., Seo, D., Kim, S., Kim, D., & Lee, Y. (2015). Analysis of medical data using the big data and R. In Advances in Computer Science and Ubiquitous Computing (pp. 867-873). Springer, Singapore. 137. Choi, J., Choi, C., Ko, H., & Kim, P. (2016). Intelligent Healthcare Service Using Health Lifelog Analysis. Journal of medical systems, 40(8), 188.

212

138. Choi, J. K., Jeon, K. H., Won, Y., & Kim, J. J. (2015). Application of big data analysis with decision tree for the foot disorder. Cluster Computing, 18(4), 1399-1404. 139. Choucair, B., Bhatt, J., & Mansour, R. (2015). A bright future: innovation transforming public health in Chicago. Journal of Public Health Management and Practice, 21(Suppl 1), S49. 140. Chouvarda, I. G., Goulis, D. G., Lambrinoudaki, I., & Maglaveras, N. (2015). Connected health and integrated care: Toward new models for chronic disease management. Maturitas, 82(1), 22-27. 141. Chowriappa, P., Dua, S., & Todorov, Y. (2014). Introduction to machine learning in healthcare informatics. In Machine Learning in Healthcare Informatics (pp. 1-23). Springer, Berlin, Heidelberg. 142. Christensson, C., Gipson, G., Thomas, T., & Weatherall, J. (2012). Text Analytics for Surveillance (TAS) An Interactive Environment for Safety Literature Review. Drug Information Journal, 46(1), 115-123. 143. Chui, K. K., Wenger, J. B., Cohen, S. A., & Naumova, E. N. (2011). Visual analytics for epidemiologists: understanding the interactions between age, time, and disease with multi-panel graphs. PloS one, 6(2), e14683. 144. Chute, C. G., Beck, S. A., Fisk, T. B., & Mohr, D. N. (2010). The Enterprise Data Trust at Mayo Clinic: a semantically integrated warehouse of biomedical data. Journal of the American Medical Informatics Association, 17(2), 131-135. 145. Clark, A., Ng, J. Q., Morlet, N., & Semmens, J. B. (2016). Big data and ophthalmic research. survey of ophthalmology, 61(4), 443-465. 146. Clark, W. R., Garzotto, F., Neri, M., Lorenzin, A., Zaccaria, M., & Ronco, C. (2016). Data analytics for continuous renal replacement therapy: historical limitations and recent technology advances. The International journal of artificial organs, 39(8), 399-406. 147. Clark, W. R., Garzotto, F., Neri, M., Lorenzin, A., Zaccaria, M., & Ronco, C. (2016). Data analytics for continuous renal replacement therapy: historical limitations and recent technology advances. Int J Artif Organs, 39(8), 399-406. doi: 10.5301/ijao.5000 148. Clasen, P. C., Fisher, A. J., & Beevers, C. G. (2015). Mood-reactive self-esteem and depression vulnerability: Person-specific symptom dynamics via smart phone assessment. PLoS one, 10(7), e0129774. 149. Clift, K., Scott, L., Johnson, M., & Gonzalez, C. (2014). Leveraging geographic information systems in an integrated health care delivery organization. The Permanente Journal, 18(2), 71. 150. Coates, J., Souhami, L., & El Naqa, I. (2016). Big data analytics for prostate radiotherapy. Frontiers in oncology, 6, 149. 151. Cohen, I. G., Amarasingham, R., Shah, A., Xie, B., & Lo, B. (2014). The legal and ethical concerns that arise from using complex predictive analytics in health care. Health affairs, 33(7), 1139-1147. 152. Cole, T. S., Frankovich, J., Iyer, S., LePendu, P., Bauer-Mehren, A., & Shah, N. H. (2013). Profiling risk factors for chronic uveitis in juvenile idiopathic arthritis: a new model for EHR-based research. Pediatric Rheumatology, 11(1), 45. 153. Cole, C., Krampis, K., Karagiannis, K., Almeida, J. S., Faison, W. J., Motwani, M., ...& Mazumder, R. (2014). Non-synonymous variations in cancer and their effects on the human proteome: workflow for NGS data biocuration and proteome-wide analysis of TCGA data. BMC bioinformatics, 15(1), 28. 154. Cole, B. K., Simmers, M. B., Feaver, R., Qualls Jr, C. W., Collado, M. S., Berzin, E., ... & Manka, D. (2015). An in vitro cynomolgus vascular surrogate system for preclinical drug assessment and human translation. Arteriosclerosis, thrombosis, and vascular biology, 35(10), 2185-2195. 155. Collins, B. (2016). Big data and health economics: strengths, weaknesses, opportunities and threats. Pharmacoeconomics, 34(2), 101-106. 156. Conway, M., & O’Connor, D. (2016). Social media, big data, and mental health: current advances and ethical implications. Current opinion in psychology, 9, 77-82.

213

157. Cook, T. S., Zimmerman, S. L., Steingall, S. R., Maidment, A. D., Kim, W., & Boonn, W. W. (2011). Informatics in radiology: RADIANCE: an automated, enterprise-wide solution for archiving and reporting CT radiation dose estimates. Radiographics, 31(7), 1833-1846. 158. Cook, T. S., & Nagy, P. (2014). Business intelligence for the radiologist: making your data work for you. Journal of the American College of Radiology, 11(12), 1238-1240. 159. Cook, T. S., Zimmerman, S. L., Steingall, S. R., Boonn, W. W., & Kim, W. (2012). An algorithm for intelligent sorting of CT-related dose parameters. Journal of digital imaging, 25(1), 179-188. 160. Cooper, G. F., Bahar, I., Becich, M. J., Benos, P. V., Berg, J., Espino, J. U., ... & Lu, X. (2015). The center for causal discovery of biomedical knowledge from big data. Journal of the American Medical Informatics Association, 22(6), 1132-1136. 161. Costa, F. F. (2014). Big data in biomedicine. Drug discovery today, 19(4), 433-440. 162. Costello, J. M., Mazwi, M. L., McBride, M. E., Gambetta, K. E., Eltayeb, O., & Epting, C. L. (2015). Critical care for paediatric patients with heart failure. Cardiology in the Young, 25(S2), 74-86. 163. Cui, L., Tao, S., & Zhang, G. Q. (2016). Biomedical ontology quality assurance using a big data approach. ACM Transactions on Knowledge Discovery from Data (TKDD), 10(4), 41. 164. Curcin, V., Woodcock, T., Poots, A. J., Majeed, A., & Bell, D. (2014). Model-driven approach to data collection and reporting for quality improvement. Journal of biomedical informatics, 52, 151-162. 165. Currie, J. (2013). “Big data” versus “big brother”: on the appropriate use of large-scale data collections in pediatrics. Pediatrics, 131(Supplement 2), S127-S132. 166. Cushing, C. C., Walters, R. W., & Hoffman, L. (2013). Aggregated N-of-1 randomized controlled trials: Modern data analytics applied to a clinically valid method of intervention effectiveness. Journal of Pediatric Psychology, 39(2), 138-150. 167. Cyganek, B., Graña, M., Krawczyk, B., Kasprzak, A., Porwik, P., Walkowiak, K., & Woźniak, M. (2016). A survey of big data issues in electronic health record analysis. Applied Artificial Intelligence, 30(6), 497-520. 168. Dang, A., & Angle, V. S. (2015). Utilizing patient registries as health technology assessment (HTA) tool. Systematic Reviews in Pharmacy, 6(1), 5. 169. Davidson, M. W., Haim, D. A., & Radin, J. M. (2015). Using networks to combine “big data” and traditional surveillance to improve influenza predictions. Scientific reports, 5, 8154. 170. Dawadi, P. N., Cook, D. J., & Schmitter-Edgecombe, M. (2016). Modeling patterns of activities using activity curves. Pervasive and mobile computing, 28, 51-68. 171. de Bono, B., Grenon, P., & Sammut, S. J. (2012). ApiNATOMY: A novel toolkit for visualizing multiscale anatomy schematics with phenotype‐related information. Human mutation, 33(5), 837-848. 172. De Silva, D., Burstein, F., Jelinek, H. F., & Stranieri, A. (2015). Addressing the complexities of big data analytics in healthcare: the diabetes screening case. Australasian Journal of Information Systems, 19. 173. Dean, D. A., Goldberger, A. L., Mueller, R., Kim, M., Rueschman, M., Mobley, D., ...& Surovec, S. (2016). Scaling up scientific discovery in sleep medicine: the National Sleep Research Resource. Sleep, 39(5), 1151-1164. 174. Dean, D. A., Goldberger, A. L., Mueller, R., Kim, M., Rueschman, M., Mobley, D., ...& Surovec, S. (2016). Scaling up scientific discovery in sleep medicine: the National Sleep Research Resource. Sleep, 39(5), 1151-1164. 175. Delen, D. (2009). Analysis of cancer data: a data mining approach. Expert Systems, 26(1), 100-112. 176. Deray, K., & Simoff, S. (2011). RETRACTED: Designing for healthy living: Supporting reflectivity on interactions in healthcare. 177. Deserno, T. M., & Marx, N. (2016). Computational electrocardiography: revisiting Holter ECG monitoring. Methods of information in medicine, 55(04), 305-311.

214

178. Devinsky, O., Dilley, C., Ozery-Flato, M., Aharonov, R., Goldschmidt, Y. A., Rosen- Zvi, M., ...& Fritz, P. (2016). Changing the approach to treatment choice in epilepsy using big data. Epilepsy & Behavior, 56, 32-37. 179. Dilsizian, S. E., & Siegel, E. L. (2014). Artificial intelligence in medicine and cardiac imaging: harnessing big data and advanced computing to provide personalized medical diagnosis and treatment. Current cardiology reports, 16(1), 441. 180. Dimitriadis, S. I., Laskaris, N. A., Bitzidou, M. P., Tarnanas, I., & Tsolaki, M. N. (2015). A novel biomarker of amnestic MCI based on dynamic cross-frequency coupling patterns during cognitive brain responses. Frontiers in neuroscience, 9, 350. 181. Dimitrov, D. V. (2016). Medical internet of things and big data in healthcare. Healthcare informatics research, 22(3), 156-163. 182. Dinov, I. D. (2016). Methodological challenges and analytic opportunities for modeling and interpreting Big Healthcare Data. Gigascience, 5(1), 12. 183. Dinov, I. D., Heavner, B., Tang, M., Glusman, G., Chard, K., Darcy, M., ...& Foster, I. (2016). Predictive big data analytics: a study of Parkinson’s disease using large, complex, heterogeneous, incongruent, multi-source and incomplete observations. PloS one, 11(8), e0157077. 184. Divita, G., Carter, M., Redd, A., Zeng, Q., Gupta, K., Trautner, B., ...& Gundlapalli, A. (2015). Scaling-up NLP pipelines to process large corpora of clinical notes. Methods of information in medicine, 54(06), 548-552. 185. Dolin, R. H., Rogers, B., & Jaffe, C. (2015). Health level seven interoperability strategy: big data, incrementally structured. Methods of information in medicine, 54(01), 75- 82. 186. Doudican, N. A., Kumar, A., Singh, N. K., Nair, P. R., Lala, D. A., Basu, K., ... & Abbasi, T. (2015). Personalization of cancer treatment using predictive simulation. Journal of translational medicine, 13(1), 43. 187. Du, Q., Zhao, W., Li, W., Zhang, X., Sun, B., Song, H., ...& Wang, Y. (2016). Massive access control aided by knowledge-extraction for co-existing periodic and random services over wireless clinical networks. Journal of medical systems, 40(7), 171. 188. Dueñas-Espín, I., Vela, E., Pauws, S., Bescos, C., Cano, I., Cleries, M., & Kaye, R. (2016). Proposals for enhanced health risk assessment and stratification in an integrated care scenario. BMJ open, 6(4), e010301. 189. Dugan, T. M., Mukhopadhyay, S., Carroll, A., & Downs, S. (2015). Machine learning techniques for prediction of early childhood obesity. Applied clinical informatics, 6(03), 506-520. 190. Duggal, P. S., Paul, S., & Tiwari, P. (2015). Analytics for the quality of fertility data using Particle Swarm Optimization. International Journal of Bio-Science and Bio- Technology, 7(1), 39-50. 191. Edelstein, P. (2013). Emerging directions in analytics. Predictive analytics will play an indispensable role in healthcare transformation and reform. Health management technology, 34(1), 16. 192. Ekström, A., Kurland, L., Farrokhnia, N., Castrén, M., & Nordberg, M. (2015). Forecasting emergency department visits using internet data. Annals of emergency medicine, 65(4), 436-442. 193. Elbers, P. W., Girbes, A., Malbrain, M. L., & Bosman, R. (2015). Right dose, right now: using big data to optimize antibiotic dosing in the critically ill. Anaesthesiology intensive therapy, 47(5), 457-463. 194. El Emam, K., Arbuckle, L., Koru, G., Eze, B., Gaudette, L., Neri, E., ...& Gluck, J. (2012). De-identification methods for open health data: the case of the Heritage Health Prize claims dataset. Journal of medical Internet research, 14(1). 195. Ellis, D. W., & Srigley, J. (2016). Does standardised structured reporting contribute to quality in diagnostic pathology? The importance of evidence-based datasets. Virchows Archiv, 468(1), 51-59. 196. Elsebakhi, E., Lee, F., Schendel, E., Haque, A., Kathireason, N., Pathare, T., ...& Al- Ali, R. (2015). Large-scale machine learning based on functional networks for biomedical

215

big data with high performance computing platforms. Journal of Computational Science, 11, 69-81. 197. Elshazly, M. B., Quispe, R., Michos, E. D., Sniderman, A. D., Toth, P. P., Banach, M., ... & Martin, S. S. (2015). Patient-Level Discordance in Population Percentiles of the Total Cholesterol to High-Density Lipoprotein Cholesterol Ratio in Comparison With Low- Density Lipoprotein Cholesterol and Non–High-Density Lipoprotein CholesterolCLINICAL PERSPECTIVE: The Very Large Database of Lipids Study (VLDL- 2B). Circulation, 132(8), 667-676. 198. Erdman, A. G., Keefe, D. F., & Schiestl, R. (2013). Grand challenge: applying regulatory science and big data to improve medical device innovation. IEEE Trans. Biomed. Engineering, 60(3), 700-706. 199. Erickson, B. J., Meenan, C., & Langer, S. (2013). Standards for business analytics and departmental workflow. Journal of digital imaging, 26(1), 53-57. 200. Ermann, J., Rao, D. A., Teslovich, N. C., Brenner, M. B., & Raychaudhuri, S. (2015). Immune cell profiling to guide therapeutic decisions in rheumatic diseases. Nature Reviews Rheumatology, 11(9), 541. 201. Evangelatos, N., Reumann, M., Lehrach, H., & Brand, A. (2016). Clinical Trial Data as Public Goods: Fair Trade and the Virtual Knowledge Bank as a Solution to the Free Rider Problem-A Framework for the Promotion of Innovation by Facilitation of Clinical Trial Data Sharing among Biopharmaceutical Companies in the Era of Omics and Big Data. Public health genomics, 19(4), 211-219. 202. Ewing, E. T., Gad, S., & Ramakrishnan, N. (2013). Gaining insights into epidemics by mining historical newspapers. Computer, 46(6), 68-72. 203. Fabian, B., Ermakova, T., & Junghanns, P. (2015). Collaborative and secure sharing of healthcare data in multi-clouds. Information Systems, 48, 132-150. 204. Fahim, M., Idris, M., Ali, R., Nugent, C., Kang, B., Huh, E. N., & Lee, S. (2014). ATHENA: a personalized platform to promote an active lifestyle and wellbeing based on physical, mental and social health primitives. Sensors, 14(5), 9313-9329. 205. Falcon, M. I., Jirsa, V., & Solodkin, A. (2016). A new neuroinformatics approach to personalized medicine in neurology: The Virtual Brain. Current opinion in neurology, 29(4), 429. 206. Fang, R., Pouyanfar, S., Yang, Y., Chen, S. C., & Iyengar, S. S. (2016). Computational health informatics in the big data age: A survey. ACM Computing Surveys (CSUR), 49(1), 12. 207. Fang, Z., Fan, X., & Chen, G. (2014). A study on specialist or special disease clinics based on big data. Frontiers of medicine, 8(3), 376-381. 208. Farruggia, A., Magro, R., & Vitabile, S. (2014). A text based indexing system for mammographic image retrieval and classification. Future Generation Computer Systems, 37, 243-251. 209. Faurholt-Jepsen, M., Busk, J., Frost, M., Vinberg, M., Christensen, E. M., Winther, O., ...& Kessing, L. V. (2016). Voice analysis as an objective state marker in bipolar disorder. Translational psychiatry, 6(7), e856. 210. Feldman, K., & Chawla, N. V. (2015). Does medical school training relate to practice? Evidence from big data. Big data, 3(2), 103-113. 211. Feldman, K., Davis, D., & Chawla, N. V. (2015). Scaling and contextualizing personalized healthcare: A case study of disease prediction algorithm integration. Journal of biomedical informatics, 57, 377-385. 212. Fernandes, L. M., O'Connor, M., & Weaver, V. (2012). Big data, bigger outcomes. Journal of AHIMA, 83(10), 38-43. 213. Fernández-Luque, L., & Bau, T. (2015). Health and social media: perfect storm of information. Healthcare informatics research, 21(2), 67-73. 214. Ferrand, D., Amyot, D., & Corrales, C. V. (2010). Towards a business intelligence framework for healthcare safety. Journal of Internet Banking and Commerce, 15(3), 1-9.

216

215. Ferranti, J. M., Langman, M. K., Tanaka, D., McCall, J., & Ahmad, A. (2010). Bridging the gap: leveraging business intelligence tools in support of patient safety and financial effectiveness. Journal of the American Medical Informatics Association, 17(2), 136-143. 216. Feuerriegel, S. (2016). Decision support in healthcare: determining provider influence on treatment outcomes with robust risk adjustment. Journal of Decision systems, 25(4), 371-390. 217. Fihn, S. D., Francis, J., Clancy, C., Nielson, C., Nelson, K., Rumsfeld, J., & Graham, G. L. (2014). Insights from advanced analytics at the Veterans Health Administration. Health Affairs, 33(7), 1203-1211. 218. Filkins, B. L., Kim, J. Y., Roberts, B., Armstrong, W., Miller, M. A., Hultner, M. L., & Steinhubl, S. R. (2016). Privacy and security in the era of digital health: what should translational researchers know and do about it?. American journal of translational research, 8(3), 1560. 219. Finney, J. M., Walker, A. S., Peto, T. E., & Wyllie, D. H. (2011). An efficient record linkage scheme using graphical analysis for identifier error detection. BMC medical informatics and decision making, 11(1), 7. 220. Fischer, T., Brothers, K. B., Erdmann, P., & Langanke, M. (2016). Clinical decision- making and secondary findings in systems medicine. BMC medical ethics, 17(1), 32. 221. Fitzhenry, F., Resnic, F. S., Robbins, S. L., Denton, J., Nookala, L., Meeker, D., ...& Matheny, M. E. (2015). Creating a common data model for comparative effectiveness with the observational medical outcomes partnership. Applied clinical informatics, 6(03), 536- 547. 222. Flaig, T. W., Potluri, R. C., Ng, Y., Todd, M. B., & Mehra, M. (2016). Treatment evolution for metastatic castration‐resistant prostate cancer with recent introduction of novel agents: retrospective analysis of real‐world data. Cancer medicine, 5(2), 182-191. 223. Flechet, M., Grandas, F. G., & Meyfroidt, G. (2016). Informatics in neurocritical care: new ideas for Big Data. Current opinion in critical care, 22(2), 87-93. 224. Fleurence, R. L., Beal, A. C., Sheridan, S. E., Johnson, L. B., & Selby, J. V. (2014). Patient-powered research networks aim to improve patient care and health research. Health Affairs, 33(7), 1212-1219. 225. Flores, M., Glusman, G., Brogaard, K., Price, N. D., & Hood, L. (2013). P4 medicine: how systems medicine will transform the healthcare sector and society. Personalized medicine, 10(6), 565-576. 226. Forkan, A. R. M., Khalil, I., Ibaida, A., & Tari, Z. (2017). BDCaM: big data for context-aware monitoring—a personalized knowledge discovery framework for assisted healthcare. IEEE transactions on cloud computing, 5(4), 628-641. 227. Forrest, C. B., Margolis, P. A., Bailey, L. C., Marsolo, K., Del Beccaro, M. A., Finkelstein, J. A., ... & Kahn, M. G. (2014). PEDSnet: a national pediatric learning health system. Journal of the American Medical Informatics Association, 21(4), 602-606. 228. Forrest, C. B., Margolis, P. A., Bailey, L. C., Marsolo, K., Del Beccaro, M. A., Finkelstein, J. A., ... & Kahn, M. G. (2014). PEDSnet: a national pediatric learning health system. Journal of the American Medical Informatics Association, 21(4), 602-606. 229. Fortino, G., Parisi, D., Pirrone, V., & Di Fatta, G. (2014). BodyCloud: A SaaS approach for community body sensor networks. Future Generation Computer Systems, 35, 62-79. 230. Frizzo-Barker, J., Chow-White, P. A., Charters, A., & Ha, D. (2016). Genomic big data and privacy: Challenges and opportunities for precision medicine. Computer Supported Cooperative Work (CSCW), 25(2-3), 115-136. 231. Gale, T. C., Chatterjee, A., Mellor, N. E., & Allan, R. J. (2016). Health worker focused distributed simulation for improving capability of health systems in Liberia. Simulation in Healthcare, 11(2), 75-81. 232. Gálvez, J. A., Ahumada, L., Simpao, A. F., Lin, E. E., Bonafide, C. P., Choudhry, D., ... & Rehman, M. A. (2013). Visual analytical tool for evaluation of 10-year perioperative transfusion practice at a children's hospital. Journal of the American Medical Informatics Association, 21(3), 529-534.

217

233. Gange, S. J., & Golub, E. T. (2015). From smallpox to big data: the next 100 years of epidemiologic methods. American journal of epidemiology, 183(5), 423-426. 234. Gao, L., Pan, H., Xie, X., Zhang, Z., Li, Q., & Han, Q. (2016). Graph modeling and mining methods for brain images. Multimedia Tools and Applications, 75(15), 9333-9369. 235. Geerts, H., Dacks, P. A., Devanarayan, V., Haas, M., Khachaturian, Z. S., Gordon, M. F., & Brain Health Modeling Initiative. (2016). Big data to smart data in Alzheimer's disease: The brain health modeling initiative to foster actionable knowledge. Alzheimer's & Dementia, 12(9), 1014-1021. 236. Gimeno-Blanes, F. J., Blanco-Velasco, M., Barquero-Pérez, Ó., García-Alberola, A., & Rojo-Álvarez, J. L. (2016). Sudden cardiac risk stratification with electrocardiographic indices-a review on computational processing, technology transfer, and scientific evidence. Frontiers in physiology, 7, 82. 237. Gittelman, S., Lange, V., Crawford, C. A. G., Okoro, C. A., Lieb, E., Dhingra, S. S., & Trimarchi, E. (2015). A new source of data for public health surveillance: Facebook likes. Journal of medical Internet research, 17(4). 238. Glaßer, S., Preim, U., Tönnies, K., & Preim, B. (2010). A visual analytics approach to diagnosis of breast DCE-MRI data. Computers & Graphics, 34(5), 602-611. 239. Glenn, T., & Monteith, S. (2014). Privacy in the digital world: medical and health data outside of HIPAA protections. Current psychiatry reports, 16(11), 494. 240. Gligorijević, V., Malod‐Dognin, N., & Pržulj, N. (2016). Integrative methods for analyzing big data in precision medicine. Proteomics, 16(5), 741-758. 241. Glueck, M., Hamilton, P., Chevalier, F., Breslav, S., Khan, A., Wigdor, D., & Brudno, M. (2016). PhenoBlocks: phenotype comparison visualizations. IEEE transactions on visualization and computer graphics, 22(1), 101-110. 242. Glurich, I., Acharya, A., Brilliant, M. H., & Shukla, S. K. (2015). Progress in oral personalized medicine: contribution of ‘omics’. Journal of oral microbiology, 7(1), 28223. 243. Godara, S., & Singh, R. (2016). Evaluation of predictive machine learning techniques as expert systems in medical diagnosis. Indian Journal of Science and Technology, 9(10). 244. Goh, W. P., Tao, X., Zhang, J., & Yong, J. (2016). Decision support systems for adoption in dental clinics: a survey. Knowledge-Based Systems, 104, 195-206. 245. Goldenholz, D. M., Moss, R., Scott, J., Auh, S., & Theodore, W. H. (2015). Confusing placebo effect with natural history in epilepsy: a big data approach. Annals of neurology, 78(3), 329-336. 246. Goldman, D., & Domschke, K. (2014). Making sense of deep sequencing. International Journal of Neuropsychopharmacology, 17(10), 1717-1725. 247. Goli-Malekabadi, Z., Sargolzaei-Javan, M., & Akbari, M. K. (2016). An effective model for store and retrieve big health data in cloud computing. Computer methods and programs in biomedicine, 132, 75-82. 248. Gong, Y., Fang, Y., & Guo, Y. (2016). Private data analytics on biomedical sensing data via distributed computation. IEEE/ACM transactions on computational biology and bioinformatics, 13(3), 431-444. 249. Goossens, K., Van Uytfanghe, K., Twomey, P. J., & Thienpont, L. M. (2015). Monitoring laboratory data across manufacturers and laboratories—A prerequisite to make “Big Data” work. Clinica Chimica Acta, 445, 12-18. 250. Goossens, N., Nakagawa, S., Sun, X., & Hoshida, Y. (2015). Cancer biomarker discovery and validation. Translational cancer research, 4(3), 256. 251. Gorenshteyn, D., Zaslavsky, E., Fribourg, M., Park, C. Y., Wong, A. K., Tadych, A., ...& Troyanskaya, O. G. (2015). Interactive big data resource to elucidate human immune pathways and diseases. Immunity, 43(3), 605-614. 252. Gotz, D., Stavropoulos, H., Sun, J., & Wang, F. (2012). ICDA: a platform for intelligent care delivery analytics. In AMIA annual symposium proceedings (Vol. 2012, p. 264). American Medical Informatics Association. 253. Gotz, D., Wang, F., & Perer, A. (2014). A methodology for interactive mining and visual analysis of clinical event patterns using electronic health record data. Journal of biomedical informatics, 48, 148-159.

218

254. Graf, T., Erskine, A., & Steele Jr, G. D. (2014). Leveraging data to systematically improve care: coronary artery disease management at Geisinger. The Journal of ambulatory care management, 37(3), 199-205. 255. Grammer, A. C., Ryals, M. M., Heuer, S. E., Robl, R. D., Madamanchi, S., Davis, L. S., ... & Lipsky, P. E. (2016). Drug repositioning in SLE: crowd-sourcing, literature-mining and Big Data analysis. Lupus, 25(10), 1150-1170. 256. Graven, M., Allen, P., Smith, I., & MacDonald, N. E. (2013). Decline in mortality with the Belize integrated patient-centred country wide health information system (BHIS) with embedded program management. International journal of medical informatics, 82(10), 954- 963. 257. Green, S., & Vogt, H. (2016). Personalizing medicine: Disease prevention in silico and in socio. Humana. Mente Journal of Philosophical Studies, 9(30), 105-145. 258. Greenspun, H., & Bercik, W. (2013). Cost-outcomes focus is essential for ACO success: an ACO's ability to effectively manage the risks associated with value-based payment depends on how precisely it can account for costs and whether it can apply predictive analytics to analyze clinical and financial outcomes. Healthcare Financial Management, 67(2), 96-103. 259. Gribov, A., Sill, M., Lück, S., Rücker, F., Döhner, K., Bullinger, L., ...& Unwin, A. (2010). SEURAT: visual analytics for the integrated analysis of microarray data. BMC medical genomics, 3(1), 21. 260. Grossglauser, M., & Saner, H. (2014). Data-driven healthcare: from patterns to actions. European journal of preventive cardiology, 21(2_suppl), 14-17. 261. Gruber, S. (2015). Targeted learning in healthcare research. Big data, 3(4), 211-218. 262. Gunapal, P. P. G., Kannapiran, P., Teow, K. L., Zhu, Z., Xiaobin You, A., Saxena, N., ...& Sim, J. H. J. (2016). Setting up a regional health system database for seamless population health management in Singapore. Proceedings of Singapore Healthcare, 25(1), 27-34. 263. Guo, J., Liu, H., & Zheng, J. (2015). SynLethDB: synthetic lethality database toward discovery of selective and sensitive anticancer drug targets. Nucleic acids research, 44(D1), D1011-D1017. 264. Haarbrandt, B., Tute, E., & Marschollek, M. (2016). Automated population of an i2b2 clinical data warehouse from an openEHR-based data repository. Journal of biomedical informatics, 63, 277-294. 265. Halamka, J. D. (2014). Early experiences with big data at an academic medical center. Health Affairs, 33(7), 1132-1138. 266. Hamad, R., Modrek, S., Kubo, J., Goldstein, B. A., & Cullen, M. R. (2015). Using “big data” to capture overall health status: Properties and predictive value of a claims-based health risk score. PloS one, 10(5), e0126054. 267. Han, H., & Liu, Y. (2016). Transcriptome marker diagnostics using big data. IET systems biology, 10(1), 41-48. 268. Hao, H., & Zhang, K. (2016). The voice of Chinese health consumers: a text mining approach to web-based physician reviews. Journal of medical Internet research, 18(5). 269. Hao, S., Jin, B. O., Shin, A. Y., Zhao, Y., Zhu, C., Li, Z., ...& Zhao, Y. (2014). Risk prediction of emergency department revisit 30 days post discharge: a prospective study. PloS one, 9(11), e112944. 270. Haque, W., Urquhart, B., Berg, E., & Dhanoa, R. (2014). Using business intelligence to analyze and share health system infrastructure data in a rural health authority. JMIR medical informatics, 2(2). 271. Harris, D. R., Henderson, D. W., Kavuluru, R., Stromberg, A. J., & Johnson, T. R. (2014). Using common table expressions to build a scalable boolean query generator for clinical data warehouses. IEEE journal of biomedical and health informatics, 18(5), 1607- 1613. 272. Harris, S. L., May, J. H., & Vargas, L. G. (2016). Predictive analytics model for healthcare planning and scheduling. European Journal of Operational Research, 253(1), 121-131.

219

273. Hart, T., & Xie, L. (2016). Providing data science support for systems pharmacology and its implications to drug discovery. Expert opinion on drug discovery, 11(3), 241-256. 274. Hasan, S., Shamsuddin, S. M., & Lopes, N. (2014). Machine learning big data framework and analytics for big data problems. Int. J. Advance Soft Compu. Appl, 6(2). 275. Hata, Y., & Nakajima, H. (2014). A Survey of Intelligent Computing in Medical and Health Care System. IEICE TRANSACTIONS on Information and Systems, 97(9), 2218- 2225. 276. He, P., Wang, P., Gao, J., & Tang, B. (2015). City-wide smart healthcare appointment systems based on cloud data virtualization paas. International Journal of Multimedia and Ubiquitous Engineering, 10(2), 371-382. 277. He, Y. (2014). Ontology-supported research on vaccine efficacy, safety and integrative biological networks. Expert review of vaccines, 13(7), 825-841. 278. Helm, E., & Paster, F. (2015). First steps towards process mining in distributed health information systems. International Journal of Electronics and Telecommunications, 61(2), 137-142. 279. Helm-Murtagh, S. C. (2014). Use of big data by blue cross and blue shield of North Carolina. North Carolina medical journal, 75(3), 195-197. 280. Higdon, R., Stewart, E., Roach, J. C., Dombrowski, C., Stanberry, L., Clifton, H., ...& Kolker, E. (2013). Predictive analytics in healthcare: Medications as a predictor of medical complexity. Big data, 1(4), 237-244. 281. Hilario, M., Kalousis, A., Pellegrini, C., & Mueller, M. (2006). Processing and classification of protein mass spectra. Mass spectrometry reviews, 25(3), 409-449. 282. Hiller, J. S. (2016). Healthy Predictions? Questions for Data Analytics in Health Care. American Business Law Journal, 53(2), 251-314. 283. Holzinger, A., & Jurisica, I. (2014). Knowledge discovery and data mining in biomedical informatics: The future is in integrative, interactive machine learning solutions. In Interactive knowledge discovery and data mining in biomedical informatics (pp. 1-18). Springer, Berlin, Heidelberg. 284. Holzinger, A., Schantl, J., Schroettner, M., Seifert, C., & Verspoor, K. (2014). Biomedical text mining: state-of-the-art, open problems and future challenges. In Interactive knowledge discovery and data mining in biomedical informatics (pp. 271- 300). Springer, Berlin, Heidelberg. 285. Horowitz, G. L., Zaman, Z., Blanckaert, N. J., Chan, D. W., Dubois, J. A., Golaz, O., ... & Marocchi, A. (1900). MODULAR ANALYTICS: a new approach to automation in the clinical laboratory. Journal of Analytical Methods in Chemistry, 2005(1), 8-25. 286. Horvath, M. M., Cozart, H., Ahmad, A., Langman, M. K., & Ferranti, J. (2009). Sharing adverse drug event data using business intelligence technology. Journal of patient safety, 5(1), 35-41. 287. Horvath, M. M., Winfield, S., Evans, S., Slopek, S., Shang, H., & Ferranti, J. (2011). The DEDUCE Guided Query tool: providing simplified access to clinical data for research and quality improvement. Journal of biomedical informatics, 44(2), 266-276. 288. Hossain, M. S., & Muhammad, G. (2016). Cloud-assisted industrial internet of things (iiot)–enabled framework for health monitoring. Computer Networks, 101, 192-202. 289. Howard, C. M., & Felton, K. W. (2013). Determining hospital risk management staffing through analytics. Journal of Healthcare Risk Management, 33(2), 36-42. 290. Howren, M. B., Vander Weg, M. W., & Wolinsky, F. D. (2014). Computerized cognitive training interventions to improve neuropsychological outcomes: evidence and future directions. Journal of comparative effectiveness research, 3(2), 145-154. 291. Hrovat, G., Stiglic, G., Kokol, P., & Ojsteršek, M. (2014). Contrasting temporal trend discovery for large healthcare databases. Computer methods and programs in biomedicine, 113(1), 251-257. 292. Hsieh, J. C., Li, A. H., & Yang, C. C. (2013). Mobile, cloud, and big data computing: contributions, challenges, and new directions in telecardiology. International journal of environmental research and public health, 10(11), 6131-6153.

220

293. Hsueh, P. Y. S., Zhu, X. X., Hsiao, M. J., Lee, S. Y., Deng, V., & Ramakrishnan, S. (2015). Automatic summarization of risk factors preceding disease progression an insight- driven healthcare service case study on using medical records of diabetic patients. World Wide Web, 18(4), 1163-1175. 294. Hu, J., Wang, F., Sun, J., Sorrentino, R., & Ebadollahi, S. (2012). A healthcare utilization analysis framework for hot spotting and contextual anomaly detection. In AMIA annual symposium proceedings (Vol. 2012, p. 360). American Medical Informatics Association. 295. Hu, L., Zhang, Y., Feng, D., Hassan, M. M., Alelaiwi, A., & Alamri, A. (2015). Design of QoS-aware multi-level MAC-layer for wireless body area network. Journal of medical systems, 39(12), 192. 296. Huang, C. W., Lu, R., Iqbal, U., Lin, S. H., Nguyen, P. A. A., Yang, H. C., ... & Jian, W. S. (2015). A richly interactive exploratory data analysis and visualization tool using electronic medical records. BMC medical informatics and decision making, 15(1), 92. 297. Huang, D. C., Wang, J. F., Huang, J. X., Sui, D. Z., Zhang, H. Y., Hu, M. G., & Xu, C. D. (2016). Towards identifying and reducing the bias of disease information extracted from search engine data. PLoS computational biology, 12(6), e1004876. 298. Huang, M., Nichols, T., Huang, C., Yu, Y., Lu, Z., Knickmeyer, R. C., ...& Alzheimer's Disease Neuroimaging Initiative. (2015). FVGWAS: Fast voxelwise genome wide association analysis of large-scale imaging genetic data. Neuroimage, 118, 613-627. 299. Huang, T., Lan, L., Fang, X., An, P., Min, J., & Wang, F. (2015). Promises and challenges of big data computing in health sciences. Big Data Research, 2(1), 2-11. 300. Huang, W., Zeng, S., Wan, M., & Chen, G. (2016). Medical media analytics via ranking and big learning: A multi-modality image-based disease severity prediction study. Neurocomputing, 204, 125-134. 301. Hughes, A., Landers, D., Arkenau, H. T., Shah, S., Stephens, R., Mahal, A., ...& Royle, J. (2016). Development and evaluation of a new technological way of engaging patients and enhancing understanding of drug tolerability in early clinical development: PROACT. Advances in therapy, 33(6), 1012-1024. 302. Hui, J., Knoop, S., & Schwarz, P. (2011). HIWAS: enabling technology for analysis of clinical data in XML documents. To be appeared in, Proc 37th VLDB, Seattle. 303. Huque, M. H., Anderson, C., Walton, R., & Ryan, L. (2016). Individual level covariate adjusted conditional autoregressive (indiCAR) model for disease mapping. International journal of health geographics, 15(1), 25. 304. Hussain, S., Bang, J. H., Han, M., Ahmed, M. I., Amin, M. B., Lee, S., ...& Parr, G. (2014). Behavior life style analysis for mobile sensory data in cloud computing through MapReduce. Sensors, 14(11), 22001-22020. 305. Iniesta, R., Stahl, D., & McGuffin, P. (2016). Machine learning, statistical learning and the future of biological research in psychiatry. Psychological medicine, 46(12), 2455-2465. 306. Iqbal, U., Hsu, C. K., Nguyen, P. A. A., Clinciu, D. L., Lu, R., Syed-Abdul, S., ... & Chang, Y. C. (2016). Cancer-disease associations: A visualization and animation through medical big data. Computer methods and programs in biomedicine, 127, 44-51. 307. Ireland, M. E., Schwartz, H. A., Chen, Q., Ungar, L. H., & Albarracín, D. (2015). Future-oriented tweets predict lower county-level HIV prevalence in the United States. Health Psychology, 34(S), 1252. 308. Issa, Naiem T., Stephen W. Byers, and Sivanesan Dakshanamurthy. "Big data: the next frontier for innovation in therapeutics and healthcare." Expert review of clinical pharmacology 7.3 (2014): 293-298. 309. Istephan, S., & Siadat, M. R. (2016). Unstructured medical image query using big data– an epilepsy case study. Journal of biomedical informatics, 59, 218-226. 310. Ivan, M. L., Trifu, M. R., Velicanu, M., & Ciurea, C. (2016). Using business intelligence tools for predictive analytics in healthcare system. International Journal of Advanced Computer Science and Applications, 7(5), 178-182.

221

311. Iwasaki, Y., Abe, T., Wada, Y., Wada, K., & Ikemura, T. (2013). Novel bioinformatics strategies for prediction of directional sequence changes in influenza virus genomes and for surveillance of potentially hazardous strains. BMC infectious diseases, 13(1), 386. 312. Jadhav, A., Andrews, D., Fiksdal, A., Kumbamu, A., McCormick, J. B., Misitano, A., ...& Pathak, J. (2014). Comparative analysis of online health queries originating from personal computers and smart devices on a consumer health information portal. Journal of medical Internet research, 16(7). 313. Janevski, A., Kamalakaran, S., Banerjee, N., Varadan, V., & Dimitrova, N. (2009, September). PAPAyA: a platform for breast cancer biomarker signature discovery, evaluation and assessment. In BMC bioinformatics (Vol. 10, No. 9, p. S7). BioMed Central. 314. Janke, A. T., Overbeek, D. L., Kocher, K. E., & Levy, P. D. (2016). Exploring the potential of predictive analytics and big data in emergency care. Annals of emergency medicine, 67(2), 227-236. 315. Jauhari, S., & Rizvi, S. A. M. (2015). An Indian eye to personalized medicine. Computers in biology and medicine, 59, 211-220. 316. Jayapandian, C. P., Chen, C. H., Bozorgi, A., Lhatoo, S. D., Zhang, G. Q., & Sahoo, S. S. (2013). Cloudwave: distributed processing of “Big Data” from electrophysiological recordings for epilepsy clinical research using Hadoop. In AMIA Annual Symposium Proceedings (Vol. 2013, p. 691). American Medical Informatics Association. 317. Jee, K., & Kim, G. H. (2013). Potentiality of big data in the medical sector: focus on how to reshape the healthcare system. Healthcare informatics research, 19(2), 79-85. 318. Jeffery, A. D. (2015). Methodological challenges in examining the impact of healthcare predictive analytics on nursing-sensitive patient outcomes. CIN: Computers, Informatics, Nursing, 33(6), 258-264. 319. Jelinek, H. F., Stranieri, A., Yatsko, A., & Venkatraman, S. (2016). Data analytics identify glycated haemoglobin co-markers for type 2 diabetes mellitus diagnosis. Computers in biology and medicine, 75, 90-97. 320. Ji, Z., Ganchev, I., O’Droma, M., Zhang, X., & Zhang, X. (2014). A cloud-based X73 ubiquitous mobile healthcare system: design and implementation. The Scientific World Journal, 2014. 321. Jia, T., Tao, H., Qin, K., Wang, Y., Liu, C., &Gao, Q. (2014). Selecting the optimal healthcare centers with a modified P-median model: a visual analytic perspective. International journal of health geographics, 13(1), 42. 322. Jiang, P., Winkley, J., Zhao, C., Munnoch, R., Min, G., & Yang, L. T. (2016). An intelligent information forwarder for healthcare big data systems with distributed wearable sensors. IEEE systems journal, 10(3), 1147-1159. 323. Jin, H., Wu, S., Vidyanti, I., Di Capua, P., & Wu, B. (2015). Predicting Depression among Patients with Diabetes Using Longitudinal Data. Methods of information in medicine, 54(06), 553-559. 324. Johnson, A. E., Ghassemi, M. M., Nemati, S., Niehaus, K. E., Clifton, D. A., & Clifford, G. D. (2016). Machine learning and decision support in critical care. Proceedings of the IEEE, 104(2), 444-466. 325. Kaggal, V. C., Elayavilli, R. K., Mehrabi, S., Pankratz, J. J., Sohn, S., Wang, Y., ...& Chaudhry, R. (2016). Toward a learning health-care system–knowledge delivery at the point of care empowered by big data and NLP. Biomedical informatics insights, 8, BII-S37977. 326. Kamdar, M. R., Zeginis, D., Hasnain, A., Decker, S., & Deus, H. F. (2014). ReVeaLD: A user-driven domain-specific interactive search platform for biomedical research. Journal of biomedical informatics, 47, 112-130. 327. Kanevsky, J., Corban, J., Gaster, R., Kanevsky, A., Lin, S., & Gilardino, M. (2016). Big data and machine learning in plastic surgery: a new frontier in surgical innovation. Plastic and reconstructive surgery, 137(5), 890e-897e. 328. Kansagra, A. P., John-Paul, J. Y., Chatterjee, A. R., Lenchik, L., Chow, D. S., Prater, A. B., ... & Smith, S. E. (2016). Big data and the future of radiology informatics. Academic radiology, 23(1), 30-42.

222

329. Kao, H. Y., Yu, M. C., Masud, M., Wu, W. H., Chen, L. J., & Wu, Y. C. J. (2016). Design and evaluation of hospital-based business intelligence system (HBIS): A foundation for design science research methodology. Computers in Human Behavior, 62, 495-505. 330. Karami, M., Safdari, R., & Rahimi, A. (2013). Effective radiology dashboards: key research findings. Radiol Manage, 35(2), 42-5. 331. Karami, M., Fatehi, M., Torabi, M., Langarizadeh, M., Rahimi, A., & Safdari, R. (2013). Enhance hospital performance from intellectual capital to business intelligence. Radiol Manage, 35(6), 30-35. 332. Karimi, N., Samavi, S., Soroushmehr, S. R., Shirani, S., & Najarian, K. (2016). Toward practical guideline for design of image compression algorithms for biomedical applications. Expert Systems with Applications, 56, 360-367. 333. Karthika, V., Anu, M. & Veeramuthu, A. 2016 An efficient attribute based crtyptographic algothithms for securing trustworthy storage and auting for healthcare big data in cloud ARPN Journal of Engineering and Applied Sciences. 334. Kass-Hout, T. A., Xu, Z., Mohebbi, M., Nelsen, H., Baker, A., Levine, J., ...& Bright, R. A. (2015). OpenFDA: an innovative platform providing access to a wealth of FDA’s publicly available data. Journal of the American Medical Informatics Association, 23(3), 596-600. 335. Katayev, A., Fleming, J. K., Luo, D., Fisher, A. H., & Sharp, T. M. (2015). Reference intervals data mining: no longer a probability paper method. American journal of clinical pathology, 143(1), 134-142. 336. Katircioglu, K., Gooby, R., Helander, M., Drissi, Y., Chowdhary, P., Johnson, M., & Yonezawa, T. (2014). Supply Chain Scenario Modeler: A Holistic Executive Decision Support Solution. Interfaces, 44(1), 85-104. 337. Kaur, K., & Rani, R. (2015). A smart polyglot solution for big data in healthcare. IT Professional, 17(6), 48-55. 338. Kaur, K., & Rani, R. (2015). Managing data in healthcare information systems: many models, one solution. Computer, 48(3), 52-59. 339. Kavitha, R., Kannan, E., & Kotteswaran, S. (2016). Implementation of cloud based Electronic Health Record (EHR) for Indian healthcare needs. Indian Journal of Science and Technology, 9(3). 340. Kawamoto, K., Martin, C. J., Williams, K., Tu, M. C., Park, C. G., Hunter, C., ...& Morris, S. J. (2014). Value Driven Outcomes (VDO): a pragmatic, modular, and extensible software framework for understanding and improving health care costs and outcomes. Journal of the American Medical Informatics Association, 22(1), 223-235. 341. Kayış, E., Khaniyev, T. T., Suermondt, J., & Sylvester, K. (2015). A robust estimation model for surgery durations with temporal, operational, and surgery team effects. Health care management science, 18(3), 222-233. 342. Kazantsev, D., Guo, E., Kaestner, A., Lionheart, W. R., Bent, J., Withers, P. J., & Lee, P. D. (2016). Temporal sparsity exploiting nonlocal regularization for 4D computed tomography reconstruction. Journal of X-ray science and technology, 24(2), 207-219. 343. Kenner, A. (2016). Asthma on the move: how mobile apps remediate risk for disease management. Health, Risk & Society, 17(7-8), 510-529. 344. Ketchersid, T. (2013). Big data in nephrology: friend or foe?. Blood purification, 36(3- 4), 160-164. 345. Khan, N., Yaqoob, I., Hashem, I. A. T., Inayat, Z., Ali, M., Kamaleldin, W., ...& Gani, A. (2014). Big data: survey, technologies, opportunities, and challenges. The Scientific World Journal, 2014. 346. Khansa, L., Forcade, J., Nambari, G., Parasuraman, S., & Cox, P. (2012). Proposing an intelligent cloud-based electronic health record system. International Journal of Business Data Communications and Networking (IJBDCN), 8(3), 57-71. 347. Khare, R., Good, B. M., Leaman, R., Su, A. I., & Lu, Z. (2015). Crowdsourcing in biomedicine: challenges and opportunities. Briefings in bioinformatics, 17(1), 23-32.

223

348. Khazaei, H., Mench-Bressan, N., McGregor, C., & Pugh, J. E. (2015). Health informatics for neonatal intensive care units: An analytical modeling perspective. IEEE journal of translational engineering in health and medicine, 3. 349. Khokhar, B., Jette, N., Metcalfe, A., Cunningham, C. T., Quan, H., Kaplan, G. G., ...& Rabi, D. (2016). Systematic review of validated case definitions for diabetes in ICD-9- coded and ICD-10-coded data in adult populations. BMJ open, 6(8), e009952. 350. Kim, H. J., & Jeon, B. (2016). How close are we to individualized medicine for Parkinson’s disease?. Expert review of neurotherapeutics, 16(7), 815-830. 351. Kim, J., & Lee, W. (2015). Stochastic decision making for adaptive crowdsourcing in medical big-data platforms. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 45(11), 1471-1476. 352. Kim, M. J., Jang, J. W., & Yu, Y. S. (2015). A Study on the Development of Real-Time Analysis Monitoring System and Its Application of Medical Ins. In Advances in Computer Science and Ubiquitous Computing (pp. 411-416). Springer, Singapore. 353. Kim, M. J., & Yu, Y. S. (2015). Development of Real-time Big Data Analysis System and a Case Study on the Application of Information in a Medical Institution. International Journal of Software Engineering and Its Applications, 9(7), 93-102. 354. Kim, M. T., Huang, R., Sedykh, A., Wang, W., Xia, M., & Zhu, H. (2015). Mechanism profiling of hepatotoxicity caused by oxidative stress using antioxidant response element reporter gene assay models and big data. Environmental health perspectives, 124(5), 634- 641. 355. Kim, S., Lee, H., & Chung, Y. D. (2017). Privacy-preserving data cube for electronic medical records: An experimental evaluation. International journal of medical informatics, 97, 33-42. 356. Kim, T. W., Park, K. H., Yi, S. H., & Kim, H. C. (2014, June). A big data framework for u-healthcare systems utilizing vital signs. In Computer, Consumer and Control (IS3C), 2014 International Symposium on (pp. 494-497). IEEE. 357. Kitakaze, M., Asakura, M., Nakano, A., Takashima, S., & Washio, T. (2015). Data mining as a powerful tool for creating novel drugs in cardiovascular medicine: the importance of a “back-and-forth loop” between clinical data and basic research. Cardiovascular drugs and therapy, 29(3), 309-315. 358. Kite, B. J., Tangasi, W., Kelley, M., Bower, J. K., & Foraker, R. E. (2015). Electronic medical records and their use in health promotion and population research of cardiovascular disease. Current Cardiovascular Risk Reports, 9(1), 422. 359. Klann, J. G., Abend, A., Raghavan, V. A., Mandl, K. D., & Murphy, S. N. (2016). Data interchange using i2b2. Journal of the American Medical Informatics Association, 23(5), 909-915. 360. Klann, J. G., Buck, M. D., Brown, J., Hadley, M., Elmore, R., Weber, G. M., & Murphy, S. N. (2014). Query Health: standards-based, cross-platform population health surveillance. Journal of the American Medical Informatics Association, 21(4), 650-656. 361. Klein, S. M. (2015). Generating Real‐Time, Actionable Outcome Measures at Blythedale Children's Hospital. Global Business and Organizational Excellence, 34(4), 6- 17. 362. Klimov, D., Shknevsky, A., & Shahar, Y. (2014). Exploration of patterns predicting renal damage in patients with diabetes type II using a visual temporal analysis laboratory. Journal of the American Medical Informatics Association, 22(2), 275-289. 363. Klonoff, D. C. (2015). Precision medicine for managing diabetes. 364. Kohli, M. D., Warnock, M., Daly, M., Toland, C., Meenan, C., & Nagy, P. G. (2014). Building blocks for a clinical imaging informatics environment. Journal of digital imaging, 27(2), 174-181. 365. Kokol, P., Blazun H. V. &Zeleznik, D.(2016) Visualising nursing data using correspondence analysis. Nurse Researcher 24(1):38-40 366. Kolker, E., & Kolker, E. (2014). Healthcare analytics: Creating a prioritized improvement system with performance benchmarking. Big Data, 2(1), 50-54.

224

367. Kolowitz, B. J., Lauro, G. R., Venturella, J., Georgiev, V., Barone, M., Deible, C., & Shrestha, R. (2014). Clinical Social Networking—A New Revolution in Provider Communication and Delivery of Clinical Information across Providers of Care?. Journal of digital imaging, 27(2), 192-199. 368. Komenda, M., Schwarz, D., Švancara, J., Vaitsis, C., Zary, N., & Dušek, L. (2015). Practical use of medical terminology in curriculum mapping. Computers in biology and medicine, 63, 74-82. 369. Komenda, M., Víta, M., Vaitsis, C., Schwarz, D., Pokorná, A., Zary, N., & Dušek, L. (2015). Curriculum mapping with academic analytics in medical and healthcare education. PloS one, 10(12), e0143748. 370. Kondziolka, D., Cooper, B. T., Lunsford, L. D., & Silverman, J. (2015). Development, implementation, and use of a local and global clinical registry for neurosurgery. Big data, 3(2), 80-89. 371. Körpeoğlu, E., Kurtz, Z., Kılınç-Karzan, F., Kekre, S., & Basu, P. A. (2014). Business analytics assists transitioning traditional medicine to telemedicine at virtual radiologic. Interfaces, 44(4), 393-410. 372. Koster, J., Stewart, E., & Kolker, E. (2016). Health care transformation: A strategy rooted in data and analytics. Academic Medicine, 91(2), 165-167. 373. Krause, J., Perer, A., & Bertini, E. (2014). INFUSE: interactive feature selection for predictive modeling of high dimensional data. IEEE transactions on visualization and computer graphics, 20(12), 1614-1623. 374. Krumholz, H. M. (2014). Big data and new knowledge in medicine: the thinking, training, and tools needed for a learning health system. Health Affairs, 33(7), 1163-1170. 375. Kuceyeski, A., Navi, B. B., Kamel, H., Relkin, N., Villanueva, M., Raj, A., ...& Iadecola, C. (2015). Exploring the brain's structural connectome: A quantitative stroke lesion‐dysfunction mapping study. Human brain mapping, 36(6), 2147-2160. 376. Kudyba, S., & Rader, M. (2010). Conceptual factors to leverage business intelligence in healthcare (Electronic medical records, six sigma and workflow management). Proceedings Of The Northeast Business & Economics Association, 428-430. 377. Kuiler, E. W. (2014). From B ig D ata to Knowledge: An Ontological Approach to Big D ata Analytics. Review of Policy Research, 31(4), 311-318. 378. Kulkarni, P., Smith, L. D., & Woeltje, K. F. (2016). Assessing risk of hospital readmissions for improving medical practice. Health care management science, 19(3), 291- 299. 379. Kumar, A., Nette, F., Klein, K., Fulham, M., & Kim, J. (2015). A visual analytics approach using the exploration of multidimensional feature spaces for content-based medical image retrieval. IEEE journal of biomedical and health informatics, 19(5), 1734- 1746. 380. Kumar, J. S., & Appavu, S. (2016). The personalized disease predication care from harm using big data analytics in healthcare. Indian Journal of Science and Technology, 9(8). 381. Kumar, P., Mohammed, S., Kim, A., & Fiaidhi, J. (2016). Revisiting Medical Entity Recognition through the Guidelines of the Aurora Initiative. International Journal of Bio- Science and Bio-Technology, 8(4), 111-124. 382. Kumar, R. B., Goren, N. D., Stark, D. E., Wall, D. P., & Longhurst, C. A. (2016). Automated integration of continuous glucose monitor data in the electronic health record using consumer technology. Journal of the American Medical Informatics Association, 23(3), 532-537. 383. Kumar, S., Abowd, G. D., Abraham, W. T., al’Absi, M., Gayle Beck, J., Chau, D. H., .& Ganesan, D. (2015). Center of excellence for mobile sensor data-to-knowledge (MD2K). Journal of the American Medical Informatics Association, 22(6), 1137-1142.. 384. Kurczy, M. E., Ivanisevic, J., Johnson, C. H., Uritboonthai, W., Hoang, L., Fang, M., ...& Tautenhahn, R. (2015). Determining conserved metabolic biomarkers from a million database queries. Bioinformatics, 31(23), 3721-3724.

225

385. Kuzu, M., Kantarcioglu, M., Durham, E. A., Toth, C., & Malin, B. (2012). A practical approach to achieve private medical record linkage in light of public resources. Journal of the American Medical Informatics Association, 20(2), 285-292. 386. La, H. J. (2016). A conceptual framework for trajectory-based medical analytics with IoT contexts. Journal of Computer and System Sciences, 82(4), 610-626. 387. Lal, S. V., Palaniappan, R., & Prakash, V. (2015). Real time nursing management system for health care industry by using xenomai kernel. Indian Journal of Science and Technology, 8(20). 388. Lary, D. J., Woolf, S., Faruque, F., & LePage, J. P. (2014). Holistics 3.0 for health. ISPRS International Journal of Geo-Information, 3(3), 1023-1038. 389. Lazarou, I., Karakostas, A., Stavropoulos, T. G., Tsompanidis, T., Meditskos, G., Kompatsiaris, I., & Tsolaki, M. (2016). A novel and intelligent home monitoring system for care support of elders with cognitive impairment. Journal of Alzheimer's Disease, 54(4), 1561-1591. 390. Lee, J., Maslove, D. M., & Dubin, J. A. (2015). Personalized mortality prediction driven by electronic medical data and a patient similarity metric. PloS one, 10(5), e0127428. 391. Lee, J., Maslove, D. M., & Dubin, J. A. (2015). Personalized mortality prediction driven by electronic medical data and a patient similarity metric. PloS one, 10(5), e0127428.Lee, K. A., Dziadkowiec, O., & Meek, P. (2014). A systems science approach to fatigue management in research and health care. Nursing outlook, 62(5), 313-321. 392. Lee, M., Park, Y. S., Kim, M. H., & Lee, J. W. (2016). A convergence data model for medical information related to acute myocardial infarction. Human-centric Computing and Information Sciences, 6(1), 15. 393. Leeper, N. J., Bauer-Mehren, A., Iyer, S. V., LePendu, P., Olson, C., & Shah, N. H. (2013). Practice-based evidence: profiling the safety of cilostazol by text-mining of clinical notes. PloS one, 8(5), e63499. 394. Levtzow, C. B., & Willis, M. S. (2013). Reducing laboratory billing defects using six sigma principles. Laboratory Medicine, 44(4), 358-371. 395. Li, D., & Budoff, M. J. (2016). Genetics paired with CT angiography in the setting of atherosclerosis. Clinical imaging, 40(5), 917-925. 396. Li, G., Zuo, X., & Liu, B. (2014). Scientific computation of big data in real-world clinical research. Frontiers of medicine, 8(3), 310-315. 397. Li, K., Guo, L., Faraco, C., Zhu, D., Chen, H., Yuan, Y., ...& Hu, X. (2012). Visual analytics of brain networks. NeuroImage, 61(1), 82-97. 398. Li, L. (2015). The potential of translational bioinformatics approaches for pharmacology research. British journal of clinical pharmacology, 80(4), 862-867. 399. Li, Y., & Guo, Y. (2016). Wiki-health: from quantified self to self- understanding. Future Generation Computer Systems, 56, 333-359. 400. Li, Z., Wei, Z., Yue, Y., Wang, H., Jia, W., Burke, L. E., ...& Sun, M. (2015). An adaptive hidden markov model for activity recognition based on a wearable multi-sensor device. Journal of medical systems, 39(5), 57. 401. Liaw, S. T., Taggart, J., Yu, H., & de Lusignan, S. (2013). Data extraction from electronic health records-existing tools may be unreliable and potentially unsafe. Australian Family Physician, 42(11), 820. 402. Liebeskind, D. S., Albers, G. W., Crawford, K., Derdeyn, C. P., George, M. S., Palesch, Y. Y., & Sacco, R. L. (2015). Imaging in StrokeNet: realizing the potential of big data. Stroke, 46(7), 2000-2006. 403. Lin, C., Song, Z., Song, H., Zhou, Y., Wang, Y., & Wu, G. (2016). Differential privacy preserving in big data analytics for connected health. Journal of medical systems, 40(4), 97. 404. Lin, C., Wang, P., Song, H., Zhou, Y., Liu, Q., & Wu, G. (2016). A differential privacy protection scheme for sensitive big data in body sensor networks. Annals of Telecommunications, 71(9-10), 465-475. 405. Lin, F., Li, Z., Hua, Y., & Lim, Y. P. (2016). Proteomic profiling predicts drug response to novel targeted anticancer therapeutics. Expert review of proteomics, 13(4), 411-420.

226

406. Lin, N., Jiang, J., Guo, S., & Xiong, M. (2015). Functional principal component analysis and randomized sparse clustering algorithm for medical image analysis. PloS one, 10(7), e0132945. 407. Lin, S., Yin, Y. A., Jiang, X., Sahni, N., & Yi, S. (2016). Multi-OMICs and genome editing perspectives on liver cancer signaling networks. BioMed research international, 2016. 408. Lin, W., Dou, W., Zhou, Z., & Liu, C. (2015). A cloud-based framework for Home- diagnosis service over big medical data. Journal of Systems and Software, 102, 192-206. 409. Ling, Z. J., Tran, Q. T., Fan, J., Koh, G. C., Nguyen, T., Tan, C. S., ...& Zhang, M. (2014). GEMINI: an integrative healthcare analytics system. Proceedings of the VLDB Endowment, 7(13), 1766-1771. 410. Lismont, J., Janssens, A. S., Odnoletkova, I., vanden Broucke, S., Caron, F., & Vanthienen, J. (2016). A guide for the application of analytics on healthcare processes: A dynamic view on patient pathways. Computers in biology and medicine, 77, 125-134. 411. Liu, F., Feng, Y., Li, Z., Pan, C., Su, Y., Yang, R., ...& Deng, N. (2014). Clinic- genomic association mining for colorectal cancer using publicly available datasets. BioMed research international, 2014. 412. Liu, J., Ma, J., Wang, J., Zeng, D. D., Song, H., Wang, L., & Cao, Z. (2016). Comorbidity analysis according to sex and age in hypertension patients in China. International journal of medical sciences, 13(2), 99. 413. Liu, L., Tian, Z., Zhang, Z., & Fei, B. (2016). Computer-aided detection of prostate cancer with MRI: technology and applications. Academic radiology, 23(8), 1024-1046. 414. Liu, P., & Wu, S. (2016). An agent-based simulation model to study accountable care organizations. Health care management science, 19(1), 89-101. 415. Liu, X., & Chen, H. (2015). Identifying adverse drug events from patient social media: a case study for diabetes. IEEE intelligent systems, 30(3), 44-51. 416. Livengood, P., Maciejewski, R., Chen, W., & Ebert, D. S. (2012). OmicsVis: an interactive tool for visually analyzing metabolomics data. Bmc Bioinformatics, 13(8), S6. 417. Livnat, Y., Rhyne, T. M., & Samore, M. (2012). Epinome: A visual-analytics workbench for epidemiology data. IEEE computer graphics and applications, 32(2), 89-95. 418. Liyanage, H., Liaw, S. T., & De Lusignan, S. (2013). Accelerating the development of an information ecosystem in health care, by stimulating the growth of safe intermediate processing of health information (IPHI). Journal of Innovation in Health Informatics, 20(2), 82-86. 419. Liu, K., Li, L., Jiang, T., Chen, B., Jiang, Z., Wang, Z., & Gu, H. (2016). Chinese public attention to the outbreak of Ebola in West Africa: evidence from the online big data platform. International journal of environmental research and public health, 13(8), 780. 420. Lopez, M. E., Carberry, K., & Macias, C. (2015). Improving appendectomy outcomes using advanced analytics and team structures. Physician leadership journal, 2(6), 32-36. 421. Lori, N. F. et al. 2016 Processing Time Reduction: an Application in Living Human High-Resolution Diffusion Magnetic Resonance Imaging Data. J Med Syst 40(11):243. 422. Lopez, M. E., Carberry, K., & Macias, C. (2015). Improving appendectomy outcomes using advanced analytics and team structures. Physician leadership journal, 2(6), 32-36 423. Low, Y. S., Caster, O., Bergvall, T., Fourches, D., Zang, X., Norén, G. N., ...& Tropsha, A. (2015). Cheminformatics-aided pharmacovigilance: application to Stevens-Johnson Syndrome. Journal of the American Medical Informatics Association, 23(5), 968-978. 424. Ltifi, H., Ben Mohamed, E., & ben Ayed, M. (2016). Interactive visual knowledge discovery from data-based temporal decision support system. Information Visualization, 15(1), 31-50 425. Lukas, A., Kumbein, F., Temml, C., Mayer, B., & Oberbauer, R. (2003). Body mass index is the main risk factor for arterial hypertension in young subjects without major comorbidity. European journal of clinical investigation, 33(3), 223-230. 426. Luo, J., Wu, M., Gopukumar, D., & Zhao, Y. (2016). Big data application in biomedical research and health care: a literature review. Biomedical informatics insights, 8, BII- S31559.

227

427. Luo, L., Li, L., Hu, J., Wang, X., Hou, B., Zhang, T., & Zhao, L. P. (2016). A hybrid solution for extracting structured medical information from unstructured data in medical records via a double-reading/entry system. BMC medical informatics and decision making, 16(1), 114. 428. Lytkin, N. I., McVoy, L., Weitkamp, J. H., Aliferis, C. F., & Statnikov, A. (2011). Expanding the understanding of biases in development of clinical-grade molecular signatures: a case study in acute respiratory viral infections. PLoS One, 6(6), e20662. 429. Maccione, A., Gandolfo, M., Zordan, S., Amin, H., Di Marco, S., Nieus, T., ...& Berdondini, L. (2015). Microelectronics, bioinformatics and neurocomputation for massive neuronal recordings in brain circuits with large scale multielectrode array probes. Brain research bulletin, 119, 118-126. 430. Maciejewski, R., Hafen, R., Rudolph, S., Tebbetts, G., Cleveland, W. S., Grannis, S. J., & Ebert, D. S. (2009). Generating synthetic syndromic-surveillance data for evaluating visual-analytics techniques. IEEE Computer Graphics and Applications, 29(3). 431. Maciejewski, R., Livengood, P., Rudolph, S., Collins, T. F., Ebert, D. S., Brigantic, R. T., ...& Sanders, S. W. (2011). A pandemic influenza modeling and visualization tool. Journal of Visual Languages & Computing, 22(4), 268-278. 432. MacRae, J., Darlow, B., McBain, L., Jones, O., Stubbe, M., Turner, N., & Dowell, A. (2015). Accessing primary care Big Data: the development of a software algorithm to explore the rich content of consultation records. BMJ open, 5(8), e008160. 433. Madan, A., Cebrian, M., Moturu, S., & Farrahi, K. (2012). Sensing the" health state" of a community. IEEE Pervasive Computing, 11(4), 36-45. 434. Mahmud, S., Iqbal, R., & Doctor, F. (2016). Cloud enabled data analytics and visualization framework for health-shocks prediction. Future Generation Computer Systems, 65, 169-181. 435. Mäkinen, V. P., Soininen, P., Forsblom, C., Parkkonen, M., Ingman, P., Kaski, K., ...& Ala‐Korpela, M. (2008). 1H NMR metabonomics approach to the disease continuum of diabetic complications and premature death. Molecular systems biology, 4(1), 167. 436. Mamary, G. (2013). Making the leap to real-time analytics: Hunterdon Healthcare implements a new decision-support system. Health management technology, 34(2), 16. 437. Manchanda, R., & Jacobs, I. (2016). Genetic screening for gynecological cancer: where are we heading?. Future Oncology, 12(2), 207-220. 438. Manchanda, R., & Jacobs, I. (2016). Genetic screening for gynecological cancer: where are we heading?. Future Oncology, 12(2), 207-220. 439. Mane, K. K., Bizon, C., Owen, P., Gersing, K., Mostafa, J., & Schmitt, C. (2011). Patient Electronic Health Data–Driven Approach to Clinical Decision Support. Clinical and translational science, 4(5), 369-371.Mancini, M. (2014). Exploiting big data for improving healthcare services. Journal of e-Learning and Knowledge Society, 10(2). 440. Mane, K. K., Bizon, C., Owen, P., Gersing, K., Mostafa, J., & Schmitt, C. (2011). Patient Electronic Health Data–Driven Approach to Clinical Decision Support. Clinical and translational science, 4(5), 369-371. 441. Mane, K. K., Bizon, C., Owen, P., Gersing, K., Mostafa, J., & Schmitt, C. (2011). Patient Electronic Health Data–Driven Approach to Clinical Decision Support. Clinical and translational science, 4(5), 369-371.Mane, K. K., Bizon, C., Schmitt, C., Owen, P., Burchett, B., Pietrobon, R., & Gersing, K. (2012). VisualDecisionLinc: A visual analytics approach for comparative effectiveness-based clinical decision support in psychiatry. Journal of Biomedical Informatics, 45(1), 101-106. 442. Marés, J., Shamardin, L., Weiler, G., Anguita, A., Sfakianakis, S., Neri, E., ...& Coveney, P. V. (2014). p-medicine: A medical informatics platform for integrated large scale heterogeneous patient data. In AMIA Annual Symposium Proceedings (Vol. 2014, p. 872). American Medical Informatics Association. 443. Margolies, L. R., Pandey, G., Horowitz, E. R., & Mendelson, D. S. (2016). Breast imaging in the era of big data: structured reporting and data mining. American journal of roentgenology, 206(2), 259-264.

228

444. Marino, D. J. (2014). Using business intelligence to reduce the cost of care. Healthcare Financial Management, 68(3), 42-46. 445. Markowetz, A., Błaszkiewicz, K., Montag, C., Switala, C., & Schlaepfer, T. E. (2014). Psycho-informatics: big data shaping modern psychometrics. Medical hypotheses, 82(4), 405-411. 446. Marshall, D. A., Burgos-Liz, L., IJzerman, M. J., Crown, W., Padula, W. V., Wong, P. K., ... & Osgood, N. D. (2015). Selecting a dynamic simulation modeling method for health care delivery research—part 2: report of the ISPOR Dynamic Simulation Modeling Emerging Good Practices Task Force. Value in health, 18(2), 147-160. 447. Marshall, D. A., Burgos-Liz, L., Pasupathy, K. S., Padula, W. V., IJzerman, M. J., Wong, P. K.,& Osgood, N. D. (2016). Transforming healthcare delivery: integrating dynamic simulation modelling and big data in health economics and outcomes research. PharmacoEconomics, 34(2), 115-126. 448. Martin, C. M., Vogel, C., Grady, D., Zarabzadeh, A., Hederman, L., Kellett, J., & O’Shea, B. (2012). Implementation of complex adaptive chronic care: the P atient J ourney R ecord system (PaJR). Journal of Evaluation in Clinical Practice, 18(6), 1226-1234. 449. Martin, M., Champion, R., Kinsman, L., & Masman, K. (2011). Mapping patient flow in a regional Australian emergency department: A model driven approach. International Emergency Nursing, 19(2), 75-85. 450. Martinaa, M., & Vaithiyanadhan, V. (2015). Proxy Re-Encryption for Secure Data Storage in Clouds. Indian Journal of Science and Technology, 8(35). 451. Martínez, P., Martínez, J. L., Segura-Bedmar, I., Moreno-Schneider, J., Luna, A., & Revert, R. (2016). Turning user generated health-related content into actionable knowledge through text analytics services. Computers in Industry, 78, 43-56. 452. Martinez, R., Ordunez, P., Soliz, P. N., & Ballesteros, M. F. (2016). Data visualisation in surveillance for injury prevention and control: conceptual bases and case studies. Injury prevention, injuryprev-2015. 453. Maslove, D. M., Dubin, J. A., Shrivats, A., & Lee, J. (2016). Errors, omissions, and outliers in hourly vital signs measurements in intensive care. Critical care medicine, 44(11), e1021-e1030. 454. Mata, P., Chamney, A., Viner, G., Archibald, D., & Peyton, L. (2015). A development framework for mobile healthcare monitoring apps. Personal and Ubiquitous Computing, 19(3-4), 623-633. 455. Mathias, J. S., Agrawal, A., Feinglass, J., Cooper, A. J., Baker, D. W., & Choudhary, A. (2013). Development of a 5 year life expectancy index in older adults using predictive mining of electronic health record data. Journal of the American Medical Informatics Association, 20(e1), e118-e124. 456. Mavandadi, S., Dimitrov, S., Feng, S., Yu, F., Yu, R., Sikora, U., & Ozcan, A. (2012). Crowd-sourced BioGames: managing the big data problem for next-generation lab-on-a- chip platforms. Lab on a chip, 12(20), 4102-4106. 457. Mazor, I., Heart, T., & Even, A. (2016). Simulating the impact of an online digital dashboard in emergency departments on patients length of stay. Journal of Decision Systems, 25(sup1), 343-353. 458. McAnany, S. J., Anwar, M. A., & Qureshi, S. A. (2015). Decision analytic modeling in spinal surgery: a methodologic overview with review of current published literature. The Spine Journal, 15(10), 2254-2270. 459. McClay, W. A., Yadav, N., Ozbek, Y., Haas, A., Attias, H. T., & Nagarajan, S. S. (2015). A real-time magnetoencephalography brain-computer interface using interactive 3D visualization and the Hadoop ecosystem. Brain sciences, 5(4), 419-440. 460. McGirt, M. J., Sivaganesan, A., Asher, A. L., & Devin, C. J. (2015). Prediction model for outcome after low-back surgery: individualized likelihood of complication, hospital readmission, return to work, and 12-month improvement in functional disability. Neurosurgical focus, 39(6), E13.

229

461. McIntyre, R. S., Cha, D. S., Jerrell, J. M., Swardfager, W., Kim, R. D., Costa, L. G., ... & Brietzke, E. (2014). Advancing biomarker research: utilizing ‘Big Data’approaches for the characterization and prevention of bipolar disorder. Bipolar disorders, 16(5), 531-547. 462. McLaughlin, N., Afsar-manesh, N., Ragland, V., Buxey, F., & Martin, N. A. (2013). Tracking and sustaining improvement initiatives: leveraging quality dashboards to lead change in a neurosurgical department. Neurosurgery, 74(3), 235-244. 463. McNabb, M., Cao, Y., Devlin, T., Baxter, B., & Thornton, A. (2012, August). Measuring MERCI: exploring data mining techniques for examining the neurologic outcomes of stroke patients undergoing endo-vascular therapy at Erlanger Southeast Stroke Center. In Engineering in Medicine and Biology Society (EMBC), 2012 Annual International Conference of the IEEE (pp. 4704-4707). IEEE. 464. McNair, D. S. (2015). Enhancing nursing staffing forecasting with safety stock over lead time modeling. Nursing administration quarterly, 39(4), 291-296. 465. McRae, M. P., Simmons, G., Wong, J., & McDevitt, J. T. (2016). Programmable bio- nanochip platform: a point-of-care biosensor system with the capacity to learn. Accounts of chemical research, 49(7), 1359-1368. 466. Medrano-Gracia, P., Cowan, B. R., Suinesiaputra, A., & Young, A. A. (2015). Challenges of cardiac image analysis in large-scale population-based studies. Current cardiology reports, 17(3), 9. 467. Mei, K., Peng, J., Gao, L., Zheng, N. N., & Fan, J. (2015). Hierarchical classification of large-scale patient records for automatic treatment stratification. IEEE journal of biomedical and health informatics, 19(4), 1234-1245. 468. Meldolesi, E., Van Soest, J., Damiani, A., Dekker, A., Alitto, A. R., Campitelli, M., & Lambin, P. (2016). Standardized data collection to build prediction models in oncology: a prototype for rectal cancer. Future Oncology, 12(1), 119-136. 469. Menger, V., Spruit, M., Hagoort, K., & Scheepers, F. (2016). Transitioning to a data driven mental health practice: Collaborative expert sessions for knowledge and hypothesis finding. Computational and mathematical methods in medicine, 2016. 470. Merelli, I., Pérez-Sánchez, H., Gesing, S., & D’Agostino, D. (2014). Managing, analysing, and integrating big data in medical bioinformatics: open problems and future perspectives. BioMed research international, 2014. 471. Mettler, T., & Vimarlund, V. (2009). Understanding business intelligence in the context of healthcare. Health informatics journal, 15(3), 254-264. 472. Metz, G. A., Ng, J. W., Kovalchuk, I., & Olson, D. M. (2015). Ancestral experience as a game changer in stress vulnerability and disease outcomes. Bioessays, 37(6), 602-611. 473. Meyer, A. M., Olshan, A. F., Green, L., Meyer, A., Wheeler, S. B., Basch, E., & Carpenter, W. R. (2014). Big Data for Population-Based Cancer Research The Integrated Cancer Information and Surveillance System. North Carolina medical journal, 75(4), 265- 269. 474. Mezghani, E., Exposito, E., Drira, K., Da Silveira, M., & Pruski, C. (2015). A semantic big data platform for integrating heterogeneous wearable data in healthcare. Journal of medical systems, 39(12), 185. 475. Miller, A. R., & Tucker, C. (2014). Health information exchange, system size and information silos. Journal of health economics, 33, 28-42. 476. Miriovsky, B. J., Shulman, L. N., & Abernethy, A. P. (2012). Importance of health information technology, electronic health records, and continuously aggregating data to comparative effectiveness research and learning health care. Journal of Clinical Oncology, 30(34), 4243-4248. 477. Mirkes, E. M., Coats, T. J., Levesley, J., & Gorban, A. N. (2016). Handling missing data in large healthcare dataset: A case study of unknown trauma outcomes. Computers in biology and medicine, 75, 203-216. 478. Misiunas, N., Oztekin, A., Chen, Y., & Chandra, K. (2016). DEANN: A healthcare analytic methodology of data envelopment analysis and artificial neural networks for the prediction of organ recipient functional status. Omega, 58, 46-54.

230

479. Mistry, P., Neagu, D., Trundle, P. R., & Vessey, J. D. (2016). Using random forest and decision tree models for a new vehicle prediction approach in computational toxicology. Soft Computing, 20(8), 2967-2979. 480. Moghimi, F. H., De Steiger, R., Schaffer, J., & Wickramasinghe, N. (2013). The benefits of adopting e-performance management techniques and strategies to facilitate superior healthcare delivery: the proffering of a conceptual framework for the context of Hip and Knee Arthroplasty. Health and technology, 3(3), 237-247. 481. Mohammed, E. A., Far, B. H., & Naugler, C. (2014). Applications of the MapReduce programming framework to clinical big data analysis: current landscape and future trends. BioData mining, 7(1), 22. 482. Mohan, M. et al. (2016). Disease diagnosis for personalized health care using map reduce technique.Source of the Document International Journal of Control Theory and Applications 483. Mongkolwat, P., Kleper, V., Talbot, S., & Rubin, D. (2014). The National Cancer Informatics Program (NCIP) Annotation and Image Markup (AIM) Foundation Model. Journal of digital imaging, 27(6), 692-701. 484. Monteith, S., Glenn, T., Geddes, J., Whybrow, P. C., & Bauer, M. (2016). Big data for bipolar disorder. International journal of bipolar disorders, 4(1), 10. 485. Motai, Y., Ma, D., Docef, A., & Yoshida, H. (2015). Smart Colonography for Distributed Medical Databases with Group Kernel Feature Analysis. ACM Transactions on Intelligent Systems and Technology (TIST), 6(4), 58. 486. Mudunuri, U. S., Khouja, M., Repetski, S., Venkataraman, G., Che, A., Luke, B. T., & Stephens, R. M. (2013). Knowledge and theme discovery across very large biological data sets using distributed queries: a prototype combining unstructured and structured data. PloS one, 8(12), e80503. 487. Müller, H., Reihs, R., Zatloukal, K., & Holzinger, A. (2014). Analysis of biomedical data with multilevel glyphs. BMC bioinformatics, 15(6), S5. 488. Müller, H., Reihs, R., Zatloukal, K., Jeanquartier, F., Merino-Martinez, R., van Enckevort, D., ...& Holzinger, A. (2015). State-of-the-art and future challenges in the integration of biobank catalogues. In Smart Health (pp. 261-273). Springer, Cham. 489. Munos, B., Baker, P. C., Bot, B. M., Crouthamel, M., Vries, G., Ferguson, I., & Ozcan, A. (2016). Mobile health: the power of wearables, sensors, and apps to transform clinical trials. Annals of the New York Academy of Sciences, 1375(1), 3-18. 490. Musen, M. A et al. (2012)The National Center for Biomedical Ontology. Journalof American Medical Information Association19(2):190-5 491. Mwangi, B., Soares, J. C., & Hasan, K. M. (2014). Visualization and unsupervised predictive clustering of high-dimensional multimodal neuroimaging data. Journal of neuroscience methods, 236, 19-25. 492. Mwangi, B., Wu, M. J., Cao, B., Passos, I. C., Lavagnino, L., Keser, Z., & Soares, J. C. (2016). Individualized prediction and clinical staging of bipolar disorders using neuroanatomical biomarkers. Biological psychiatry: cognitive neuroscience and neuroimaging, 1(2), 186-194. 493. Nagy, P. G., Warnock, M. J., Daly, M., Toland, C., Meenan, C. D., & Mezrich, R. S. (2009). Informatics in radiology: automated Web-based graphical dashboard for radiology operational business intelligence. Radiographics, 29(7), 1897-1906. 494. Nakas, C. T., Schütz, N., Werners, M., & Leichtle, A. B. (2016). Accuracy and calibration of computational approaches for inpatient mortality predictive modeling. PloS one, 11(7), e0159046. 495. Nash, M., Pestrue, J., Geier, P., Sharp, K., Helder, A., & McAlearney, A. S. (2010). Leveraging information technology to drive improvement in patient satisfaction. Journal for Healthcare Quality, 32(5), 30-40. 496. Nath, C., Albaghdadi, M. S., & Jonnalagadda, S. R. (2016). A natural language processing tool for large-scale data extraction from echocardiography reports. PloS one, 11(4), e0153749.

231

497. Nepal, S., Ranjan, R., & Choo, K. K. R. (2015). Trustworthy processing of healthcare big data in hybrid clouds. IEEE Cloud Computing, 2(2), 78-84. 498. Newman, D., Herrera, C. N., & Parente, S. T. (2014). Overcoming barriers to a research-ready national commercial claims database. The American journal of managed care, 20(11 Spec No. 17), eSP25-30. 499. Ng, K., Ghoting, A., Steinhubl, S. R., Stewart, W. F., Malin, B., & Sun, J. (2014). PARAMO: a PARAllel predictive MOdeling platform for healthcare analytic research using electronic health records. Journal of biomedical informatics, 48, 160-170. 500. Nielson, J. L., Guandique, C. F., Liu, A. W., Burke, D. A., Lash, A. T., Moseanko, R., & Brock, J. H. (2014). Development of a database for translational spinal cord injury research. Journal of neurotrauma, 31(21), 1789-1799. 501. Nielson, J. L., Haefeli, J., Salegio, E. A., Liu, A. W., Guandique, C. F., Stück, E. D., ... & Brock, J. H. (2015). Leveraging biomedical informatics for assessing plasticity and repair in primate spinal cord injury. Brain research, 1619, 124-138. 502. Nielson, J. L., Paquette, J., Liu, A. W., Guandique, C. F., Tovar, C. A., Inoue, T., ... & Lum, P. Y. (2015). Topological data analysis for discovery in preclinical spinal cord injury and traumatic brain injury. Nature communications, 6, 8581. 503. Nippert, K. H., & Graves, B. (2012). Transforming labor-management practices through real-time analytics. Healthcare financial management: journal of the Healthcare Financial Management Association, 66(6), 118-20. 504. Noor, A. M., Holmberg, L., Gillett, C., & Grigoriadis, A. (2015). Big Data: the challenge for small research groups in the era of cancer genomics. British journal of cancer, 113(10), 1405. 505. Nouraei, S. A. R., Virk, J. S., Hudovsky, A., Wathen, C., Darzi, A., & Parsons, D. (2015). Accuracy of clinician-clinical coder information handover following acute medical admissions: implication for using administrative datasets in clinical outcomes management. Journal of Public Health, 38(2), 352-362. 506. O’Connor, K., Pimpalkhute, P., Nikfarjam, A., Ginn, R., Smith, K. L., & Gonzalez, G. (2014). Pharmacovigilance on twitter? Mining tweets for adverse drug reactions. In AMIA annual symposium proceedings (Vol. 2014, p. 924). American Medical Informatics Association. 507. Obeid, I., & Picone, J. (2016). The temple university hospital eeg data corpus. Frontiers in neuroscience, 10, 196. 508. Oberg, A. L., McKinney, B. A., Schaid, D. J., Pankratz, V. S., Kennedy, R. B., & Poland, G. A. (2015). Lessons learned in the analysis of high-dimensional data in vaccinomics. Vaccine, 33(40), 5262-5270. 509. O'Dea, B., Wan, S., Batterham, P. J., Calear, A. L., Paris, C., & Christensen, H. (2015). Detecting suicidality on Twitter. Internet Interventions, 2(2), 183-188. 510. O’Driscoll, A., Belogrudov, V., Carroll, J., Kropp, K., Walsh, P., Ghazal, P., & Sleator, R. D. (2015). HBLAST: Parallelised sequence similarity–A Hadoop MapReducable basic local alignment search tool. Journal of biomedical informatics, 54, 58-64. 511. O’Driscoll, A., Daugelaite, J., & Sleator, R. D. (2013). ‘Big data’, Hadoop and cloud computing in genomics. Journal of biomedical informatics, 46(5), 774-781. 512. Oh, W., Kim, E., Castro, M. R., Caraballo, P. J., Kumar, V., Steinbach, M. S., & Simon, G. J. (2016). Type 2 diabetes mellitus trajectories and associated risks. Big data, 4(1), 25- 30. 513. Okimoto, G., Zeinalzadeh, A., Wenska, T., Loomis, M., Nation, J. B., Fabre, T., & Kwee, S. (2016). Joint analysis of multiple high-dimensional data types using sparse matrix approximations of rank-1 with applications to ovarian and liver cancer. BioData mining, 9(1), 24. 514. Omar, A. M. S., Bansal, M., & Sengupta, P. P. (2016). Advances in echocardiographic imaging in heart failure with reduced and preserved ejection fraction. Circulation research, 119(2), 357-374. 515. Ow, G. S., Tang, Z., & Kuznetsov, V. A. (2016). Big data and computational biology strategy for personalized prognosis. Oncotarget, 7(26), 40200.

232

516. Özdemir, V., & Kolker, E. (2016). Precision nutrition 4.0: A big data and ethics foresight analysis—Convergence of agrigenomics, nutrigenomics, nutriproteomics, and nutrimetabolomics. Omics: a journal of integrative biology, 20(2), 69-75. 517. Ozkaynak, M., Dziadkowiec, O., Mistry, R., Callahan, T., He, Z., Deakyne, S., & Tham, E. (2015). Characterizing workflow for pediatric asthma patients in emergency departments using electronic health records. Journal of biomedical informatics, 57, 386- 398. 518. Ozminkowski, R. J., Wells, T. S., Hawkins, K., Bhattarai, G. R., Martel, C. W., & Yeh, C. S. (2015). Big data, little data, and care coordination for medicare beneficiaries with medigap coverage. Big data, 3(2), 114-125. 519. Pah, A. R., Rasmussen-Torvik, L. J., Goel, S., Greenland, P., & Kho, A. N. (2015). Big data: what is it and what does it mean for cardiovascular research and prevention policy. Current Cardiovascular Risk Reports, 9(1), 424. 520. Pan, T., & Fang, K. (2010). An Effective Information Support System for Medical Management: Indicator Based Intelligence System. International Journal of Computers and Applications, 32(1), 119-124. 521. Pan, X. (2016). Application of improved ant colony algorithm in intelligent medical system: from the perspective of big data. CHEMICAL ENGINEERING, 51. 522. Panaccio, M. P., Cummins, G., Wentworth, C., Lanes, S., Reynolds, S. L., Reynolds, M. W., & Koren, A. (2015). A common data model to assess cardiovascular hospitalization and mortality in atrial fibrillation patients using administrative claims and medical records. Clinical epidemiology, 7, 77. 523. Park, S. J., Saito‐Adachi, M., Komiyama, Y., & Nakai, K. (2016). Advances, practice, and clinical perspectives in high‐throughput sequencing. Oral diseases, 22(5), 353-364. 524. Park, S. L., Pantanowitz, L., & Parwani, A. V. (2013). Quality assurance in anatomic pathology. Diagnostic Histopathology, 19(12), 438-446. 525. Passos, I. C., Mwangi, B., Cao, B., Hamilton, J. E., Wu, M. J., Zhang, X. Y., & Soares, J. C. (2016). Identifying a clinical signature of suicidality among patients with mood disorders: a pilot study using a machine learning approach. Journal of affective disorders, 193, 109-116. 526. Patadia, V. K., Nimke, D., Stefansdottir, G., Benjoya, J., Allard-Schrijer, L., Holbrook, C., ...& Saccomanno, C. F. (2015). A Business Intelligence Solution to Pharmacovigilance Signal Tracking and Management: One Mid-Size Pharma’s Experience. Pharmaceutical Medicine, 29(4), 197-201. 527. Patching, H. M., Hudson, L. M., Cooke, W., Garcia, A. J., Hay, S. I., Roberts, M., & Moyes, C. L. (2015). A supervised learning process to validate online disease reports for use in predictive models. Big data, 3(4), 230-237. 528. Paten, B., Diekhans, M., Druker, B. J., Friend, S., Guinney, J., Gassner, N., & Massie, M. (2015). The NIH BD2K center for big data in translational genomics. Journal of the American Medical Informatics Association, 22(6), 1143-1147. 529. Paulus, J. K., Wessler, B. S., Lundquist, C., Lai, L. L., Raman, G., Lutz, J. S., & Kent, D. M. (2016). Field synopsis of sex in clinical prediction models for cardiovascular disease. Circulation: Cardiovascular Quality and Outcomes, 9(2 suppl 1), S8-S15. 530. Pendry, K. (2015). The use of big data in transfusion medicine. Transfusion medicine, 25(3), 129-137. 531. Percival, J., McGregor, C., Percival, N., & James, A. (2015). Enabling the integration of clinical event and physiological data for real-time and retrospective analysis. Information Systems and e-Business Management, 13(4), 693-711. 532. Perer, A., Wang, F., & Hu, J. (2015). Mining and exploring care pathways from electronic medical records with visual analytics. Journal of biomedical informatics, 56, 369- 378. 533. Perer, A., & Sun, J. (2012). MatrixFlow: temporal network visual analytics to track symptom evolution during disease progression. In AMIA annual symposium proceedings (Vol. 2012, p. 716). American Medical Informatics Association.

233

534. Petushi, S., Marker, J., Zhang, J., Zhu, W., Breen, D., Chen, C., & Garcia, F. U. (2008). A visual analytics system for breast tumor evaluation. Analytical and quantitative cytology and histology, 30(5), 279-290. 535. Pfeifer, L., Stein, K., Fink, U., Welker, A., Wetzl, B., Bastian, P., & Wolfbeis, O. S. (2005). Improved routine bio-medical and bio-analytical online fluorescence measurements using fluorescence lifetime resolution. Journal of fluorescence, 15(3), 423-432. 536. Pfeiffer, D. U., & Stevens, K. B. (2015). Spatial and temporal epidemiological analysis in the Big Data era. Preventive veterinary medicine, 122(1-2), 213-220. 537. Phillips, K. A., Trosman, J. R., Kelley, R. K., Pletcher, M. J., Douglas, M. P., & Weldon, C. B. (2014). Genomic sequencing: assessing the health care system, policy, and big-data implications. Health affairs, 33(7), 1246-1253. 538. Piette, J. D., Sussman, J. B., Pfeiffer, P. N., Silveira, M. J., Singh, S., & Lavieri, M. S. (2013). Maximizing the value of mobile health monitoring by avoiding redundant patient reports: Prediction of depression-related symptoms and adherence problems in automated health assessment services. Journal of medical internet research, 15(7). 539. Pillai, R. R., Divekar, R., Brasier, A., Bhavnani, S., & Calhoun, W. J. (2012). Strategies for molecular classification of asthma using bipartite network analysis of cytokine expression. Current allergy and asthma reports, 12(5), 388-395. 540. Pinsky, M. R., & Dubrawski, A. (2014). Gleaning knowledge from data in the intensive care unit. American journal of respiratory and critical care medicine, 190(6), 606-610. 541. Pivovarov, R., Albers, D. J., Hripcsak, G., Sepulveda, J. L., & Elhadad, N. (2014). Temporal trends of hemoglobin A1c testing. Journal of the American Medical Informatics Association, 21(6), 1038-1044. 542. Pollom, R. K., Balbach, J., & Jones, K. A. (2007). Clinical analytics equal better systemwide outcomes. Nursing management, 38(12), 44-46.. 543. Poon, C. C., Lo, B. P., Yuce, M. R., Alomainy, A., & Hao, Y. (2015). Body sensor networks: In the era of big data and beyond. IEEE reviews in biomedical engineering, 8, 4- 16. 544. Popescu, D., & Borangiu, A. (2014). Digital signal processing for knowledge based sonotubometry of eustachian tube function. Journal of Control Engineering and Applied Informatics, 16(3), 56-64. 545. Post, A. R., Kurc, T., Cholleti, S., Gao, J., Lin, X., Bornstein, W., & Saltz, J. H. (2013). The Analytic Information Warehouse (AIW): A platform for analytics using electronic health record data. Journal of biomedical informatics, 46(3), 410-424. 546. Post, A. R., Kurc, T., Willard, R., Rathod, H., Mansour, M., Pai, A. K., & Saltz, J. H. (2013). Temporal abstraction-based clinical phenotyping with eureka!. In AMIA Annual Symposium Proceedings (Vol. 2013, p. 1160). American Medical Informatics Association. 547. Potharaju, S. P., & Sreedevi, M. (2016). An improved prediction of kidney disease using SMOTE. Indian Journal of Science and Technology, 9(31). 548. Potter, M. A., Schuh, R. G., Pomer, B., & Stebbins, S. (2013). The adaptive response metric: toward an all-hazards tool for planning, decision support, and after-action analytics. Journal of public health management and practice, 19, S49-S54. 549. Poulin, C., Shiner, B., Thompson, P., Vepstas, L., Young-Xu, Y., Goertzel, B., & McAllister, T. (2014). Predicting the risk of suicide by analyzing the text of clinical notes. PloS one, 9(1), e85733. 550. Powell, L., Hindman, A., & McMillan, B. (2009). Responsibility-based A/R reporting: how one health system drove performance with analytics: analytics can help organizations produce meaningful reports that engage management and staff alike and drive A/R performance. Healthcare Financial Management, 63(9), 54-60. 551. Preissner, S., Kostka, E., Mokross, M., Kersten, N. V., Blunck, U., & Preissner, R. (2015). DBEndo: a web-based endodontic case management tool. BMC research notes, 8(1), 685. 552. Premarathne, U., Abuadbba, A., Alabdulatif, A., Khalil, I., Tari, Z., Zomaya, A., & Buyya, R. (2016). Hybrid cryptographic access control for cloud-based EHR systems. IEEE Cloud Computing, (4), 58-64.

234

553. Prevedello, L. M., Andriole, K. P., Hanson, R., Kelly, P., & Khorasani, R. (2010). Business intelligence tools for radiology: creating a prototype model using open-source tools. Journal of digital imaging, 23(2), 133-141. 554. Price, L. E., Shea, K., & Gephart, S. (2015). The Veterans Affairs's corporate data Warehouse: uses and implications for nursing research and practice. Nursing administration quarterly, 39(4), 311-318. 555. Priya, M., & Ranjith Kumar, P. (2015). A novel intelligent approach for predicting atherosclerotic individuals from big data for healthcare. International Journal of Production Research, 53(24), 7517-7532. 556. Procházka, A., Schätz, M., Vyšata, O., & Vališ, M. (2016). Microsoft kinect visual and depth sensors for breathing and heart rate analysis. Sensors, 16(7), 996. 557. Puppala, M., He, T., Chen, S., Ogunti, R., Yu, X., Li, F., ...& Wong, S. T. (2015). METEOR: an enterprise health informatics environment to support evidence-based medicine. IEEE Transactions on Biomedical Engineering, 62(12), 2776-2786. 558. Purkayastha, S., & Braa, J. (2013). Big data analytics for developing countries–using the cloud for operational BI in health. The Electronic Journal of Information Systems in Developing Countries, 59(1), 1-17. 559. Qin, C., Tao, L., Phang, Y. H., Zhang, C., Chen, S. Y., Zhang, P., ...& Chen, Y. Z. (2015). The assessment of the readiness of molecular biomarker-based mobile health technologies for healthcare applications. Scientific reports, 5, 17854. 560. Quinn, C. C., Clough, S. S., Minor, J. M., Lender, D., Okafor, M. C., & Gruber-Baldini, A. (2008). WellDoc™ mobile diabetes management randomized controlled trial: change in clinical and behavioral outcomes and patient and physician satisfaction. Diabetes technology & therapeutics, 10(3), 160-168. 561. Quwaider, M., & Jararweh, Y. (2016). A cloud supported model for efficient community health awareness. Pervasive and Mobile Computing, 28, 35-50. 562. Raghu, A., Praveen, D., Peiris, D., Tarassenko, L., & Clifford, G. (2015). Engineering a mobile health tool for resource-poor settings to assess and manage cardiovascular disease risk: SMARThealth study. BMC medical informatics and decision making, 15(1), 36. 563. Raghupathi, V., & Raghupathi, W. (2013). Exploring the relationship between ICTs and public health at country level: a health analytics approach. International Journal of Healthcare Information Systems and Informatics (IJHISI), 8(3), 1-22. 564. Raghupathi, V., & Raghupathi, W. (2014). An Unstructured Information Management Architecture Approach to Text Analytics of Cancer Blogs. International Journal of Healthcare Information Systems and Informatics (IJHISI), 9(2), 16-33. 565. Raidou, R. G., van der Heide, U. A., Dinh, C. V., Ghobadi, G., Kallehauge, J. F., Breeuwer, M., & Vilanova, A. (2015, June). Visual analytics for the exploration of tumor tissue characterization. In Computer Graphics Forum (Vol. 34, No. 3, pp. 11-20). 566. Rajalakshmi, K., & Nirmala, K. (2016). Heart disease prediction with mapreduce by using weighted association classifier and k-means. Indian Journal of Science and Technology, 9(19). 567. Rajkumar, N., Vimal, K. R., Nathiya, M., & Silambarasan, K. (2014). Mining association rules in big data for e-healthcare information system. Research Journal of Applied Sciences, Engineering and Technology, 8(8), 1002-1008. 568. Ram, S., Zhang, W., Williams, M., & Pengetnze, Y. (2015). Predicting asthma-related emergency department visits using big data. IEEE J. Biomedical and Health Informatics, 19(4), 1216-1223. 569. Ramnath, V. R., Ho, L., Maggio, L. A., & Khazeni, N. (2014). Centralized monitoring and virtual consultant models of tele-ICU care: a systematic review. Telemedicine and e- Health, 20(10), 936-961. 570. Ratliff, J. K., Balise, R., Veeravagu, A., Cole, T. S., Cheng, I., Olshen, R. A., & Tian, L. (2016). Predicting occurrence of spine surgery complications using “big data” modeling of an administrative claims database. JBJS, 98(10), 824-834.

235

571. Ratwani, R. M., & Fong, A. (2014). ‘Connecting the dots’: leveraging visual analytics to make sense of patient safety event reports. Journal of the American Medical Informatics Association, 22(2), 312-317. 572. Razavian, N., Blecker, S., Schmidt, A. M., Smith-McLallen, A., Nigam, S., & Sontag, D. (2015). Population-level prediction of type 2 diabetes from claims data and analysis of risk factors. Big Data, 3(4), 277-287. 573. Reilly-Harrington, N. A., Sylvia, L. G., Rabideau, D. J., Gold, A. K., Deckersbach, T., Bowden, C. L., ...& Friedman, E. S. (2016). Tracking medication changes to assess outcomes in comparative effectiveness research: A bipolar CHOICE study. Journal of affective disorders, 205, 159-164. 574. Ren, G., & Krawetz, R. (2015). Applying computation biology and “big data” to develop multiplex diagnostics for complex chronic diseases such as osteoarthritis. Biomarkers, 20(8), 533-539. 575. Rexit, R., Tsui, F. R., Espino, J., Chrysanthis, P. K., Wesaratchakit, S., & Ye, Y. (2015). An analytics appliance for identifying (near) optimal over-the-counter medicine products as health indicators for influenza surveillance. Information Systems, 48, 151-163. 576. Richardson, A., Signor, B. M., Lidbury, B. A., & Badrick, T. (2016). Clinical chemistry in higher dimensions: machine-learning and enhanced prediction from routine clinical chemistry data. Clinical biochemistry, 49(16-17), 1213-1220. 577. Ritt, E. (2013). Embedding a culture of analytics in nursing practice. Nurse Leader, 11(3), 48-50. 578. Robbins, M. J., & Jacobson, S. H. (2015). Analytics for vaccine economics and pricing: insights and observations. Expert review of vaccines, 14(4), 605-616. 579. Robertson, S. P., Quon, H., Kiess, A. P., Moore, J. A., Yang, W., Cheng, Z., ...& Sharabi, A. (2015). A data‐mining framework for large scale analysis of dose‐outcome relationships in a database of irradiated head and neck cancer patients. Medical physics, 42(7), 4329-4337. 580. Robson, B., & Boray, S. (2015). Implementation of a web based universal exchange and inference language for medicine: Sparse data, probabilities and inference in data mining of clinical data repositories. Computers in biology and medicine, 66, 82-102. 581. Rodger, J. A. (2015). Discovery of medical Big Data analytics: Improving the prediction of traumatic brain injury survival rates by data mining Patient Informatics Processing Software Hybrid Hadoop Hive. Informatics in Medicine Unlocked, 1, 17-26. 582. Rodrigues Jr, J. F., Paulovich, F. V., de Oliveira, M. C., & de Oliveira Jr, O. N. (2016). On the convergence of nanotechnology and Big Data analysis for computer-aided diagnosis. Nanomedicine, 11(8), 959-982. 583. Rojas, C. C., Patton, R. M., & Beckerman, B. G. (2011). Characterizing mammography reports for health analytics. Journal of medical systems, 35(5), 1197-1210. 584. Ronquillo, J. G., Baer, M. R., & Lester, W. T. (2016). Sex-specific patterns and differences in dementia and Alzheimer’s disease using informatics approaches. Journal of women & aging, 28(5), 403-411. 585. Roski, J., Bo-Linn, G. W., & Andrews, T. A. (2014). Creating value in health care through big data: opportunities and policy implications. Health affairs, 33(7), 1115-1122. 586. Rosow, E., Adam, J., Coulombe, K., Race, K., & Anderson, R. (2003). Virtual instrumentation and real-time executive dashboards: Solutions for health care systems. Nursing Administration Quarterly, 27(1), 58-76. 587. Roy, S., LaFramboise, W. A., Nikiforov, Y. E., Nikiforova, M. N., Routbort, M. J., Pfeifer, J., ...& Pantanowitz, L. (2016). Next-generation sequencing informatics: challenges and strategies for implementation in a clinical environment. Archives of pathology & laboratory medicine, 140(9), 958-975. 588. Rumpf, R. W., Wolock, S. L., & Ray, W. C. (2014). StickWRLD as an interactive visual pre-filter for canceromics-centric expression quantitative trait locus data. Cancer informatics, 13, CIN-S14024. 589. Rumsfeld, J. S., Joynt, K. E., & Maddox, T. M. (2016). Big data analytics to improve cardiovascular care: promise and challenges. Nature Reviews Cardiology, 13(6), 350.

236

590. Rusin, C. G., Acosta, S. I., Shekerdemian, L. S., Vu, E. L., Bavare, A. C., Myers, R. B., ... & Penny, D. J. (2016). Prediction of imminent, severe deterioration of children with parallel circulations using real-time processing of physiologic data. The Journal of thoracic and cardiovascular surgery, 152(1), 171-177. 591. Ryan, J., Doster, B., Daily, S., & Lewis, C. (2014). A balanced perspective to perioperative process management aligned to hospital strategy. International Journal of Healthcare Information Systems and Informatics (IJHISI), 9(4), 1-19. 592. Ryan, J., Hendler, J., & Bennett, K. P. (2015). Understanding Emergency Department 72-hour revisits among medicaid patients using electronic healthcare records. Big data, 3(4), 238-248. 593. Sacks, J. A., Zehe, E., Redick, C., Bah, A., Cowger, K., Camara, M., ...& Liu, A. (2015). Introduction of mobile health tools to support Ebola surveillance and contact tracing in Guinea. Global Health: Science and Practice, 3(4), 646-659. 594. Sáez, C., Rodrigues, P. P., Gama, J., Robles, M., & García-Gómez, J. M. (2015). Probabilistic change detection and visualization methods for the assessment of temporal stability in biomedical data quality. Data Mining and Knowledge Discovery, 29(4), 950- 975. 595. Safari, L., & Patrick, J. D. (2014). Restricted natural language based querying of clinical databases. Journal of biomedical Informatics, 52, 338-353. 596. Safari, L., & Patrick, J. D. (2013, July). Mapping query terms to data and schema using content based similarity search in clinical information systems. In Engineering in Medicine and Biology Society (EMBC), 2013 35th Annual International Conference of the IEEE (pp. 4779-4782). IEEE. 597. Safari, L., & Patrick, J. D. (2013, July). A temporal model for clinical data analytics language. In Engineering in Medicine and Biology Society (EMBC), 2013 35th Annual International Conference of the IEEE (pp. 3218-3221). IEEE. 598. Sahoo, S. S., Jayapandian, C., Garg, G., Kaffashi, F., Chung, S., Bozorgi, A., ...& Zhang, G. Q. (2013). Heart beats in the cloud: distributed analysis of electrophysiological ‘Big Data’using cloud computing for epilepsy clinical research. Journal of the American Medical Informatics Association, 21(2), 263-271. 599. Sahoo, S. S., Wei, A., Valdez, J., Wang, L., Zonjy, B., Tatsuoka, C., ...& Lhatoo, S. D. (2016). NeuroPigPen: A Scalable Toolkit for Processing Electrophysiological Signal Data in Neuroscience Applications Using Apache Pig. Frontiers in neuroinformatics, 10, 18. 600. Sakr, S., & Elgammal, A. (2016). Towards a comprehensive data analytics framework for smart healthcare services. Big Data Research, 4, 44-58. 601. Salem, J., Borgmann, H., Bultitude, M., Fritsche, H. M., Haferkamp, A., Heidenreich, A., ...& Tsaur, I. (2016). Online discussion on# KidneyStones: a longitudinal assessment of activity, users and content. PloS one, 11(8), e0160863. 602. Salomi, M., & Balamurugan, S. A. A. (2016). Need, Application and Characteristics of Big Data Analytics in Healthcare-A Survey. Indian Journal of Science and Technology, 9(16). 603. Samuels, J. G., McGrath, R. J., Fetzer, S. J., Mittal, P., & Bourgoine, D. (2015). Using the electronic health record in nursing research: Challenges and opportunities. Western journal of nursing research, 37(10), 1284-1294. 604. Carrasco, R. S. M. (2016). Detection of Adverse Reaction to Drugs in Elderly Patients through Predictive Modeling. IJIMAI, 3(6), 52-56. 605. Sanchez-Morillo, D., Fernandez-Granero, M. A., & Leon-Jimenez, A. (2016). Use of predictive algorithms in-home monitoring of chronic obstructive pulmonary disease and asthma: a systematic review. Chronic respiratory disease, 13(3), 264-283. 606. Sandhu, R., Sood, S. K., & Kaur, G. (2016). An intelligent system for predicting and preventing MERS-CoV infection outbreak. The Journal of Supercomputing, 72(8), 3033- 3056. 607. Sapna, S., & Kumar, M. P. (2015). Diagnosis of disease from clinical big data using neural network. Indian Journal of Science and Technology, 8(24).

237

608. Satagopam, V., Gu, W., Eifes, S., Gawron, P., Ostaszewski, M., Gebel, S., ...& Schneider, R. (2016). Integration and visualization of translational medicine data for better understanding of human diseases. Big data, 4(2), 97-108. 609. Sathiyavathi, R. (2015). A Survey: Big Data Analytics on Healthcare System. Contemporary Engineering Sciences, 8(3), 121-125. 610. Sawand, A., Djahel, S., Zhang, Z., & Nait-Abdesselam, F. (2015). Toward energy- efficient and trustworthy eHealth monitoring system. China Communications, 12. 611. Schatz, B. R. (2015). National Surveys of population health: big data analytics for mobile health monitors. Big Data, 3(4), 219-229. 612. Schilsky, R. L., Michels, D. L., Kearbey, A. H., Yu, P. P., & Hudis, C. A. (2014). Building a rapid learning health care system for oncology: the regulatory framework of CancerLinQ. J Clin Oncol, 32(22), 2373-9. 613. Schissler, A. G., Gardeux, V., Li, Q., Achour, I., Li, H., Piegorsch, W. W., & Lussier, Y. A. (2015). Dynamic changes of RNA-sequencing expression for precision medicine: N- of-1-pathways Mahalanobis distance within pathways of single subjects predicts breast cancer survival. Bioinformatics, 31(12), i293-i302. 614. Schmeling, H., Horneff, G., Benseler, S. M., & Fritzler, M. J. (2014). Pharmacogenetics: can genes determine treatment efficacy and safety in JIA?.Nature Reviews Rheumatology, 10(11), 682. 615. Schneeweiss, S., Shrank, W. H., Ruhl, M., & Maclure, M. (2015). Decision-making aligned with rapid-cycle evaluation in health care. International journal of technology assessment in health care, 31(4), 214-222. 616. Schouten, P. (2013) Big data in health care. Healthcare financial management: journal of the Healthcare Financial Management Association 617. Schuh, R. G., Basque, M., & Potter, M. A. (2014). The effects of funding change and reorganization on patterns of emergency response in a local health agency. Public Health Reports, 129(6_suppl4), 166-172. 618. Schultz, T. J., Crock, C., Hansen, K., Deakin, A., & Gosbell, A. (2014). Piloting an online incident reporting system in A ustralasian emergency medicine. Emergency Medicine Australasia, 26(5), 461-467. 619. Schumacher, A., Rujan, T., & Hoefkens, J. (2014). A collaborative approach to develop a multi-omics data analytics platform for translational research. Applied & translational genomics, 3(4), 105-108. 620. Sedlmayr, M., Würfl, T., Maier, C., Häberle, L., Fasching, P., Prokosch, H. U., & Christoph, J. (2016). Optimizing R with SparkR on a commodity cluster for biomedical research. computer methods and programs in biomedicine, 137, 321-328. 621. Sehgal, V., Seviour, E. G., Moss, T. J., Mills, G. B., Azencott, R., & Ram, P. T. (2015). Robust selection algorithm (RSA) for multi-omic biomarker discovery; integration with functional network analysis to identify miRNA regulated pathways in multiple cancers. PloS one, 10(10), e0140072. 622. Selman, D. H. (2013). Paradigm of prediction: Predictive analytics to prevent congestive heart failure (Doctoral dissertation, Walden University). 623. Sengupta, P. P. (2013). Intelligent platforms for disease assessment: novel approaches in functional echocardiography. JACC: Cardiovascular Imaging, 6(11), 1206-1211. 624. Sengupta, P. P., Huang, Y. M., Bansal, M., Ashrafi, A., Fisher, M., Shameer, K., ...& Dudley, J. T. (2016). A cognitive machine learning algorithm for cardiac imaging: a pilot study for differentiating constrictive pericarditis from restrictive cardiomyopathy. Circulation. Cardiovascular imaging, 9(6). 625. Serghiou, S., Patel, C. J., Tan, Y. Y., Koay, P., & Ioannidis, J. P. (2016). Field-wide meta-analyses of observational associations can map selective availability of risk factors and the impact of model specifications. Journal of clinical epidemiology, 71, 58-67. 626. Serhani, M. A., El Menshawy, M., & Benharref, A. (2016). SME2EM: Smart mobile end-to-end monitoring architecture for life-long diseases. Computers in biology and medicine, 68, 137-154.

238

627. Serhani, M. A., Benharref, A., & Nujum, A. R. (2014, August). Intelligent remote health monitoring using evident-based DSS for automated assistance. In Engineering in Medicine and Biology Society (EMBC), 2014 36th Annual International Conference of the IEEE (pp. 2674-2677). IEEE. 628. Shams, I., Ajorlou, S., & Yang, K. (2015). A predictive analytics approach to reducing 30-day avoidable readmissions among patients with heart failure, acute myocardial infarction, pneumonia, or COPD. Health care management science, 18(1), 19-34. 629. Sharma, G. B., Robertson, D. D., Laney, D. A., Gambello, M. J., & Terk, M. (2016). Machine learning based analytics of micro-MRI trabecular bone microarchitecture and texture in type 1 Gaucher disease. Journal of biomechanics, 49(9), 1961-1968. 630. Sharma, V., Stranieri, A., Burstein, F., Warren, J., Daly, S., Patterson, L., ...& Wolff, A. (2016). Group decision making in health care: A case study of multidisciplinary meetings. Journal of Decision Systems, 25(sup1), 476-485. 631. Shen, C., & Li, X. (2016). On the uncertainty of individual prediction because of sampling predictors. Statistics in medicine, 35(12), 2016-2030. 632. Shen, C. C., Chang, R. E., Hsu, C. J., & Chang, I. C. (2017). How business intelligence maturity enabling hospital agility. Telematics and Informatics, 34(1), 450-456. 633. Shi, X., Li, W., Song, J., Hossain, M. S., Rahman, S. M. M., & Alelaiwi, A. (2016). Towards interactive medical content delivery between simulated body sensor networks and practical data center. Journal of medical systems, 40(10), 214. 634. Shi, X., & Wang, S. (2015). Computational and data sciences for health-GIS. Annals of GIS, 21(2), 111-118. 635. Shneiderman, B., Plaisant, C., & Hesse, B. W. (2013). Improving healthcare with interactive visualization. Computer, 46(5), 58-66. 636. Siddiqui, S. A., Zhang, Y., Feng, Z., & Kos, A. (2016). A pulse rate estimation algorithm using PPG and smartphone camera. Journal of medical systems, 40(5), 126. 637. Siebert, J. C., Munsil, W., Rosenberg-Hasson, Y., Davis, M. M., & Maecker, H. T. (2012). The Stanford Data Miner: a novel approach for integrating and exploring heterogeneous immunological data. Journal of translational medicine, 10(1), 62. 638. Simpao, A. F., Ahumada, L. M., Desai, B. R., Bonafide, C. P., Gálvez, J. A., Rehman, M. A., ... & Shelov, E. D. (2014). Optimization of drug–drug interaction alert rules in a pediatric hospital's electronic health record system using a visual analytics dashboard. Journal of the American Medical Informatics Association, 22(2), 361-369. 639. Simpao, A. F., Ahumada, L. M., Gálvez, J. A., & Rehman, M. A. (2014). A review of analytics and clinical informatics in health care. Journal of medical systems, 38(4), 45. 640. Simpao, A. F., Ahumada, L. M., & Rehman, M. A. (2015). Big data and visual analytics in anaesthesia and health care. British journal of anaesthesia, 115(3), 350-356. 641. Singh, S. P., & Sawhney, T. G. (2006). Predictive analytics and the new world of retail healthcare. Health management technology, 27(1), 46. 642. Singla, D., Dhanda, S. K., Chauhan, J. S., Bhardwaj, A., Brahmachari, S. K., & Raghava, G. P. (2013). Open source software and web services for designing therapeutic molecules. Current topics in medicinal chemistry, 13(10), 1172-1191. 643. Singleton, C. A. (2014). MS in the analysis of biosimilars. Bioanalysis, 6(12), 1627- 1637. 644. Sinha, S., Song, J., Weinshilboum, R., Jongeneel, V., & Han, J. (2015). KnowEnG: a knowledge engine for genomics. Journal of the American Medical Informatics Association, 22(6), 1115-1119. 645. Sir, M. Y., Dundar, B., Steege, L. M. B., & Pasupathy, K. S. (2015). Nurse–patient assignment models considering patient acuity metrics and nurses’ perceived workload. Journal of biomedical informatics, 55, 237-248. 646. Skledar, S. J., Niccolai, C. S., Schilling, D., Costello, S., Mininni, N., Ervin, K., & Urban, A. (2013). Quality-improvement analytics for intravenous infusion pumps. American Journal of Health-System Pharmacy, 70(8), 680-686.

239

647. Skripcak, T., Belka, C., Bosch, W., Brink, C., Brunner, T., Budach, V., ...& Gulliford, S. (2014). Creating a data exchange strategy for radiotherapy research: towards federated databases and anonymised public datasets. Radiotherapy and Oncology, 113(3), 303-309. 648. Sloane, E. B., Rosow, E., Adam, J., & Shine, D. (2006, August). JEDI-an executive dashboard and decision support system for lean global military medical resource and logistics management. In Engineering in Medicine and Biology Society, 2006. EMBS'06. 28th Annual International Conference of the IEEE (pp. 5440-5443). IEEE. 649. Smith, D. H., Hicks, R. R., Johnson, V. E., Bergstrom, D. A., Cummings, D. M., Noble, L. J., ...& Tortella, F. C. (2015). Pre-clinical traumatic brain injury common data elements: toward a common language across laboratories. Journal of neurotrauma, 32(22), 1725- 1735. 650. Som, D., Tak, M., Setia, M., Patil, A., Sengupta, A., Chilakapati, C. M. K., ...& Badwe, R. (2016). A grid matrix-based Raman spectroscopic method to characterize different cell milieu in biopsied axillary sentinel lymph nodes of breast cancer patients. Lasers in medical science, 31(1), 95-111. 651. Song, T. M., & Ryu, S. (2015). Big data analysis framework for healthcare and social sectors in Korea. Healthcare informatics research, 21(1), 3-9. 652. Souliotis, K., Kani, C., Papageorgiou, M., Lionis, D., & Gourgoulianis, K. (2016). Using big data to assess prescribing patterns in Greece: the case of chronic obstructive pulmonary disease. PloS one, 11(5), e0154960. 653. Spaulding, T. J., & Raghu, T. S. (2013). Impact of CPOE usage on medication management process costs and quality outcomes. INQUIRY: The Journal of Health Care Organization, Provision, and Financing, 50(3), 229-247. 654. Spence, J., & Seargeant, D. (2015). service line analytics in the new era. Healthcare financial management: journal of the Healthcare Financial Management Association, 69(8), 44-47. 655. Spitzer, A. R., Ellsbury, D., & Clark, R. H. (2015). The Pediatrix BabySteps® Data Warehouse—A Unique National Resource for Improving Outcomes for Neonates. The Indian Journal of Pediatrics, 82(1), 71-79. 656. Splendiani, A., Gündel, M., Austyn, J. M., Cavalieri, D., Scognamiglio, C., & Brandizi, M. (2011). Knowledge sharing and collaboration in translational research, and the DC- THERA Directory. Briefings in bioinformatics, 12(6), 562-575. 657. Spruit, M., Vroon, R., & Batenburg, R. (2014). Towards healthcare business intelligence in long-term care: an explorative case study in the Netherlands. Computers in Human Behavior, 30, 698-707. 658. Rao, V. S. H., & Jonnalagedda, M. V. (2012). Insurance Dynamics–A Data Mining Approach for Customer Retention in Health Care Insurance Industry. Cybernetics and Information Technologies, 12(1), 49-60. 659. Srinivasan, U. (2014). Anomalies Detection in Healthcare Services. It Professional, 16(6), 12-15. 660. Srinivasan, U., & Arunasalam, B. (2013). Leveraging big data analytics to reduce healthcare costs. IT professional, 15(6), 21-28. 661. Venkatraman, S., Bala, H., Venkatesh, V., & Bates, J. (2008). Six strategies for electronic medical records systems. Communications of the ACM, 51(11), 140-144. 662. Stadler, J. G., Donlon, K., Siewert, J. D., Franken, T., & Lewis, N. E. (2016). Improving the efficiency and ease of healthcare analysis through use of data visualization dashboards. Big Data, 4(2), 129-135. 663. Stead, M., & Halford, J. J. (2016). A proposal for a standard format for neurophysiology data recording and exchange. Journal of clinical neurophysiology: official publication of the American Electroencephalographic Society, 33(5), 403. 664. Steinberg, G. B., Church, B. W., McCall, C. J., Scott, A. B., & Kalis, B. P. (2014). Novel predictive models for metabolic syndrome risk: a “big data” analytic approach. Am J Manag Care, 20(6), e221-e228.

240

665. Steinhubl, S. R., Marriott, M. P., & Wegerich, S. W. (2015). Remote sensing of vital signs: a wearable, wireless “Band-Aid” sensor with personalized analytics for improved Ebola patient care and worker safety. Global Health: Science and Practice, 3(3), 516-519. 666. Stewart, A. M., Gerlai, R., & Kalueff, A. V. (2015). Developing highER-throughput zebrafish screens for in-vivo CNS drug discovery. Frontiers in behavioral neuroscience, 9, 14. 667. Stojanovic, A., & Kessler, K. (2011). Case study: solutions for multidisciplinary decision making. Journal of Medical Marketing, 11(1), 60-70. 668. Stolper, C. D., Perer, A., & Gotz, D. (2014). Progressive visual analytics: User-driven visual exploration of in-progress analytics. IEEE Transactions on Visualization and Computer Graphics, 20(12), 1653-1662. 669. Suchard, M. A., Simpson, S. E., Zorych, I., Ryan, P., & Madigan, D. (2013). Massive parallelization of serial inference algorithms for a complex generalized linear model. ACM Transactions on Modeling and Computer Simulation (TOMACS), 23(1), 10. 670. Suciu, G., Suciu, V., Martian, A., Craciunescu, R., Vulpe, A., Marcu, I., ...& Fratu, O. (2015). Big data, internet of things and cloud convergence–an architecture for secure e- health applications. Journal of medical systems, 39(11), 141. 671. Suh, M. K., Moin, T., Woodbridge, J., Lan, M., Ghasemzadeh, H., Bui, A., ...& Sarrafzadeh, M. (2012, August). Dynamic self-adaptive remote health monitoring system for diabetics. In Engineering in Medicine and Biology Society (EMBC), 2012 Annual International Conference of the IEEE (pp. 2223-2226). IEEE. 672. Settele, J., Scholes, R., Betts, R. A., Bunn, S., Leadley, P., Nepstad, D., ...& Root, T. (2015). Terrestrial and inland water systems. In Climate Change 2014 Impacts, Adaptation and Vulnerability: Part A: Global and Sectoral Aspects. Cambridge University Press. 673. Sukumar, S. R., Natarajan, R., & Ferrell, R. K. (2015). Quality of Big Data in health care. International journal of health care quality assurance, 28(6), 621-634. 674. 675. Sun, L., Yamin, M., Mushi, C., Liu, K., Alsaigh, M., & Chen, F. (2014). Information analytics for healthcare service discovery. Journal of healthcare engineering, 5(4), 457-478. 676. Sung, J., Hale, V., Merkel, A. C., Kim, P. J., & Chia, N. (2016). Metabolic modeling with Big Data and the gut microbiome. Applied & translational genomics, 10, 10-15. 677. Suresh, S. (2016). Big data and predictive analytics: applications in the care of children. Pediatric Clinics, 63(2), 357-366. 678. Svechtarova, M. I., Buzzacchera, I., Toebes, B. J., Lauko, J., Anton, N., & Wilson, C. J. (2016). Sensor devices inspired by the five senses: a review. Electroanalysis, 28(6), 1201- 1241. 679. Szlezák, N., Evers, M., Wang, J., & Pérez, L. (2014). The role of big data and advanced analytics in drug discovery, development, and commercialization. Clinical Pharmacology & Therapeutics, 95(5), 492-495. 680. Taber, D. J., Palanisamy, A. P., Srinivas, T. R., Gebregziabher, M., Odeghe, J., Chavin, K. D., ...& Baliga, P. K. (2015). Inclusion of dynamic clinical data improves the predictive performance of a 30-day readmission risk model in kidney transplantation. Transplantation, 99(2), 324. 681. Taglang, G., & Jackson, D. B. (2016). Use of “big data” in drug discovery and clinical trials. Gynecologic oncology, 141(1), 17-23. 682. Takeuchi, H., & Kodama, N. (2014, August). Validity of association rules extracted by healthcare-data-mining. In Engineering in Medicine and Biology Society (EMBC), 2014 36th Annual International Conference of the IEEE (pp. 4960-4963). IEEE. 683. Tamblyn, R. et al. (2015)Evaluating the impact of an integrated computer-based decision support with person-centered analytics for the management of asthma in primary care: a randomized controlled trial. Journal of American Medical Information Association 22(4):773-83 684. Taylor, R. A., Pare, J. R., Venkatesh, A. K., Mowafi, H., Melnick, E. R., Fleischman, W., & Hall, M. K. (2016). Prediction of In‐hospital Mortality in Emergency Department

241

Patients With Sepsis: A Local Big Data–Driven, Machine Learning Approach. Academic emergency medicine, 23(3), 269-278. 685. Wah, T. Y., & Sim, O. S. (2009). Development of a data warehouse for lymphoma cancer diagnosis and treatment decision support. WSEAS Transactions on Information Science and Applications, 6(3), 530-543 686. Tehrani, N. (2016). Health Relationship Management Services (HRMS): A new healthcare paradigm using the 5Rs. Int J Biomed, 6(1), 87. 687. Thomas, A., & Peterson, L. E. (2014). Reduction of costs for anemia-management drugs associated with the use of ferric citrate. International journal of nephrology and renovascular disease, 7, 191. 688. Thomas, J. G., & Bond, D. S. (2015). Behavioral response to a just-in-time adaptive intervention (JITAI) to reduce sedentary behavior in obese adults: Implications for JITAI optimization. Health Psychology, 34(S), 1261. 689. Thompson, S., Varvel, S., Sasinowski, M., & Burke, J. P. (2016). From value assessment to value cocreation: informing clinical decision-making with medical claims data. Big data, 4(3), 141-147. 690. Toerper, M. F., Flanagan, E., Siddiqui, S., Appelbaum, J., Kasper, E. K., & Levin, S. (2015). Cardiac catheterization laboratory inpatient forecast tool: a prospective evaluation. Journal of the American Medical Informatics Association, 23(e1), e49-e57. 691. Toews, M., Wachinger, C., Estepar, R. S. J., & Wells, W. M. (2015, June). A feature- based approach to big data analysis of medical images. In International Conference on Information Processing in Medical Imaging (pp. 339-350). Springer, Cham. 692. Toga, A. W., Foster, I., Kesselman, C., Madduri, R., Chard, K., Deutsch, E. W., ...& Ames, J. (2015). Big biomedical data as the key resource for discovery science. Journal of the American Medical Informatics Association, 22(6), 1126-1131. 693. Torosyan, Y., Hu, Y., Hoffman, S., Luo, Q., Carleton, B., & Marinac-Dabic, D. (2016). An in silico framework for integrating epidemiologic and genetic evidence with health care applications: ventilation-related pneumothorax as a case illustration. Journal of the American Medical Informatics Association, 23(4), 711-720. 694. Torous, J., Staples, P., & Onnela, J. P. (2015). Realizing the potential of mobile mental health: new methods for new data in psychiatry. Current psychiatry reports, 17(8), 61. 695. Torres, E. B., Isenhower, R. W., Nguyen, J., Whyatt, C., Nurnberger, J. I., Jose, J. V., ... & Cole, J. (2016). Toward precision psychiatry: statistical platform for the personalized characterization of natural behaviors. Frontiers in neurology, 7, 8. 696. Torres, E. B., & Lande, B. (2015). Objective and personalized longitudinal assessment of a pregnant patient with post severe brain trauma. Frontiers in human neuroscience, 9, 128. 697. Tran, T., Nguyen, T. D., Phung, D., & Venkatesh, S. (2015). Learning vector representation of medical objects via EMR-driven nonnegative restricted Boltzmann machines (eNRBM). Journal of biomedical informatics, 54, 96-105. 698. Tremblay, M. C., Hevner, A. R., & Berndt, D. J. (2012). Design of an information volatility measure for health care decision making. Decision Support Systems, 52(2), 331- 341. 699. Trifiletti, D. M., & Showalter, T. N. (2015). Big data and comparative effectiveness research in radiation oncology: synergy and accelerated discovery. Frontiers in oncology, 5, 274. 700. Tsai, C. W., Chiang, M. C., Ksentini, A., & Chen, M. (2016). Metaheuristic algorithms for healthcare: open issues and challenges. Computers & Electrical Engineering, 53, 421- 434. 701. Tsang, W., Salgo, I. S., Medvedofsky, D., Takeuchi, M., Prater, D., Weinert, L., ...& Lang, R. M. (2016). Transthoracic 3D echocardiographic left heart chamber quantification using an automated adaptive analytics algorithm. JACC: Cardiovascular Imaging, 9(7), 769-782. 702. Tu, J. V., Chu, A., Donovan, L. R., Ko, D. T., Booth, G. L., Tu, K., ...& Kapral, M. K. (2015). The Cardiovascular Health in Ambulatory Care Research Team (CANHEART):

242

using big data to measure and improve cardiovascular health and healthcare services. Circulation: Cardiovascular Quality and Outcomes, 8(2), 204-212. 703. Turakhia, M. P., & Kaiser, D. W. (2016). Transforming the care of atrial fibrillation with mobile health. Journal of Interventional Cardiac Electrophysiology, 47(1), 45-50. 704. Uddin, S., Kelaher, M., & Srinivasan, U. (2016). A framework for administrative claim data to explore healthcare coordination and collaboration. Australian Health Review, 40(5), 500-510. 705. Valdes, G., Solberg, T. D., Heskel, M., Ungar, L., & Simone II, C. B. (2016). Using machine learning to predict radiation pneumonitis in patients with stage I non-small cell lung cancer treated with stereotactic body radiation therapy. Physics in Medicine & Biology, 61(16), 6105. 706. Van Poucke, S., Thomeer, M., Heath, J., & Vukicevic, M. (2016). Are randomized controlled trials the (g) old standard? From clinical intelligence to prescriptive analytics. Journal of medical Internet research, 18(7). 707. Van Poucke, S., Zhang, Z., Schmitz, M., Vukicevic, M., Vander Laenen, M., Celi, L. A., & De Deyne, C. (2016). Scalable predictive analysis in critically ill patients using a visual open data analysis platform. PloS one, 11(1), e0145791. 708. Varma, N., Piccini, J. P., Snell, J., Fischer, A., Dalal, N., & Mittal, S. (2015). The relationship between level of adherence to automatic wireless remote monitoring and survival in pacemaker and defibrillator patients. Journal of the American College of Cardiology, 65(24), 2601-2610. 709. Varma, N., Piccini, J. P., Snell, J., Fischer, A., Dalal, N., & Mittal, S. (2015). The relationship between level of adherence to automatic wireless remote monitoring and survival in pacemaker and defibrillator patients. Journal of the American College of Cardiology, 65(24), 2601-2610. 710. Vatsalan, D., & Christen, P. (2016). Privacy-preserving matching of similar patients. Journal of biomedical informatics, 59, 285-298. 711. Vawdrey, D. K. (2008). Assessing usage patterns of electronic clinical documentation templates. In AMIA annual symposium proceedings (Vol. 2008, p. 758). American Medical Informatics Association. 712. Vehlow, C., Kao, D. P., Bristow, M. R., Hunter, L. E., Weiskopf, D., & Görg, C. (2015). Visual analysis of biological data-knowledge networks. BMC bioinformatics, 16(1), 135. 713. Velianoff, G. D. (2014). Advancing the Evolution of Healthcare: Information Technology in a Person-Focused Population Health Model. Journal of Nursing Administration, 44(7/8), 381-387. 714. Velsko, S., & Bates, T. (2016). A conceptual architecture for national biosurveillance: moving beyond situational awareness to enable digital detection of emerging threats. Health security, 14(3), 189-201. 715. Verlingue, L., Alt, M., Kamal, M., Sablin, M. P., Zoubir, M., Bousetta, N., ...& Le Tourneau, C. (2014). Challenges for the implementation of high-throughput testing and liquid biopsies in personalized medicine cancer trials. Personalized medicine, 11(5), 545- 558. 716. Viangteeravat, T., Anyanwu, M. N., Nagisetty, V. R., & Kuscu, E. (2011). Automated generation of massive image knowledge collections using Microsoft Live Labs Pivot to promote neuroimaging and translational research. Journal of clinical bioinformatics, 1(1), 18. 717. Viangteeravat, T., Huang, E. Y., & Wade, G. (2013, October). Giving raw data a chance to talk: A demonstration of de-identified Pediatric Research Database (PRD) and exploratory analysis techniques for possible research cohort discovery and identifiable high risk factors for readmission. In BMC bioinformatics (Vol. 14, No. 17, p. A5). BioMed Central. 718. Vinodhini, M. A., & Vanitha, R. (2016). A Knowledge Discovery Based Big Data for Context aware Monitoring Model for Assisted Healthcare. International Journal of Applied Engineering Research, 11(5), 3241-3246.

243

719. Viceconti, M., Hunter, P. J., & Hose, R. D. (2015). Big data, big knowledge: big data for personalized healthcare. IEEE J. Biomedical and Health Informatics, 19(4), 1209-1215. 720. Voisin, S., Pinto, F., Morin‐Ducote, G., Hudson, K. B., & Tourassi, G. D. (2013). Predicting diagnostic error in radiology via eye‐tracking and image analytics: Preliminary investigation in mammography. Medical physics, 40(10). 721. von Eckardstein, A., Roth, H. J., Jones, G., Preston, S., Szekeres, T., Imdahl, R., ... & Feldmann, L. (2013). Cobas 8000 Modular analyzer series evaluated under routine-like conditions at 14 sites in Australia, Europe, and the United States. Journal of laboratory automation, 18(4), 306-327. 722. von Landesberger, T., Andrienko, G., Andrienko, N., Bremm, S., Kirschner, M., Wesarg, S., & Kuijper, A. (2013). Opening up the “black box” of medical image segmentation with statistical shape models. The Visual Computer, 29(9), 893-905. 723. Von Landesberger, T., Bremm, S., Kirschner, M., Wesarg, S., & Kuijper, A. (2013). Visual analytics for model-based medical image segmentation: Opportunities and challenges. Expert Systems with Applications, 40(12), 4934-4943. 724. von Landesberger, T., Basgier, D., & Becker, M. (2016). Comparative local quality assessment of 3D medical image segmentations with focus on statistical shape model-based algorithms. IEEE transactions on visualization and computer graphics, 22(12), 2537-2549. 725. Voss, E. A., Makadia, R., Matcho, A., Ma, Q., Knoll, C., Schuemie, M., ...& Ryan, P. B. (2015). Feasibility and utility of applications of the common data model to multiple, disparate observational health databases. Journal of the American Medical Informatics Association, 22(3), 553-564. 726. Vranic, B. Z., & Vandamme, T. F. (2015). Preliminary study of an offline simultaneous determination of metoprolol tartrate and hydrochlorothiazide in powders and tablets by reflectance near-infrared spectroscopy. Pharmaceutical development and technology, 20(1), 99-104. 727. Vukicevic, M., Radovanovic, S., Kovacevic, A., Stiglic, G., & Obradovic, Z. (2015, August). Improving hospital readmission prediction using domain knowledge based virtual examples. In International Conference on Knowledge Management in Organizations (pp. 695-706). Springer, Cham. 728. Vukićević, M., Radovanović, S., Milovanović, M., & Minović, M. (2014). Cloud based metalearning system for predictive modeling of biomedical data. The Scientific World Journal, 2014. 729. Wadsworth, T., Graves, B., Glass, S., Harrison, A. M., Donovan, C., & Proctor, A. (2009). Using business intelligence to improve performance: Cleveland Clinic tracks KPIs daily to measure progress toward achieving the organization's strategic objectives. This effort has helped reduce labor costs and other expenses--and improve quality of care. Healthcare Financial Management, 63(10), 68-73. 730. Wai, A. A. P., Duc, P. D., Syin, C., & Haihong, Z. (2014, August). iBEST: Intelligent balance assessment and stability training system using smartphone. In Engineering in Medicine and Biology Society (EMBC), 2014 36th Annual International Conference of the IEEE (pp. 3683-3686). IEEE. 731. Walcott, B. P., Nahed, B. V., Kahle, K. T., Redjal, N., & Coumans, J. V. (2011). Determination of geographic variance in stroke prevalence using Internet search engine analytics. Neurosurgical focus, 30(6), E19. 732. Walsh, B. K., Smallwood, C. D., Rettig, J. S., Thompson, J. E., Kacmarek, R. M., & Arnold, J. H. (2016). Categorization in mechanically ventilated pediatric subjects: a proposed method to improve quality. Respiratory care, respcare-04723. 733. Walsh, C., & Hripcsak, G. (2014). The effects of data sources, cohort selection, and outcome definition on a predictive model of risk of thirty-day hospital readmissions. Journal of biomedical informatics, 52, 418-426. 734. Wanderer, J. P., Gruss, C. L., & Ehrenfeld, J. M. (2015). Using Visual Analytics to Determine the Utilization of Preoperative Anesthesia Assessments. Applied clinical informatics, 6(04), 629-637.

244

735. Wang, F. (2015). Adaptive semi-supervised recursive tree partitioning: the ART towards large scale patient indexing in personalized healthcare. Journal of biomedical informatics, 55, 41-54. 736. Wang, K., Shao, Y., Shu, L., Zhu, C., & Zhang, Y. (2016). Mobile big data fault- tolerant processing for ehealth networks. IEEE Network, 30(1), 36-42. 737. Wang, L. W., Qu, A. P., Yuan, J. P., Chen, C., Sun, S. R., Hu, M. B., ... & Li, Y. (2013). Computer-based image studies on tumor nests mathematical features of breast cancer and their clinical prognostic value. PLoS One, 8(12), e82314. 738. Wang, M. (2013).Predefined three tier business intelligence architecture in healthcare enterprise. Journal of Medical Systems 37(2) 739. Wang, S., Pandis, I. &Johnson, D. 2014 Optimising parallel R correlation matrix calculations on gene expression data using MapReduce. BMC Bioinformatics 740. Wang, T. D. et al. 2011 Extracting insights from electronic health records: case studies, a visual analytics process model, and design recommendations. Journal of Medical Systems 35(5):1135-52. 741. Wang, W. &Krishnan, E. 2014 Big data and clinicians: a review on the state of the science. JMIR Med Inform 2(1) 742. Wang, Y. &Qian, X. 2016 Stochastic block coordinate Frank-Wolfe algorithm for large-scale biological network alignment. EURASIP J Bioinform Syst Biol 2016(1):9 743. Ward, M. J., Marsolo K. A. &Froehle C. M. 2014 Applications of Business Analytics in Healthcare. Business Horizons (2014) 57, 571—582 744. Warrington, L., Absolom, K. & G. Velikova, G. 2015 Integrated care pathways for cancer survivors - a role for patient-reported outcome measures and health informatics. Acta Oncol 54(5):600-8 745. Weir, M. H. et al. 2016 Effect of Surface Sampling and Recovery of Viruses and Non- Spore-Forming Bacteria on a Quantitative Microbial Risk Assessment Model for Fomites. Environ Sci Technol 50(11):5945-52 746. Wernick, M.N., Yang, Y. &Brankov, J. G. 2010 Machine Learning in Medical Imaging. IEEE Signal Processing Magazine 747. Wessler, B. S., Lai, L. Y. H. & Kramer, W. (2015) Clinical Prediction Models for Cardiovascular Disease Tufts Predictive Analytics and Comparative Effectiveness Clinical Prediction Model Database. Circulation -Cardiovascular Quality and outcomes 748. Westreich, S. T. et al. 2016 SAMSA: a comprehensive metatranscriptome analysis pipeline. BMC Bioinformatics 17(1):399 749. Whitworth, M. H. 2006 Designing the Response to an Anthrax Attack. Interfaces 36(6):562-568. 750. Wilbanks, B. A. &Langford, P. A. 2014 A review of dashboards for data analytics in nursing. Comput Inform Nurs 32(11):545-9 751. Wills, M. J. 2014 "Decisions through data analytics in healthcare." Student essay.

245

752. Wolf, M., Kim, H. & Van der Schaar, M. 2015 Caring Analytics for Adults With Special Needs. IEEE Design& Test 753. Wood, W. A., Bennett, A. V. &Basch, E.2015 Emerging uses of patient generated health data in clinical research. Mol Oncol 9(5):1018-24 754. Woodbridge, J. et al.2016 Improving Biomedical Signal Search Results in Big Data Case-Based Reasoning Environments. Pervasive Mob Comput 28:69-80 755. Wozney, L. et al. 2016 Usability, learnability and performance evaluation of Intelligent Research and Intervention Software: A delivery platform for eHealth interventions. Health Informatics J 22(3):730-43 756. Wu, F. M. 2016 Using health information technology to manage a patient population in accountable care organizations. Journal of Health Organization 757. Wu, J. et al. 2015 Building a medical research cloud in the EASI-CLOUDS project. Concurrency and Computation: Practice and Experience 27(16):4465-4477 758. Wu, J. et al. 2016 Automated Fovea Detection in Spectral Domain Optical Coherence Tomography Scans of Exudative Macular Disease. Int J Biomed Imaging 2016:7468953 759. Wu, W., Nagarajan, S.& Chen, Z. 2016 Bayesian Machine Learning: EEG/MEG Signal Processing Measurements (vol 33, pg 14, 2016). IEEE Signal Processing Magazine 760. Wyber, R. et al. 2015 Big data in global health: improving health in low- and middle- income countries. Bull World Health Organ 93(3):203-8 761. Wynn Jr. &Pratt, D.E.2014 The promises and challenges of innovating through big data and analytics in healthcare. Cutter IT Journal. 762. Xia, W. et al. 2015 R-U policy frontiers for health data de-identification. J Am Med Inform Assoc 22(5):1029-41 763. Xiang, W., Wang, G. &Pickering, M. 2016 Big Video Data for Light-Field-Based 3D Telemedicine. IEEE NETWORK 764. Xie, Y. Schreier, G. &Chang, D. C. W. 2015 Predicting Days in Hospital Using Health Insurance Claims. IEEE Journal of Biomedical Informatics 765. Xing, E.P., Curtis, R. E.&Schoenherr, G. 2014 GWAS in a Box: Statistical and Visual Analytics of Structured Associations via GenAMap. PLOS ONE 766. Xu, Z. et al. 2015 Knowle: A semantic link network based system for organizing large scale online news events. Future Generation Computer Systems 43-44:40-50. 767. Yang, C., Kerr, A. &Stankovic, V. 2016 Human Upper Limb Motion Analysis for Post- Stroke Impairment Assessment Using Video Analytics. IEEE ACCESS 768. Yang, M., Kiang, M.& Shang, W. 2015 Filtering big data from social media--Building an early warning system for adverse drug reactions. J Biomed Inform 54:230-40

246

769. Yang, S. et al. 2015 A Unified Framework and Platform for Designing of Cloud- Based Machine Health Monitoring and Manufacturing Systems. Journal of Manufacturing Science and Engineering 137(4) 770. Yang, S., Santillana, M. & Kou, S.C.2015 Accurate estimation of influenza epidemics using Google search data via ARGO. Proc Natl Acad Sci 112(47) 771. Yang, S., Njoku, M. &Mackenzie, C. F. 2014 'Big data' approaches to trauma outcome prediction and autonomous resuscitation. British Journal of Hospital Medicine 772. Yang, W., Lipsitch, M.&J. Shaman, J. 2015 Inference of seasonal and pandemic influenza transmission dynamics. Proc Natl Acad Sci U S A 112(9):2723-8 773. Yang, Y. Y. et al. 2014 Leveraging text analytics in patent analysis to empower business decisions – A competitive differentiation of kinase assay technology platforms by I2E text mining software. World Patent Information 39:24-34 774. Yao, Q. et al. 2015 Design and development of a medical big data processing system based on Hadoop. J Med Syst 39(3):23 775. Yasodha, P., & Ananthanarayanan, N. R. 2015 Analysing Big Data to Build Knowledge Based System for Early Detection of Ovarian Cancer. Indian Journal of Science and Technology 8(14). 776. Yildirim, P., Majnaric, L. & Ekmekci, O. I. 2014 Knowledge discovery of drug data on the example of adverse reaction prediction. BMC Bioinformatics 777. Yip, K. Pang, S. &Chan, K. 2016Improving outpatient phlebotomy service efficiency and patient experience using discrete-event simulation. International Journal of Health Care Quality Assurance 778. Yoo, C., Ramirez, L. & Liuzzi, J. 2014 Big data analysis using modern statistical and machine learning methods in medicine. Int Neurourol J 18(2):50-7 779. Yoo, S., Hwang,H.&Jheon, S. 2016 Hospital information systems: experience at the fully digitized Seoul National University Bundang Hospital. J Thorac Dis 8(Suppl 8):S637-41 780. You, Q., Shiaofen F., & Jake Y. C. 2008 Gene Terrain: Visual Exploration of Differential Gene Expression Profiles Organized in Native Biomolecular Interaction Networks. Information Visualization 9(1):1-12 781. Young, S. D., Rivers, C. & Lewis, B. 2014 Methods of using real-time social media technologies for detection and remote monitoring of HIV outcomes. Prev Med 63:112-5 782. Yu, K., Zhang, J. & Chen, M. 2014 Mining hidden knowledge for drug safety assessment: topic modeling of LiverTox as a case study. BMC Bioinformatics 783. Zarate, O. A. et al. 2016 Balancing Benefits and Risks of Immortal Data: Participants' Views of Open Consent in the Personal Genome Project. Hastings Cent Rep 46(1):36-45

247

784. Zarkogianni, K. , Litsa, E. &Mitsis, K. 2016 A Review of Emerging Technologies for the Management of Diabetes Mellitus. IEEE Transactions on Biomedical Engineering 785. Zeng, T. et al. 2014 Edge biomarkers for classification and prediction of phenotypes. Sci China Life Sci 57(11):1103-14 786. Zeng-Treitler, Q. et al. 2016 The effect of simulated narratives that leverage EMR data on shared decision-making: a pilot study. BMC Res Notes 9:359 787. Zenty, T.F., Bieber, E.J.&Hammack, E.R. 2014 University Hospitals: creating the infrastructure for quality and value through accountable care. Frontiers of health services management 788. Zhang, F. et al. 2015 A task-level adaptive MapReduce framework for real-time streaming data in healthcare applications. Future Generation Computer Systems 43-44:149- 160 789. Zhang, H. et al. 2015 Splitting Large Medical Data Sets based on Normal Distribution in Cloud Environment. IEEE Transactions on Cloud Computing:1-1 790. Zhang, J. &Zhang, B. 2014 Clinical research of traditional Chinese medicine in big data era. Front Med 8(3):321-7 791. Zhang, P.&V. Brusic, 2014 Mathematical modeling for novel cancer drug discovery and development. Expert Opin Drug Discov 9(10):1133-50 792. Zhang, X. Dou, W. &Pei, J. 2015 Proximity-Aware Local-Recoding Anonymization with MapReduce for Scalable Big Data Privacy Preservation in Cloud. IEEE Transactions on Computers 793. Zhang, X., Yang, L. T., Liu, C., & Chen, J. (2014). A scalable two-phase top-down specialization approach for data anonymization using mapreduce on cloud. IEEE Transactions on Parallel and Distributed Systems, 25(2), 363-373Zhang, Y. et al. 2016 Application and Exploration of Big Data Mining in Clinical Medicine. Chin Med J (Engl) 129(6):731-8. 794. Zhang, Y. et al. 2016 SenStore: A Scalable Cyberinfrastructure Platform for Implementation of Data-to-Decision Frameworks for Infrastructure Health Management. Journal of Computing in Civil Engineering 795. Zhang, Y., Qiu, M. &Tsai, C. 2016Health-CPS: Healthcare Cyber-Physical System Assisted by Cloud and Big Data. EEE Systems Journal 796. Zhang, Z., Gotz, D. &Perer, A. 2014 Iterative cohort analysis and exploration. Information Visualization 14(4):289-307 797. Zhao, L. P. &Bolouri, H. 2016Object-oriented regression for building predictive models with high dimensional omics data from translational studies. J Biomed Inform 60:431-45

248

798. Zhao, W., Zou, W. &Chen, J. 2015 Topic modeling for cluster analysis of large biological and medical datasets. 11th Annual Conference of the MidSouth-Computational- Biology-and-Bioinformatics-Society (MCBIOS) 799. Zhao, Y.&Castellanos, F. X. 2016 Annual Research Review: Discovery science strategies in studies of the pathophysiology of child and adolescent psychiatric disorders-- promises and limitations. J Child Psychol Psychiatry 57(3):421-39 800. Zhao, Y. et al. 2016 Connectome-scale group-wise consistent resting-state network analysis in autism spectrum disorder. Neuroimage Clin 12:23-33 801. Zhou, X. ,Niu, J. &Ji, W. 2016 Precision test for precision medicine: opportunities, challenges and perspectives regarding pre-eclampsia as an intervention window for future cardiovascular disease. American Journal of Translational Research 802. Zhu, F. et al. 2016 COMPASS: A computational model to predict changes in MMSE scores 24-months after initial assessment of Alzheimer's disease. Sci Rep 6:34567 803. Zhu, K. et al. 2015 Predicting 30-day Hospital Readmission with Publicly Available Administrative Database. A Conditional Logistic Regression Modeling Approach. Methods Inf Med 54(6):560-7 804. Zhu, N., Diethe, T. &Camplani, M. 2015 Bridging e-Health and the Internet of Things: The SPHERE Project. IEEE Intelligent Systems 805. Ziuzianski, P., Furmankiewicz, M. &Soltysik-Piorunkiewicz, A. 2014 E-health artificial intelligence system implementation: Case study of knowledge management dashboard of epidemiological data in Poland. International Journal of Biology and Biomedical Engineering.

List of mobile applications

1 Prognosis : Your Diagnosis 2 InSimu Patient 3 Diagnose 4 VisualDx 5 Differential Diagnosis 6 Common Differential Diagnosis 7 FREE Nursing Care Plans and Diagnosis 8 Nursing Diagnosis and Care Plans FREE 9 Clinical Medicine Differential Diagnosis 10 CURRENT Medical Diagnosis and Treatment 2020 11 Quick Medical Diagnosis & Treatment 12 Clinical Cases Diagnosis 13 Differential Diagnosis Mnemonics

249

14 Eye Diagnosis 15 Medical Differential Diagnosis 16 Clinical Sense - Improve Your Clinical Skills 17 MobileDDx - Pocket Differential Diagnosis Tool 18 Nursing Diagnosis List 19 Symptom to Diagnosis An Evidence Based Guide 20 Nursing Diagnosis 21 Manual of Nursing Diagnosis 22 USMLE Step1 & 2 CK Differential Diagnosis 23 Pediatric Disease and Treatment 24 Diagnosis of Oral Ulceration 25 Clinicals – History & Physical 26 Dermion : Differential diagnosis for dermatologist 27 ECG Cases 28 Human Dx 29 All Respiratory Disease and Treatment 30 CURRENT Diagnosis & Treatment Psychiatry 31 Nursing Diagnosis Ref Manual - Sparks and Taylor's 32 All Blood Disease and Treatment A-Z 33 ECG 100 Clinical Cases 34 PID Phenotypical Diagnosis 35 Diagnosaurus DDx 36 Clinical Dermatology (Colored & Illustrated Atlas) 37 Diseases Dictionary (FREE) 38 OphthDDx - Eye Diseases Differential Diagnosis 39 All Skin Diseases Atlas & Treatments 40 Nursing Diagnosis Flashcards 41 CURRENT Diagnosis & Treatment Neurology 42 Tcm Clinic Aid Trial 43 Differential Dx Free 44 Current Essentials of Medicine 45 Common Symptom Guide 46 Differential Diagnosis (DD) 47 X-Ray Differential Diagnosis 48 WikiMed - Offline Medical Wikipedia 49 MedEx - Clinical Examination 50 QuickEMrapid 51 ddxof 52 Nurse's Pocket Guide 53 Diseases Treatments Dictionary 54 Diseases and Disorders; Nursing Therapeutic Manual 55 iCU Notes - a free Critical Care Medicine resource 56 RN Pocket Guide 57 Medical Doctors & Nurses Differential Diagnosis 58 Signs & Symptoms

250

59 Medical Records 60 ECG Interpretation Made Easy 61 A to Z ECG Interpretation 62 A to Z ECG Interpretation 63 Ferri's Clinical Advisor 64 MSF Medical Guidelines 65 Harrison's Manual of Medicine 66 Visual Diagnosis in Emergency & Critical Care Medi 67 History Taking 68 semDDx 69 Nursing Diagnoses: Definitions and Classification 70 MSD Manual Professional 71 Oxford Medical Dictionary 72 Medical Terminology Dictionary:Search&Vocabulary 73 Nursing & Medical Quiz 74 PDM 75 Diabetes Diagnostics 76 WikiMed mini - Offline Medical Wikipedia 77 Fun Medical Quiz 78 Medical Lab Tests 79 Clinical Medicine 100 Cases 80 All Medical Mnemonics (Colored & Illustrative) 81 Ophthalmology & Optometry Guide 82 HandbooK of Laboratory and Diagnostic Tests 83 Medical Learning App for Doctors 84 Lab Test Reference Range (Free & Offline) 85 Nutrition Guide for Clinicians 86 Medical Cases Management 87 The Chief Complaint 88 Medical Quiz 89 Miniris 90 Clinical Examination Tips 91 Bates' Physical Examination 92 Radiology Assistant 93 Ferri's Clinical Advisor "5 books in 1" format App 94 TBI Prognosis Calculator 95 Short Cases in Medicine 96 Osmosis Med 97 UBC Radiology 98 Medical FlashNotes 99 MDCalc Medical Calculator 100 Calculate by QxMD 101 MEDizzy - Medical Community 102 Clinical Skills 103 Explain Medicine

251

104 Dr. Najeeb Lectures 105 Teach Me Anatomy 106 Mosby's Drug Reference for Health Professions 107 Anatomist - Anatomy Quiz Game 108 ACLS Simulator v2018 109 100+ & Short Cases in Clinical Medicine 110 Clinical Practice Guidelines (CPG) Malaysia 111 ABG Acid-Base Eval 112 ACC Guideline Clinical App 113 Clinical Treatment: Internal Medicine 114 Medical & Surgical Instruments 115 Anatomy 3D Atlas 116 Anesthesiologist 117 Child-Pugh score 118 Touch Surgery: Surgical Videos 119 Medicine Dash - Hospital Time Management Game 120 OSCE Reference Guide 121 BMJ OnExamination Exam Revision - Free Questions 122 JACC Journals 123 Medscape CME & Education 124 Internal Organs in 3D (Anatomy) 125 MedSchool 126 Resuscitation 127 Medical Calculators 128 ACLS Rhythm Tutor 129 Medical News Online 130 MediCalc® 131 MediCode: AHA ACLS, BLS & PALS 132 Glasgow Coma Scale (GCS): Consciousness Level 133 Physiology GURU 134 JoinTriage 135 AbcMedicalNotes 2020 136 Case Medical Research 137 Docquity- The Doctors' Network 138 Clinical Pediatrics 139 Clinical Skills and Examinations 140 Medical MCQs 141 Clinical Signs 142 doctorsgate - Secure Messenger 143 VirtualClinic Cases 144 ACP Clinical Guidelines 145 Medical Formulas 146 Doctor At Work (Plus) - Patient Medical Records 147 HealthTap for Doctors 148 Best Medical Pearls

252

149 MedLearn | Medical Education 150 Patient Medical Records & Appointments for Doctors 151 Clinical Lab Science Review 152 Docty — Medical & Laboratory Reference Values 153 MedsBla - Medical Messenger 154 MedLab Tutor 155 Airway Ex: Played by Anesthesiologists 156 The Lancet 157 Guideline Central - Clinical Practice Guidelines 158 Kdigo Mobile 159 GP Antibiotics 160 Complete Anatomy Platform 2020 161 Hidoc Dr. - Medical Learning App for Doctors 162 Curofy - Medical Cases, Chat, Appointment 163 OPD App - For Doctors 164 Doctor Assistant 165 Simpl - Simulated Patient Monitor 166 Dentist Manager: patient organiser software 167 List of my patients 168 MedShr: Discuss Clinical Cases

253