Consumer and Personal Health Records 17

Kevin Johnson , Holly Brugge Jimison , and Kenneth D. Mandl

After reading this chapter, you should know the • Defi ne patient portals, personal health records, answers to these questions: patient-controlled health records, and per- • What is the central tenet of consumer health sonal health applications. informatics? • What are the issues that biomedical informat- ics research can address relating to consumer 17.1 Introduction communication, decision-making, and infor- mation access needs? Complexity and collaboration characterize • Although technology is a major focus of the health care in the early twenty-fi rst century. discussion around consumer health informat- Complexity arises from a deeper or more ics, we realize that any technology of interest sophisticated understanding of health and dis- today will be obsolete in the next 10 years. eases, wherein etiological models must take However, the issues this technology is trying to into account both molecular processes and address will continue to be prevalent. With that physical environments. Collaboration refl ects in mind, what are the larger biomedical infor- not only inter-professional collaboration, but matics topics this technology is being used to also a realization that successful attainment of address? optimal well being and effective management of disease processes necessitate active engage- ment of clinicians, lay persons, concerned fam- ily members, and society as a whole. This K. Johnson , MD, MS (*) Department of Biomedical Informatics , chapter introduces technologies forming the Vanderbilt University School of Medicine, foundation of consumer health informatics , 2209 Garland Ave , Nashville 37232 , TN , USA including personal health records (PHR), e-mail: [email protected] patient portals , personal health applications , H. B. Jimison , PhD, FACMI and social networking . We will illustrate how Consortium on Technology for Proactive Care, each of these tools embraces the complexity of Colleges of Computer & Information Sciences and Health Sciences, Northeastern University , health care and makes possible the collabora- 360 Huntington Ave, WVH 476, tions necessary to engage consumers in health Boston 02115, MA , USA promotion, disease management, and preventive e-mail: [email protected] K. D. Mandl , MD, MPH Children’s Hospital Informatics Program , Harvard Medical School, Boston Children’s Hospital , The authors acknowledge and thank the co-authors of the 300 Longwood Avenue , Boston 02446 , MA , USA previous chapter edition, Patricia Flatley Brennan and e-mail: [email protected] Justin Starren.

E.H. Shortliffe, J.J. Cimino (eds.), Biomedical Informatics, 517 DOI 10.1007/978-1-4471-4474-8_17, © Springer-Verlag London 2014 518 K. Johnson et al. care. In the fi rst part of this chapter, we intro- Patient engagement is not purely driven by the duce four common health care challenges that patient. Seminal work by sociologists such as consumer health informatics addresses: con- Goffman ( 1959 ) and psychologists such as sumer engagement, information sharing, com- Skinner ( 1938 ) and Deci and Ryan ( 1985 ) munication and decision facilitation. We then observed that consumers may become engaged describe the history of consumer health infor- because of an innate desire to be rewarded by matics and the tools commonly available to participation or through external rewards. These address the above challenges. We conclude the observers of human behavior have reinforced the chapter by describing many of the emerging notion that in many individuals, innate desire trends in the discipline and some of the con- motivates engagement without the apparent pres- tinuing opportunities. ence of an extrinsic reward or punishment as a consequence of engaging in the activity. These observers also note, importantly, that not every 17.1.1 The Challenge of Improving person will have an innate desire to engage, or Consumer Engagement may lack the desire as the perception of the cost of engaging increases. For these individuals, Contemporary consumers are under greater extrinsic motivators (money, material goods, and pressure to participate in their own care than praise) may catalyze behavior change. It is also have patients at any time since the emergence the case for many of these individuals that peers, of modern health care. Patient participation family members, and even others in their social takes many forms, including shared decision- network may provide a surrogate for the patient’s making, self-care, and collaborative practices lack of engagement by actively participating in (Mead and Bower 2000 ). The role of the con- information gathering and health care decision sumer as a full partner in health promotion and making. All of these observations may be disease management has never been more nec- addressed by consumer health informatics essary than now. In the distributed managed- researchers. care models of health care, consumers serve as their own case managers, brokering care from generalists, specialists, and ancillary groups. 17.1.2 The Challenge of Improving Rarely do consumers receive all of their clinical the Information Available services from a single provider. Although to Consumers health information technology, such as the elec- tronic health record, provides an integrated data It has long been recognized that information repository and communication service, the sharing can lead to improved decision making view of the patient from a single electronic and health care system vigilance. Pamplets, new- health record often will not include a record of paper articles, and phone communication about services provided through the continuum of disease outbreaks were once the only methods to care. Moreover, for some patients, the inability accomplish information dissemination. These to access this record may represent a barrier to methods were suboptimal in their ability to optimal care delivery (Detmer 2003 ). Consumer address challenges with , numer- health informatics as a subdiscipline is founded acy, widespread access, appropriately timed on this need for access. A central tenet of con- access, or proper formatting/tailoring (for exam- sumer health informatics is that although not all ple, so that they could benefi t adolescents as well patients will participate, patients participating as adults, or men as well as women.) The advent in their health care leads to higher quality care of the Internet, interactive television and a host than that which is achievable without patient of widely available home technologies provide a participation. rich source of tools to address this challenge. 17 Consumer Health Informatics and Personal Health Records 519

17.1.3 The Challenge of Improving 17.2 Historical Perspective the Communication Among of Consumer Health Consumers and Providers Informatics

Consumer health informatics also strives to 17.2.1 The Rise of Consumer- improve consumer-provider interactions by both Oriented Health helping to prepare patients for meaningful dis- Communications cussions with their clinicians and by facilitating shared decision making. A full discussion of the In essence, what we think of as consumer engage- efforts in this area may be found in the chapter on ment refl ects a shift from the patient as the silent Telemedicine (Chap. 18 ). recipient of ministrations from a wise, benefi cent Consumer-to-consumer communication has clinician to an active collaborator whose values, been an important source of relevant information preferences, and lifestyle not only alter predispo- for patients with diseases of all sorts. Before the sition to certain illnesses but also shape the char- Internet became available, this type of communi- acteristics of desirable treatments. cation was diffi cult, though many health care Patients, family members and the general pub- institutes, churches, and communities developed lic have long-played an active role in health care support groups for common diseases or interests and have actively sought information from health (Lieberman 1988 ). professionals and government agencies. During the early twentieth century, the US Federal Children’s Bureau served as a major source of 17.1.4 The Challenge of Improving health information for the public. Mothers could Consumer Decision-Making write to this federal agency, asking questions about normal child development, nutrition, and One of the most important aspects of consumer disease management. Written materials, such as health informatics has to do with moving beyond letters and pamphlets served as the primary providing background health information toward mechanism for delivering information that sup- helping consumers make quality health decisions ported lay people in their handling of health chal- and appropriate actions. Before the increased lenges. Patient education companies such as availability of home computing and the Internet, Krames1 would partner with organizations like patients facing diffi cult decisions used informa- the American Heart Association to provide gen- tion available in public libraries or sent to them eral printed material on heart disease or with the (via stamped, self-addressed envelopes!) They American Cancer Society to provide information relied on conversations and advice from family, on cancer. As television became commonplace, close friends, church members, other patients, or broadcast and reproduction media including tele- doctors. Importantly, they made decisions often vision, videotapes and audiotapes, were quickly with no attention given to the evidence support- harnessed by health professionals as a means for ing each treatment alternative, without an objec- presenting complex information and health care tive assessment of risks and benefi ts, and without instructions to people in their homes. Consumer recognizing their own personal characteristics health broadcasts, and eventually dedicated that might impact the outcome of the chosen channels covering topics such as nutrition, fi tness, approach. and women’s health issues, became common and Each of these challenges to consumer engage- were broadcast on local daytime television and ment has rationalized work in the fi eld of con- radio networks. This mechanism continues today, sumer health informatics since its inception. We and often uses consumer panels, live phoned in will now briefl y review the history of this sub discipline. 1 www.krames.com (accessed 4/23/13). 520 K. Johnson et al.

these advocates for integrating technology into the examination room were intrigued by the aspect of computer use both by patients for their needs, as well as by patients to assist with health care professional needs.

17.2.2 Early Advances in Consumer Information Sharing and Decision Making

The legacy of the self-help movement of the 1970s and the consumerism of the 1980s was growth in the importance of the patient as a full participant in health care. At the time, the failure of bioscience research to produce defi nitive evidence-based guidelines for most common diseases required that medical science be combined with patient prefer- ences (e.g., between surgical and radiation thera- pies). Patients participated by self-monitoring, by Fig. 17.1 Early use of computing in consumer health evaluating and choosing therapeutic strategies informatics, here taking a medical history directly from a from a set of acceptable alternatives, by imple- patient (Slack, WV, et al. ( 1966 ). NEJM with permission) menting the therapies, and by evaluating their effects. In addition, social and clinical changes in questions, expert guests, audience participation, the manner in which care was provided shifted and other ways to provide general advice for activities that were once the purview of licensed consumers. professionals to patients and their family caregiv- Computers and telecommunications have ers. Furthermore, there was growing belief that been used since the early 1940’s for patient behavioral interventions and alternative therapies assessment, patient education, and clinical infor- held great promise as adjuncts, or even replace- mation sharing. In the 1950’s Collen and col- ments, for traditional medical therapies. leagues at Kaiser Permanente created a health In the 1980s clinicians and health educators appraisal system that prompted for patient data capitalized on the increasingly common personal and returned a systematic risk appraisal (Collen computers as vehicles for . et al. 1964 ). Then in the 1960’s Warner Slack and Initially, computers were used primarily for colleagues at the University of Wisconsin used a computer- assisted learning programs, providing mainframe computer system as a health assess- general coaching regarding topics such as nutri- ment tool. Patients sat at a cathode ray tube tion and home care of the elderly. As shown in (CRT) terminal and responded to text questions, Fig. 17.2 , The Body Awareness Resource Network receiving a printed summary of their health (BARN), developed in the 1980s by Gustafson appraisal at the end of the session (Slack et al. and colleagues at the University of Wisconsin, 1967 ). Figure 17.1 shows an early example of the engaged adolescents in game-like interactions to mainframe-based tool developed by Slack. At help them learn about growth and development, Massachusetts General Hospital in the late develop healthy attitudes towards avoidance of 1950’s, computer-driven telephone systems were risky behaviors, and rehearse strategies for nego- used to conduct home-based follow-up with post- tiating the complex interpersonal world of adoles- surgery cardiac patients, calling them daily to cence (Bosworth et al. 1983 ). Later developments obtain pulse readings. From the very beginning, in the 1980s and 1990s moved beyond computer 17 Consumer Health Informatics and Personal Health Records 521

Fig. 17.2 BARN topic index BARNY’S TOPIC INDEX and use by teens. Bosworth a ( 1983 ). This picture shows WHO TO teens interacting with a game CALL FOR BODY CARE STRESS on the an early graphical computer. The fi gure on the left is the topic index as HELP displayed on the screen (With 2 3 permission from Gustafson, 1 D, personal communication)

4 5 BERR DRUGS SMOKING 6 or 7

SEX EXIT

Choose a number 1–6 (or 7 to quit) and press return. b

aided learning to values clarifi cation and risk and expensive to build, they were relatively easy appraisal exercises, thus capitalizing on the com- to disseminate, and were associated with measur- putational power as well as the visual display able changes in knowledge (Lieberman 2001 ) capabilities of the computer. As the Internet and, in some cases, symptom management—such became more available in homes, Internet support as with nausea in pediatric cancer (Redd et al. groups gathered momentum. One such example, 1987 ) and anxiety management (Patel et al. 2006 ). Hopkins Teen Central developed by Ravert and colleagues ( 2001 ), allowed otherwise isolated children with cystic fi brosis to meet virtually and 17.2.3 Early Advances in Consumer- to discuss health and developmental issues that Consumer Communication impacted healthy decision making. This addi- tional computational power eventually found its Early electronic newsgroups, bulletin boards, way into computer games (Kato and Beale 2006 ). and a precursor to the World Wide Web known Although these games were both time-consuming as Gopher connected patients to information 522 K. Johnson et al. about various rare and chronic diseases (Grunfeld 17.2.4 Early Advances in Consumer and Ho 1997 ; Grunfeld et al. 1996 ; National Decision-Facilitation Research Council (U.S.). Committee on Enhancing the Internet for Health and Biomedical Given that most health care activities actually Applications: Technical Requirements and take place in the home, away from clinics and Implementation Strategies. 2000 ) One of the ear- hospitals (Ferguson 1996 ), it has always been liest such examples was patient handouts made important that individuals be active and informed available through the National Heart, Lung, and participants in their own care. A variety of Blood Institute (Grunfeld and Ho 1997 ). This approaches have been used to provide “just-in- medium of distribution was attractive to many time” information and decision assistance to con- health care institutions because of the nature of sumers. As noted above, pamphlets and videos the Internet, which allowed patients to retrieve were created to explain treatment options and information that was of interest without requir- medical procedures to patients, as background ing the author to actively disseminate this infor- for decision-making. This was followed by an era mation to a previously-identifi ed group of of interactive video systems to help patients people. Ferguson was heavily infl uential in the understand the risks and benefi ts associated with 1980’s and 90’s in creating and analyzing online treatment options, but also to help defi ne their social support groups for patients (Ferguson values for possible future health outcomes. The 1996 ). Later, Gustafson and colleagues in prime examples of this type of system originated Wisconsin demonstrated that online social sup- with the Foundation for Informed Medical port could reduce anxiety and hopelessness, with Decision Making.3 As early as 1973, Wennberg suggestive evidence for improving coping skills and others discovered that the rates of many and life expectancy (Ferguson 1997 ). Soon, the expensive surgeries and other treatments would idea of the Internet support group became an vary from location to location throughout the active area of development, continuing through U.S. (Wennberg and Gittelsohn 1973 ). This vari- today (Eysenbach et al. 2004 ). ation seemed to occur for medical conditions At the same time, this increasing use of the where there were multiple viable treatment Internet raised concerns about the quality of the options and choices depended more on physician material available to consumers (Kershaw 2003 ). and resource availability than need or patient This concern, in turn, fueled an industry of characteristics. Wennberg and Mulley focused on groups consisting of content editors, clinicians, developing interactive video consumer decision and technologists charged with creating legible aids that focused on these conditions (e.g., pros- and responsive consumer information that could tate cancer, breast cancer, back pain, etc.), dis- be subscribed to on the Web, as well as others covering that patient preferences and priorities focused on rating the quality of this information. for possible health outcomes could vary dramati- One notable group that promotes the deployment cally from person to person, and could be critical of useful online health information is the Health to defi ning an optimal decision. on the Net Foundation (HON),2 pioneered by an international panel of experts and supported by the numerous European groups. Although 17.3 Current Trends many of these pioneering companies have long since vanished, the access to consumer health 17.3.1 Consumer-Facing Software information has remained a core function of the Internet—with 56 % of American adults seeking There is an enormous amount of momentum in information about a personal health concern consumer activation. In part, this momentum is annually (Tu and Cohen 2008 ). fueled by lessons discovered by pioneers in the

2 www.hon.ch (accessed 4/23/13). 3 www.informedmedicaldecisions.org (accessed 4/23/13). 17 Consumer Health Informatics and Personal Health Records 523

Table 17.1 Various modes of consumer engagement with health care technology Mode of engagement Defi nition Examples Communication Support for patient-to-patient, computer-to- Patient portals patient and patient-to-provider knowledge or Patient-physician secure email information dissemination Online support groups Social networking sites Data storage A patient-centered and managed repository for Personal health records patient-entered data or “liquid” health-related Data portals for devices, health systems information and, pharmaceutical companies Behavior management Tools to support personal health goals, often by Weight management tools combining data storage, care protocols, Physical activity tools information dissemination, and communication Medication reminder systems Prepare people to participate in ‘close call’ Interactive tools for Breast Ca, Prostate decisions that involve weighing benefi ts, harms, Ca, Back Pain, End of Life, Heart and scientifi c uncertainty Disease

fi eld. In addition, this momentum is the conse- 17.3.2 Communication quence of a contemporary health care environ- ment that is more diffuse and involves more 17.3.2.1 Patient-Centered people than ever before. The home and the Communication community are fast becoming the most com- For decades, as the health care system has had to mon sites where health care is provided. wrestle with chronic illness and as an increasing Information technologies necessary to support knowledge base required practicing technically patients and their family caregivers must not sophisticated and high quality care, the process of only migrate from the inpatient institution to disease management has become more and more the community but also be populated with specialized. It is not uncommon for a patient to see information resources that help to guide patients up to seven different providers to manage the aver- in complex health care decision- making, to age chronic illness. Patients frequently express communicate with health professionals and to frustration about the inability for electronic health access the clinical records and health science records to support record- sharing. This frustration knowledge necessary to help them comprehend has led to recommendations by the Institute of their health states and participate in appropriate Medicine (IOM) that recognize the central role of treatments. patients as a repository of their own medical infor- With the ready availability of home computers mation in their “To Err is Human” book series and cellular phone platforms that support Web (Aspden and Institute of Medicine (U.S.). browsing and extensible software via “apps,” Committee on Data Standards for Patient Safety. there is a wealth of what is called “consumer- 2004 ; Institute of Medicine Committee on facing” software now available to consumers. Improving the Patient Record ( 2001 ); Kohn et al. Although the types of consumer-facing health 2000 ), culminating the explicit notion that quality technology will come and go over the years, health care should be “patient-centered” among these technologies can be categorized into four other attributes, and noting that six out of ten rules modes in which consumers engage with health to improve health care quality depend on patient care through the use of technology. Table 17.1 involvement in their care (Institute of Medicine summarizes these modes. Each mode of engage- (U.S.). Committee on Quality of Health Care in ment will be discussed below, along with some America. 2001 ). The role of patients in communi- examples illustrating the potential for this mode cating also was a central component of early of engagement. visions for the National Health Information 524 K. Johnson et al.

Examples of content for the three diemensions and their overlap

• Patient ID Health care Provider • Health Industry Personal Health Dimension Dimension • Health Insurance • Provider notes • Consent forms • Nonshared personal information • Clinical orders • Medication alerts • Self-care trackers • Practice guidelines • Audit logs • Decision-support • Personal library programs

• De-identified information • Mandatory reporting • Community directories • Public health services • Inspection reports • Vital statistics • Survey data • Public education • Population health risks materials • Communicable diseases • Neighborhood • Socioeconomic conditions environmental hazards • Registries

Population Health Dimension

• Infrastructure data • Planning and policy documents • Surveillance systems • Health disparities data

Fig. 17.3 Information dimensions. Information for hhs.gov/sp/nhii/documents/NHIIReport2001/ health: a strategy for building the National Health default.htm ) Information Infrastructure (2001) (Source: http://aspe.

Infrastructure . F i g u r e 17.3 depicts the informa- health record. For example, patient over-the- tion dimensions outlined in this work, which counter medications, home therapy for allergies included a personal health dimension that empha- or trauma, many health screening forms, and data sized data required for patients to communicate derived in physiologic monitoring in the home among or on behalf of their providers. may provide valuable information to caregivers. In addition to this lack of what has been Part of the STEEEP (“Safe, Timely, Effective, termed “data liquidity” (Courtney 2011 ) – repre- Effi cient, Equitable, Patient-Centered”) care senting the ideal state of fast, free-fl owing, and framework by the IOM (Institute of Medicine interoperable data - there are data important to (U.S.). Committee on Quality of Health Care in health that may not be in the typical electronic America. 2001 ), patient-centeredness implies that 17 Consumer Health Informatics and Personal Health Records 525 liquidity is facilitated by patients being able to negative fi ndings, coupled with the widespread access and direct other’s access to their information acceptance of ISGs among consumer organiza- according to clearer laws and protocols that pro- tions suggests that this mode of communication vide consistency and transparency. This ensures should be considered as a part of armamentarium that clinicians will have the right information to for biomedical informatics research and health make informed decisions at the point of care. information technology tools. Further research in this area includes the use of group videoconfer- 17.3.2.2 Electronic Support Groups encing and socialization interventions to more As mentioned above, concomitant with the era of closely simulate the face-to-face support groups home computers and later, the Internet, there was that have been shown to be effective. an immediate recognition that technology had removed geographic and physiologic obstacles to 17.3.2.3 Social Networks communication. Patients who had rare diseases, T h e p h e n o m e n o n o f social networking h a s or people in need of support to help make tough grown rapidly and is a new form of patient- patient decisions or to face lifestyle altering conditions communication. Such sites, epitomized by suddenly had a way to communicate with one Facebook ( w w w . f a c e b o o k . c o m ) , a r e o n l i n e v i r - another. One of the earliest electronic support tual communities where participants describe groups was developed by Johnson and Ravert, themselves with member-entered attributes, who recognized the ease with which adolescents establish or break connections to other members, adopted technology. Hopkins Teen Central was communicate, and share information. This simple an internet support group developed initially strategy results in an exponential number of com- using a device known as WebTV™, which pro- munity members sharing one or more attributes, vided Internet access via phone lines and house- allowing users to connect with people with simi- hold televisions (Johnson et al. 2001 ). This lar conditions. The for-profi t online health-related project proved extremely successful. It demon- social networking community Patients Like Me strated, for example, the power of patient-to- has demonstrated that individuals with a severe patient email, as well as the role of moderated chronic disease—amyotrophic lateral sclerosis— health discussions in answering questions com- are highly willing, even without compensation, to mon to patients with a shared medical need. contribute data and observations to a patient com- Electronic support groups, also known as munity (Frost and Massagli 2008 ) to accelerate Internet Support Groups (ISGs), have since learning about their disease. The site has no ties to become an almost ubiquitous way for patients the conventional health care system and short- comfortable with the Internet to communicate circuits the traditional research enterprise, reward- with each other. One study has suggested that ing participants, not just researchers, with over 28 % of Internet users have visited at least knowledge. The patient outcomes of diverse ther- one ISG (Horrigan et al. 2001 ). ISGs use four apies are collected using crowd sourcing, where modes of communication alone or in combina- patients contribute their information to a common tion: e-mail lists, instant messaging, bulletin database that can be queried to obtain summaries boards, and chat rooms. While the sustained of an aggregated experience of their peers. availability of these sites suggests their value to Social networking Web sites share most or all patients, a recent meta-analysis of ISGs for of the features of electronic support groups, and depression noted the poor quality of the evidence. even some data commonly provided through a Chat room use appeared to be associated with portal (through creating an affi liation with a lower levels of depression. Patients with breast group who externalizes public or private infor- cancer appeared to get the most value from ISGs, mation). Social networking platforms combined although the results all came from one site. These with personal health records provide a means for equivocal results will benefi t from more sophisti- social network members to share and aggregate cated multicenter trials; however, the lack of data obtained from the traditional health system, 526 K. Johnson et al.

Fig. 17.4 These computer users are part of a cognitive physical exercise, and socialization. These photos show health coaching study where they use their home comput- two study participants communicating with each other ers to communicate with their coach, receive feedback from their homes using Skype as part of the socialization from home monitoring data on cognitive exercise, sleep, protocol (Jimison et al. (2007 ) with permission) and to do so in a private manner (Eysenbach her lifetime of health data (Riva et al. 2001 ). The 2008 ; Weitzman et al. 2011 a). Guardian Angel “Manifesto” was posted when One of the features of health-related online the Web was only 2 years old at the then readily social networks is the rapid dissemination of infor- available 2-letter domain name www.ga.org . The mation across a network; however, there is great Manifesto presented a comprehensive vision of a variability in the quality of discourse on health- Web-based process that would be spawned at a related social networking sites. As shown in person’s birth and care for them through manage- Fig. 17.4 , conversations may be moderated (in this ment of health records, decision support, insur- case by a health coach). Conversations also may ance benefi ts, education and communication. be unmoderated and commercial infl uences may Over the past two decades, this vision has been enter the discourse without transparency. There realized through advances in personal health are also concerns around privacy. Compared with records. the restrictive institutional consents and compacts with patients that limit use of data and specimens 17.3.3.1 Portals under federal regulations applicable to much fed- As the gains acceptance, erally sponsored research, online social networks its relevance to individual patients also grows. are generally governed by no more than a terms of Many hospitals and clinics have begun providing use statement, often subject to change without direct patient access to the clinical record. These notice in 30 days. These privacy policies may be portals are defi ned as consumer-facing systems diffi cult to fi nd and not written in language acces- tethered to electronic health records, allowing sible by a population with a broad range of health patient views of clinical or claims data in a single literacy (Weitzman et al. 2011 b). Industry stan- institutional electronic health record system or dards governing safety and privacy of online payer system (Conn and Lubell 2006 ). Portals health-related social networking are yet to emerge. provide motivated individuals with a way to elec- tronically access sections of their records to recall salient instructions or obtain results of tests 17.3.3 Patient Access to Health (Markle Foundation 2004 ). Some of the fi rst such Information personal health records (PHRs) were Columbia’s PatCIS system (Cimino et al. 1998 ) and Beth Szolovits’ Guardian Angel Proposal represents Israel Deaconess’s PatientSite, developed in 1999 one of the fi rst, if not the fi rst, example of recog- (Weingart et al. 2006 ). Some of the most widely nizing the role of the patient as a curator of his or deployed are Epic’s MyChart (which is used 17 Consumer Health Informatics and Personal Health Records 527

Fig. 17.5 Example screen snapshot from PatCIS, a portal developed at Columbia University. Cimino ( 1998 ) (Source: Cimino, J, personal communication, 2011)

across large HMO’s with Epic electronic health the challenges and opportunities associated with record deployments, including Kaiser and Group deploying this technology, noting important Health Cooperative) and My HealtheVet (used by issues such as authentication, policies around over half a million US Veterans (Nazi et al. minors desiring patient portal access to a PHR, 2008 )). Many of these portals also include func- decisions about how much of the EHR should be tions besides viewing EHR information, such as available, and a host of other issues. This article, secure physician-patient messaging, appointment as well as two by Osborn et al. ( 2010 , 2011 ) artic- scheduling, and viewing and managing medical ulates the pervasive way that even simple shifts bills. Users of the Group Health Cooperative from a provider-centric to patient- centric view patient portal most frequently viewed test results, precipitate ripples in the processes and culture of requested medication refi lls, participated in health care. secure messaging with their provider(s), and viewed after-visit summaries (Ralston et al. 2007 ). 17.3.3.2 Personal Health Records The range of functions available in patient portal The PHR has become the foundation for devel- systems varies widely. oping tools to store data and to facilitate its Figure 17.5 shows the PatCIS system devel- reuse in ways patients fi nd engaging. According oped at Columbia University (Cimino et al. 2002 ) to the Markle Foundation, a PHR is “An elec- Critical aspects of security, privacy and accurate tronic application through which individuals identifi cation of patients must be addressed, as can access, manage and share their health infor- well as compliance with government regulations mation, and that of others for whom they are (see Chap. 5 ). Additionally, designers must take authorized, in a private, secure, and confi dential special care to insure that the language used in environment”(Markle Foundation 2004 ). presenting personal health data enhances the The idea of a personal repository for medical patient’s understanding of his or her health con- information is far from new. Families with infants cerns. Many of these portals also provide secure have used an immunization record book for messaging systems, allowing patients to commu- decades. The immunization blue book is a quint- nicate concerns about their clinical record to their essential, effi cient system with portable informa- health care providers. One article comparing four tion that supports entry by multiple providers and patient portals (Halamka et al. 2008 ) demonstrated storage by the patient. Clayton Christensen, who 528 K. Johnson et al. invented the concept of “disruptive innovation,” Medicine (National Institutes of Health) under summarizes the widely held promise of this tech- the Next Generation Internet Initiative nology in his book, the Innovator’s Prescription: (National Library of Medicine) and the Advanced A Disruptive Solution for Health Care Networks programs (National Library of (Christensen et al. 2009 ). “We cannot overstate Medicine). PING was later renamed to Indivo how important PHRs are to the effi cient (Mandl et al. 2007 ) in 2006. Figure 17.6 provides functioning of a low-cost, high quality health care an overview of a PCHR architecture. system…. We think that the Indivo system After diffusion of the PCHR model at [described below], or something like it is a good two Harvard Medical School invitational place to start.” PHRs are an inversion of the cur- conferences,4 Indivo became the reference model rent approach to medical records in that they are for subsequent PCHRs, including Microsoft’s created by, and reside with patients who grant per- HealthVault, which used Indivo software code, missions to institutions, clinicians, researchers, GoogleHealth (now defunct), which imple- public health agencies, and other users of medical mented the model on its own servers with its own information. By contrast, electronic health records code, and the Dossia consortium of large employ- store data in thousands of individual silos which ers, which contracted with the Indivo creators to later need to be united through some form of create a version for deployment to populations of health information exchange. PHRs are designed employees of the consortium founders, including to enable users to acquire copies of their data Wal-Mart, AT&T and Intel. from every site of care, a capability that will Despite the initial enthusiasm for the PCHR become commonly leveraged under the HHS’s model, uptake of this technology has been grad- Meaningful Use c e r t i fi cation requiring EHRs to ual, primarily because PCHRs work best when provide an export to patients. In some ways, this they can readily obtain a copy of data from the model advances information fl ow far more than EHR. Although many EHRs and other sources of models requiring inter-institutional data sharing patient data are willing to provide this access to agreements. Hence, data from two competing patients, there are numerous barriers to the vision health care networks may reside in the same PHR of the PCHR that have, to date, negated this will- without cumbersome agreements between those ingness to share data. First and foremost, EHR two networks. The patient asserts her claim to the vendors have been slow to allow data liquidity. data for each network independently. This Important data may reside in unstructured clinical consumer- driven model of data aggregation may notes, text blobs, or even scanned images—all promote data liquidity far more than competing formats that compromise liquidity. Data may be approaches, such as health information exchanges structured but may not be suffi ciently demarcated (Adler-Milstein et al. 2008 ), which require cen- to separate the entry into usable pieces. The evo- tralized management of data sharing agreements lution from a less to more structure prescribing between networks and institutions. standard in recent years is a good example of this barrier. Data within an EHR may not conform to 17.3.3.3 Personally-Controlled Health published standards for interoperability. For Records example, although the system may share labora- Closely replicating the patient-centered function- tory data, it may not code the data in a standard ality and convenience of the immunization blue such as LOINC . Therefore, a receiving system book is the personally controlled health record may not be able to merge these data with other (PCHR), a special instance of personal health laboratory data. Finally, data may conform to a records (Halamka, et al. 2008 ; Pagliari et al. content standard, but may use a proprietary form 2007 ). The fi rst instance, called the Personal of a messaging standard, making transit diffi cult. Internetworked Notary and Guardian or PING (Riva, et al. 2001 ) and was developed in 4 www.pchri2006.org and www.pchri2007.org (accessed 1998, with funding from the National Library of 4/23/13). 17 Consumer Health Informatics and Personal Health Records 529

Ecosystem of Disease Wellness Clinical Social applications management applications research networking making use of tools management tools an individual's or a tools population's PCHR data

Interface to applications

Information access controls (personally controlled) PCHR Platform Secure storage of lifelong persistent data across all sources

Interface for importing data (personally controlled)

Subscription agents for standards-based exchange Hospital or Pharmacy Primary care Payer claims Laboratory Data providers health practice or system network

F i g . 1 7 . 6 An architecture for a personally controlled than models requiring inter-institutional data-sharing health record system. The personally controlled health agreements. PCHR users acquire copies of their data across record (PCHR) has a secure repository of data with subse- sites of care. Importantly, patients can not only share data quent accesses controlled by the patient or her proxy. with family members and physicisins, but also with ecosys- PCHRs facilitate consumer-driven data aggregation in a tem of apps – modular, self-contained, software applica- manner analogous to a health care version of the fi nancial tions – that can interact with users and contribute to, Quicken product, thus advancing information fl ow far more analyze, and monitor their personal health data stores

The belief in the importance of an interoper- built around data stored in individual health able platform promoting data liquidity has been records. Analogous to the iPhone™ and validated through Project HealthDesign ( w w w . Android™ platforms, PCHRs expose those data, projecthealthdesign.org ). The PCHR, similar to under patient control, to third party apps across how Quicken and MINT.com function for fi nan- an open application programming interface cial data, is a tool enabling a patient to collect (Mandl and Kohane 2008 , 2009 ). Hence, PCHRs copies of her data across sites of care and over are designed to spawn apps that provide addi- time. This functionality is similar to that pro- tional functionality without stifl ing innovation. vided by PHRs. Where PCHRs diverge is the A platform promoting market competition notion of “control” of the data. First, PCHR data among apps is intended to address shortcomings can be given attributes based on the semantics of of current HIT in terms of user interface and util- the information content or based on the patient ity (Kim and Johnson 2004 ). The major PCHR preferences. The patient may, for example, hide platform providers offer a programming inter- specifi c encounter data, or may allow only a sub- face that third parties can use to provide “value- set of authorized users to see all data in the added” software-based applications and services PCHR. For example, the PCHR model enables such as interpretation of laboratory tests, refer- the patient to authorize access to individuals rals, and customized medical advice. Employers, including clinical providers, family members, governmental and nongovernmental organiza- health care proxies, and researchers information tions and health centers can create substitutable (views or even copies of the record) or to intel- applications that connect to PCHRs, upon patient ligent software agents (“apps”). Thus, beyond its approval, through the interface. Healthvault lists properties as a “data sponge”, the PCHR is envi- scores of such Web apps on its site. The SMART sioned as a consumer gateway to a rich ecosystem (Substitutable Medical Applications, Reusable 530 K. Johnson et al.

Technologies) Platforms project ( w w w . s m a r t - Table 17.2 Models of health behavior change platforms.org ), a part of the Strategic Health IT Name and source Summary Advanced Research Projects (SHARP) Program, Self-effi cacy An individual’s impression of extends the Indivo PCHR model and seeks to (Bandura 1977 ) one’s own knowledge and skill to develop a common application programming perform any task, based on prior success, physical ability, and interface across EHRs and PCHRS that supports outside sources of persuasion. the ready creation and use of substitutable appli- Predicts the amount of effort a cations across the health care system. person will expend to change behavior. It is a key component of other theories, such as the Theory of Planned Behavior. 17.3.4 Behavior Management Social cognitive Behavior change is determined theory (Bandura by personal, environmental and Tools to assist in the achievement of personal 1989 ) behavioral elements, which are health goals have had a long history of develop- interdependent. Theory of planned A link between attitudes and ment in the consumer sector. Many patients have behavior (Ajzen behavior. It asserts that behaviors believed in the concept of health prevention 1985 ) viewed positively and supported through wellness activities (active lifestyle devel- by others (subjective norm) are opment, stress reduction, weight control) long more likely to have higher levels of motivation and more likely to before health care professionals endorsed this be performed. mode of self-care. However, with the advent of Transtheoretical/ This model asserts that preventive medicine and data supporting the role stages of change behavioral change is a 5-step that wellness activities can play in maintaining model (Prochaska process, between which a person health, many of these tenets have become a part 2005 ) may oscillate before achieving complete change. of the armamentarium for disease management and are an area of discovery supported by the Agency for Health care Research and Quality interoperable and theoretically-based behavior (AHRQ) and other Federal institutes. change tools. This initiative demonstrated many A complete discussion of foundational models tools that could help consumers with behavior of behavior change is beyond the scope of this change, as will be discussed later. These demon- chapter, but key works are listed in Table 17.2 . strations leveraged the widespread adoption of Researchers and educators capitalize on these “smart” phone technology across geographic and theories in helping to reduce risky behaviors socio-economic divides. This widespread adop- (e.g., cigarette smoking, unprotected sexual tion, coupled with easy to use software devel- intercourse, and unhealthy eating) and to pro- opment environments, enables the development mote desirable health behaviors (i.e., referred to of personal health applications that operate as as behavior change.) stand- alone or integrated tools available to most Many of the early games targeted at behavior consumers. change (described above) attempted to remove the stigma (attitude) associated with engaging in healthy behaviors (such as taking medications to 17.3.5 Consumer Decision-Making combat a chronic illness.) Beginning in 2007, the Robert Wood Johnson Social trends, coupled with the introduction of Foundation, in the Project HealthDesign Initiative managed care and the rapid growth of computer (Brennan et al. 2007 ) catalyzed the development tools, networks, and multimedia, led to both an of personal health applications, with the belief explosion of need for health care information by that a properly developed common platform the lay public and a dramatic rise in the use of infor- would be essential to the spread of intelligent, mation technology to meet that need. Lay persons 17 Consumer Health Informatics and Personal Health Records 531 need information about health promotion, illness patient with a given condition and with treatment prevention, and disease management. At pres- alternatives, it is the patient who must provide ent, special computer programs, health- focused values and preferences associated with these pos- CD-ROMs, and health-related Internet- based Web sible outcomes. These interactive decision aids sites all provide information likely to be useful to provide descriptions and testimonials about the lay public in participating in health care man- future outcomes from the patient perspective agement and treatment decisions. These applica- (usually video or quotes of patients describing tions of medical informatics technologies focus on their experiences). Evidence from the literature the patient as the primary user. has shown that these types of decision aids We have seen decision tools develop for help- increase consumers’ knowledge (O’Connor et al. ing parents know when to take their child to the 2009 ) and satisfaction with their treatment deci- doctor or start treatment at home, decision tools sion (Whelan et al. 2004 ); in addition, when for helping someone decide to have a screening patients and physicians use these types of deci- test, and tools for deciding between surgery or sion tools, the patients are more likely to choose drug therapy. Simpler decision aids are offered the less invasive treatment. in paper form. These consumer health decision Currently, consumer health decision aids are aids sometimes take the form of decision trees, offered for the full spectrum from assistance for where answers to questions lead patients to a rec- routine health decisions in the home to intricate ommendation. Many of the newer tools provide sophisticated systems for serious conditions with dynamic assistance for health behavior change complex treatment options. Many new health interventions through mobile devices. Other sys- decision aids are still provided in paper-based tems, typically offered on the Web, are based on form, but most are interactive and easily deliv- complex decision models that support tailored ered in a Web format. Challenges still remain in risk information and utility assessment tools integrating these systems into clinical work fl ow, aimed at measuring individual patient preferences easily explaining and assessing patient prefer- for multiple possible health outcomes. O’Connor ences, and in tailoring the risk and probability and colleagues reviewed a variety of patient deci- information and recommended protocols to indi- sion aids (O’Connor et al. 2009 ), defi ning these as vidual patients. tools that “prepare people to participate in ‘close call’ decisions that involve weighing benefi ts, harms, and scientifi c uncertainty.” In their review 17.3.6 Consumer Information of the literature they found that consumer deci- Access sion aids improve knowledge, reduce decisional confl ict (both in relation to feeling uninformed Patients need quick, private access to accurate and in regard to personal values), reduce passiv- information that can calm fears and ensure direct ity, and reduce indecision. Tools such as Weight access to currently accepted treatment for their Watchers Online5 represent more reference- illness. Unfortunately, most existing health care based decision aids, while numerous tools, such delivery systems are not designed to satisfy these as Vandemheen’s decision aid to help patients “just-in-time” information requests. Instead of with cystic fi brosis consider lung transplantation providing this information just in time, it is pro- as an option, use more sophisticated approaches vided “just in case”, typically by a busy clinician (Vandemheen et al. 2010 ). as a part of a standardized monolog that covers These researchers noted that while the physi- frequently asked questions. Computer technol- cians and evidence from the literature provide ogy can supplement clinicians’ teaching with expertise about the diagnosis and prognosis for a more detailed information that can be referenced repeatedly by a patient in the privacy of his home, 5 www.weightwatchers.com (accessed 4/23/13). and at the time it is most relevant. 532 K. Johnson et al.

17.3.6.1 Passive Information Access populations that they treat. These resources Consumer health informatics resources pro- tend to complement and extend the clinical vide substantive and procedural knowledge services offered by the professional group and about health problems and promising inter- may be based on a desire to ensure adherence ventions. These information resources may be with accepted therapies or to triage and man- derived from carefully conducted, theoretically age access to care for common health problems. grounded, resource-intensive processes, or they Examples include Kaiser Permanente Health may be extracted from information published for Facts, a program designed to help Kaiser mem- other purposes to save time or money. For this bers answer questions about common health reason, information resources vary in content problems, and the Mayo Health Advisor, a com- and sophistication. Some of these resources are mercially available CDROM that any interested little more than digitized brochures, presenting person can purchase and use to help manage his in an electronic form exactly what can be found health at home. Examples of other commercially in available printed materials. Others include available programs that have a professional ori- interactive or multimedia presentations of infor- entation include Health Wise and Health Desk. mation about specifi c conditions and appropriate Figure 17.7 shows an example of professional- actions to prevent, cure, or ameliorate the prob- developed CHI resources accessible via the lem. Multimedia presentations capitalize on the Internet. features of computer systems to enrich the pre- Consumer health informatics resources devel- sentation of health-related materials with pic- oped from a self-help perspective complement tures, short movies, and drawings. and augment those provided by the formal health Consumer health informatics resources pro- care delivery system. A self-help perspective is vide patients with condition-specifi c and disease- generally more inclusive than a professional per- specifi c information about the problems they spective. The information may address daily liv- face. Some resources explain the etiology and ing concerns and lifestyle issues along with, or in natural history of disease in terms comprehensi- place of, content deemed credible by established ble to lay people. Other resources provide proce- medical authorities. dural information, explain diagnostic procedures Many CHI resources represent a combina- or services, detail expected treatment activities, tion of professional and self-help perspectives. and provide any relevant warnings and precau- Web- based resources, such as those provided by tions. The presentation of CHI is heavily infl u- the Fred Hutchinson Cancer Research Center, enced by the perspective of the system developer. provide pointers to other Web sites that rep- Medically oriented clinical resources demon- resent professional or self-help perspectives. strate an emphasis on locally accepted medical Commercial vendors, such as HealthGate Data practice. Community health-oriented resources Corporation, provide access via a Web site to are more likely to include information relevant to professional-developed and self-help–oriented living in the community with a specifi c disease or CHI resources for a subscription or transaction- condition. based fee. Consumer health informatics resources origi- nate from two major perspectives: professional 17.3.6.2 Active Information Access and self-help. Professional-developed con- In addition to the passive models that require sumer resources are those developed by health patients to search for information, there is an care clinicians and their organizations. Health increasing number of personal health applica- care organizations—such as HMOs, managed- tions that actively provide information based on care companies, and group practices—develop specifi c patient needs. The Project HealthDesign information resources as a service to the patient teams developed many demonstrations of active 17 Consumer Health Informatics and Personal Health Records 533

Fig. 17.7 Professionally developed consumer health information. Note the informal language style and easy naviga- tion (Photo provided courtesy of JDRF ( http://kids.jdrf.org/index.cfm )) access (Brennan, et al. 2007 ). For example, a team socialization, and intelligent medication at the University of Rochester developed a “con- r e m i n d i n g . F i g u r e 17.8 depicts the variety of sen- versational assistant” to provide patients with a sors available for home care (Pavel et al. 2010 ; daily checkup and information to mitigate exac- Jimison et al. 2007 ). With tailored interfaces to a erbations of their heart disease (Ferguson et al. common database, these new approaches enable 2010 ). Another project, developed by Siek and coaches and informal caregivers to help man- Ross at the University of Colorado, allowed older age large panels of patients remotely and enable adults to manage complex medication regimens at patients and family members to participate as home, using active information access to denote active and informed members of the care team. possible side effects, duplicate therapies and Researchers also have begun leveraging text special instructions related to dosing (Siek et al. messaging as a medium to provide information 2011 ). Jimison and researchers at the Oregon actively to patients. For example, in the project Center for Aging & Technology additionally use MyMediHealth, Johnson and colleagues provide monitoring data from sensors in the home to pro- medication reminders to patients in real time duce automated feedback and intelligent alerts for using text messaging to their phones. When a coaching interventions in the areas of cognitive dose is missed, MyMediHealth is able to escalate exercise, physical exercise, sleep management, its medication reminder to a parent or other adult 534 K. Johnson et al.

Fig. 17.8 A variety of sensors in the home can provide medication dispensers, etc.). This fi gure shows a sample feedback for health behavior change coaching systems apartment layout with sensors for monitoring older adults (motion sensors, pressure mats, contact switches, wireless in the home (Jimison et al. (2007 ) with permission)

(Slagle et al. 2010 ). Studies also have demon- Act of 2009 (ARRA).”, 2009) may help increase strated the utility of appointment or preventive the momentum behind this model. The HITECH care reminders in reducing clinic no-show rates Act provides that, for covered entities using or (Brannan et al. 2011 ; Leong et al. 2006 ) maintaining an EHR, “the individual shall have a right to obtain from such covered entity a copy of such information in an electronic format.” By 17.4 Opportunities and 2013, this feature of EHRs will be required under Challenges the fi nal rule for meaningful use of certifi ed elec- tronic health record technology. In preparation for 17.4.1 Health Information this, early stage efforts have arisen to promote data Technology for Economic liquidity through the very well-marketed Blue and Clinical Health Act Button initiative from the Department Veterans of 2009 Affairs (Chopra et al. 2010 ), and the Direct Project, a federally initiated, health specifi c imple- T h e Health Information Technology for mentation of the SMTP protocol t o e n a b l e p o i n t - Economic and Clinical Health Act of 2009 to-point communication of health information in a (HITECH) (“Health Information Technology for secure, standards-based ( w w w . d i r e c t p r o j e c t . o r g ). Economic and Clinical Health (HITECH) Act, Unfortunately these approaches to data liquidity Title XIII of Division A and Title IV of Division came too late for the Google Health project, which B of the American Recovery and Reinvestment was closed down after not gaining a foothold in 17 Consumer Health Informatics and Personal Health Records 535 the health space. While patients and institutions resources. Furthermore, the quality of such infor- may have had issues viewing Google as a trusted mation varies widely not only in terms of the intermediary to store and manage health data, an extent to which it is accepted by the formal medi- at least equally major issue for the service was cal community but also in its basic clinical or sci- that even the early adopters, the technophiles who entifi c accuracy. Therefore, a key issue in CHI lies wanted to make this work, experienced challenges in determining the quality and relevance of health populating their PHRs with data. information found on Web sites. Credentialing or Hence we can expect that even before 2013 a certifi cation by recognized bodies, such as “tectonic shift in the health information econ- respected health care providers or clinical profes- omy” (Mandl and Kohane 2008 ) will begin, sional associations, represents one approach to mediated by a change in the locus of control of ensuring the quality of health information avail- health information from institutions to individu- able to consumers. This approach differs from the als. While this shift will be largely driven by a HON approach (see Sect. 17.2.3 ) i n t h a t i t b a s e s need to improve clinical care processes, it will its quality ratings on reviews and evaluations con- also have a deep impact on population health ducted by established knowledgeable sources. It research. The ability to reach out to populations has the advantage of delivering an imprimatur to a directly, thereby disintermediating traditional Web site, which informs the user that the informa- institutions, will produce very large cohorts of tion presented meets a standard of quality. individuals who can share EHR data and provide Credentialing is most useful when the credential detailed self-reported information about their itself is accompanied by a statement indicating care and health states. And patients are willing to the perspectives and biases of those granting it. share data for aggregate knowledge, research, Information presented by alternative therapies and public health (Weitzman et al. 2010 ). A and other non-clinical groups is no less suscepti- major area of focus for research and policy needs ble to bias than is information presented by pro- to be development of a properly aligned compact fessional sources. with patients to share data for research and public Inherent in the credentialing approach are three health by engaging them on their own terms; the disadvantages. First, the challenge to ensure that PCHR is a technology designed to do just that. every information element—every link in a deci- As CHI and evolve from being research sion program or pathway in a Web site—is tested novelties, to being the way health care is delivered, exceeds the resources available to do so. In many many challenges must be overcome. Some of these cases, the credentialing approach rests on certifi - challenges arise because the one patient, one doctor cation of the group or individuals providing the model no longer applies. Basic questions of identity information rather than approval of the content and trust become paramount. At the same time, the itself. Second, the credentialing approach leaves shifting focus from treating illness to managing control of the authority for health care informa- health and wellness requires that clinicians know not tion in the hands of traditional care providers, only the history of the individuals they treat but also refl ecting both the expertise and the biases of information about the social and environmental con- established medical sources. Third, credentialing text within which those individuals reside. alone is inherently contradictory to healthcare consumerism, which empowers the consumer to make choices consistent with her own worldview. 17.4.2 Information Credentialing A source’s credential is just an additional piece of information that may be considered in making The amount of information available to consumers personal health decisions. is growing rapidly. This volume of information An approach to evaluating the quality and can be overwhelming. Therefore, consumers may rele vance of CHI resources that is consistent with need help in sifting through the mass of available a philosophy of patient participation is based on 536 K. Johnson et al. teaching patients and lay people how to evaluate posted for search engines – the most common CHI resources. Consumers with suffi cient liter- tool that consumers use for fi nding health infor- acy and evaluation skills can locate CHI resources mation. Companies associated with the search and determine these resources’ relevance to their engines have the capability to infer a good deal individual health concerns. about the health status of the user or the user’s family from the search terms used and from the sites visited. However, users very rarely investi- 17.4.3 Privacy and Security gate the privacy policies and also rarely under- stand the implications of how data about them Because of the public, shared nature of the might be used. Internet, its resources are widely accessible by More complicated privacy situations arise citizens and health care organizations. This pub- with the new technologies for home monitoring. lic nature also presents challenges to the security Data is collected in the home and then typically of data transmitted along the Internet. The open- encrypted and transferred to a secure server in ness of the Internet leaves the transmitted data another location. A wide variety of data may be vulnerable to interception and inappropriate collected (blood pressures, blood glucose, activ- access. In spite of signifi cant improvements in ity, location, sleep quality, etc.) and then poten- the security of Web browsing several areas, tially shared with a variety of care team members including protection against viruses, authentica- (care manager, clinical, specialist, family care- tion of individuals and the security of email, giver, etc.). Not all care team members should remain problematic. necessarily have the same access to data. Role- One of the most important challenges and based access representations are beginning to be responsibilities associated with the new develop- used to make it easier and more meaningful for ments in interactive health systems for patients is patients to indicate their preferences on who (if ensuring the privacy and security of health infor- anyone) should see what data about them. mation in a way that protects the security of the There has recently been a dramatic upsurge in data and truly incorporates patient preferences the use of mobile applications for consumer for data sharing. Consumer health Web sites health monitoring and data sharing. However, the often offer users the capability of entering per- privacy and security techniques for this area are sonal data in order to obtain tailored feedback. still lagging behind the development and use of This ranges from calculators of body mass index the systems. In 2010 there was a 250 % jump in to ongoing weight management or smoking ces- mobile phone viruses and spyware, and 61 % of sation programs. These Web sites typically dis- the reported smartphone infections were spyware play their privacy practices. Consumers also can that could monitor communication from the look for certifi cates of accreditation from organi- device (Bela and Hamel 2010 ). Health care sys- zations such as URAC, 6 TRUSTe 7 and HON. 8 tems are working rapidly to address the privacy These accreditation groups each have a set of and security issues as technology and consumers’ policies to protect consumers’ privacy and certify way of using it change. that processes that encourage high quality accu- rate information is presented. Given that the pri- vacy policies for consumer health Web sites are 17.4.4 Technology Digital Divide often quite long and complicated; this approach allows consumers to feel confi dent that the site Ensuring every citizen access to the Internet rep- follows best practices. Privacy policies are also resents a second important challenge to the abil- ity to use it for public health and consumer health 6 www.urac.org (accessed 4/23/13). purposes. Access to the Internet presently 7 www.truste.com (accessed 4/23/13). requires computer equipment that may be out of 8 www.hon.ch (accessed 4/23/13). reach for persons with marginal income levels. 17 Consumer Health Informatics and Personal Health Records 537

The digital divide with respect to access to process of assessing patient preferences (utilities computer- based health care information occurs on outcomes) and communicating this to the cli- for many reasons, including economic disparity, nician. These issues are all critical for improving unequal broadband access in various geographi- health care decision-making. cal locations, and cultural issues. Majority- language literacy and the physical capability to type and read present additional requirements for 17.5 Future Directions effective use of the Internet. Preventing unequal access to health care resources delivered via the The advances in consumer health informatics Internet will require that health care agencies over the past 20 years have been made under work with other social service and educational assumptions about the availability of home com- groups to make available the technology neces- puting power, the general willingness of patients sary to capitalize on this electronic environment to communicate, and the ways in which data for health care. informs knowledge. The future of the fi eld is most likely to be shaped by disruptions in each of these assumptions. 17.4.5 Workfl ow Integration

Systems that are designed to facilitate shared 17.5.1 Home Computing Power decision making with a physician are diffi cult to integrate into routine clinical workfl ow. This In 1988, inspired by discussions with social sci- integration requires overcoming two main chal- entists, philosophers, anthropologists, and com- lenges. First, many care providers are reluctant to puter scientists, Mark Weiser, then at Xerox review patient-generated information until it is a PARC asserted that a “third wave” of computing reimbursable activity. Therefore, the preferred would transform society. The fi rst wave, main- workfl ow is to have these data brought in by the frame computers, allowed many people to access patient for review at the time of a patient encoun- one computer. This wave was complemented by ter. This approach may be helpful for many medi- the personal computer wave that supported one cal conditions, but is potentially less effi cient for person on one computer. The third wave, “ubiq- the patient who may be able to improve health uitous computing” further complemented the fi rst outcomes sooner by more timely professional and second waves by allowing one person to have review of data. many computers. This era will usher in “calm The second challenge relates to the health technology,” where technology becomes embed- information technology used by clinicians. ded into all aspects of our life and allows us to Ideally, data that are created by patients should be focus on the task, not the tool. Examples of this easily imported into the physician’s repository technology have been incorporated into the plots and provided as a view alongside clinician- of science fi ction movies for years. generated data. However this model has yet to be The average person now uses dozens of standardized. This model would be a requirement computer- enabled devices, ranging from smart- to achieve clinician buy-in with timely data phones and cars to coffee-makers, digital ther- review, and is very important even for the visit- mometers, HDTVs, and thermostats. The based data review if decisions are made based on availability of computers, coupled with the power these data. of the technology, will be leveraged by consumer Additionally, it has been diffi cult to reliably health informatics researchers. Already, pro- and meaningfully explain outcomes and proba- grams such as Robert Wood Johnson’s Project bilities to patients in a way that helps them weigh HealthDesign9 have demonstrated how home risks and benefi ts of complex scenarios. Finally, much more work needs to be done to improve the 9 www.projecthealthdesign.org (accessed 4/23/13). 538 K. Johnson et al. biosensors, mobile computing platforms, and privacy, it has become clear that we value com- other innovations can transform personal health munication and hope that there is privacy through applications. anonymity. As computer power increases and the integra- One example of patient willingness to com- tion of computer technology into everyday life municate about health is the success of tools like becomes more commonplace, we anticipate that researchmatch.org10 developed by Harris and col- health information will follow suit. Why doesn’t leagues. This tool allows patients to voluntarily a pill container send data directly to a personal create a profi le that is then used to alert them to health record when opened? Why doesn’t that studies in which they might be able to participate. pill container use wifi , speech recognition, and Patients using this system trust that the security publically available and appropriately tagged in place is suffi cient and that any risk of a secu- data to answer consumer questions about side rity breach is offset by the potential personal or effects or special instructions? Why can’t a need societal gains associated with patient research. to refi ll that medication be transmitted to a sched- We can expect various tools to leverage ule automatically, or included as a reminder when patient’s willingness to communicate. For exam- a car starts? It is likely that Federal groups such ple, social networks might be used to connect a as the US Army’s Telemedicine and Advanced person with a home health need to a person who Technology Resource Center (TATRC) or the is willing to help. Online forums or classifi ed National Science Foundation will lead the way advertisements that already connect patients to toward this level of interoperability and seamless patients for physician referrals or durable equip- information dissemination to consumers. As data ment sales could do more to help patients fi nd liquidity is increased, infl uential consumer health information or other commodities. technologies will fl ourish. In the mean time, efforts such as Google Health™ stumble waiting for data from the health system to enter the hands 17.5.3 Data Informing Knowledge of consumers. In fact, it is clear that Google Health and personal data repositories may follow One of the most exciting, though potentially the path of other notable technologies, such as alarming consequences of our extensive use of personal digital assistants or the Apple Newton the Web for shopping, communicating, and learn- MessagePad™, which emerged before support- ing is that each of us leaves behind a profi le of ing technologies or market awareness was there who we are, what we like or dislike, what we to sustain them. In the case of personal health know or don’t know, and what we want or already data repositories, the key driver may be these per- have. When combined with data mining and nat- sonal health applications integrated into daily ural language processing techniques, it is possi- life. ble to create highly targeted and predictive personal knowledge. It is this ability that search engines exploit to create a profi le of each searcher 17.5.2 Patient Communication and to improve the relevance of retrieved results. This technology also is likely to improve the It is estimated that 83 % of U.S. adults have a cell ways in which consumers and consumer-facing phone of some kind, and that 42 % of them own technologies operate. We can expect the use of a smartphone ( Smith 2011 ). In addition to cellu- these massive data sets (also called “big data”) to lar technology, social networking adoption has impact how medical care is personalized. Data exploded in the past decade. The widespread use created by consumers, coupled with ubiquitous of these technologies affi rms the belief that computing, might provide just–in-time nutri- patients are more willing than ever to communi- tional consults, over the counter medication cate about their health issues. In study after study about the tradeoff between communication and 10 www.researchmatch.org (accessed 4/23/13). 17 Consumer Health Informatics and Personal Health Records 539 advice, or advice that might prevent illnesses, Medicine, 358 (16), 1732–1737. This is a recent review of such as convenient locations to receive a fl u vac- the transformation taking place in health care as a result of cine or when to begin medications for seasonal widespread introduction of personally controlled health allergies. records. While the direction that consumer health informatics will take in the future is at best, edu- cated speculation, it is clear that as long as Questions for Discussion patient-provider partnerships are endorsed, tech- 1. What is the role of the health system in nology will be a third partner in ensuring that monitoring the quality of discourse on activated consumers manage their health and dis- online social networks? ease effectively. 2. What is the optimal model for personal health records? Should personal health Suggested Readings records display advertisements? 3. Which populations of consumers would The 1994 Guardian Angel Manifesto. www.ga.org . This be most likely to use personal health site describes a collaborative project begun in 1994 to records? “put power and responsibility for health care more into 4. Which consumer technologies do you the hands of patients.” It also provides a comprehen- think will be most infl uential in con- sive set of references to many other projects. sumer-focused health informatics? Susanna Fox’s blogs and reports. http://www.pewinternet. 5. What is the right balance between pri- org/Experts/Susannah-Fox.aspx . This site provides a vacy of personal health information and collection of presentations from form Pew Internet and ready access to it? For example for an American Life Project, which studies the technology- related cultural shifts in health care. unconscious patient in the emergency department? Mandl, K. D., & Kohane, I. S. (2008). The architecture of Personally Controlled Health Records. Tectonic shifts in the health information economy. New England Journal of