Pan American Journal Review of Public Health

Open data and public health

Marcelo D’Agostino,1 Noah O. Samuel,2 Maria Janina Sarol,2 Federico G. de Cosio,1 Myrna Marti,1 Tianyu Luo,3 Ian Brooks,2 and Marcos Espinal1

Suggested citation D’Agostino M, Samuel NO, Sarol MJ, de Cosio FG, Marti M, Luo T, et al. and public health. Rev Panam Salud Publica. 2018;42:e66. https://doi.org/10.26633/RPSP.2018.66

ABSTRACT This article provides an overview of the intersection of open data and public health by first defining data, public health data, and other key concepts and relevant termi- nologies. There are differing perceptions on the urgency and importance of the of public health data. It has been established that disease outbreaks such as happened during the Ebola and Zika virus epidemics are indicative of the need for countries to develop a framework that will provide guidance for the management of public health data. Such a framework should ensure that data collected during public health emergencies are accessible to the appropriate authorities and in a form that can help with timely decision-making during such public health crises. In this article, we highlight available open data policies across many countries, includ- ing in the Americas. Our analysis shows that there are currently no articulated policy guide- lines for the collection and management of public health data across many countries, especially in Latin America. We propose that any national data governance strategy must address poten- tial benefits, possible risks, examples of data that could be shared, and the attributes of such data. Finally, we stress that the key concern in the Americas should be the development of regional frameworks for open data in public health that can be adopted or adapted by each coun- try through appropriate national or subnational policies and strategies.

Keywords Information systems; data sources; public health; health systems; confidentiality.

In this article, we provide some basics public for free, and in formats that al- to base decisions point to the need for about open data and open health data in low a variety of uses. open public health data. particular, and we discuss how these “Open data” is a general umbrella Documented and openly shared pub- concepts apply to the public health sec- term that is used to describe all forms of lic health data can help to prevent cata- tor. Open data as a concept is widely data made available to the public for strophic events. Such data can also help applicable in various sectors, and its im- free. Public health data should be in- health researchers and, by extension, en- portance has been articulated in the ex- cluded in this category. Unlike personal hance the health of a city, a state, a na- tant literature. Governments around the health data (whose improper release tion, or the world at large. world are already creating policies to might violate an individual’s privacy), The drive for data openness is gaining enhance data openness by providing public health data are often aggregated impetus in various quarters. For exam- detailed descriptions of how data relat- information that can help in making de- ple, the concept of openness has built ing to public welfare should be man- cisions regarding health issues. The ad- noticeable strength within the technol- aged. There is growing emphasis on vantages and benefits that can accrue ogy environment, with a focus on open making these data available to the from having appropriate data on which source application development and

1 Pan American Health Organization, Washington, 2 School of Information Sciences, University of 3 School of Information Studies, Syracuse D.C., United States of America. Send correspondence Illinois at Urbana-Champaign, Champaign, Illinois, University, Syracuse, New York, United States of to Marcelo D’Agostino, at [email protected] United States of America. America.

This is an article distributed under the terms of the Attribution-NonCommercial-NoDerivs 3.0 IGO License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited. No modifications or commercial use of this article are permitted. In any reproduction of this article there should not be any suggestion that PAHO or this article endorse any specific organization or products. The use of the PAHO logo is not permitted. This notice should be preserved along with the article’s original URL.

Rev Panam Salud Publica 42, 2018 1 Review D’Agostino et al. • Open data and public health code sharing. The scientific community KEY CONCEPTS AND of data by the citizens in ways that create has also been pushing for open access to TERMINOLOGIES additional usefulness; and (c) participa- scholarly work, and a few initiatives tory governance, ensuring that citizens have been launched in the past few years Open data have access to data upon which govern- to support this (1). In addition, govern- ment decisions are based. ment agencies around the world are cre- Although the term “open data” seems Some initiatives measure government ating policies to enhance access and use intuitive and self-explanatory, individu- performance around the world regard- of government data by the public at als and organizations may have different ing OGD. large (2). conception of openness. Some concepts The seven countries that ranked high- In the research literature, there are dif- and principles relevant to some sectors est in the 2016 Global Open Data Index ferent viewpoints on the urgency and may not apply to other sectors. For in- (GODI), which is run by the Open the importance of the concept of open- stance, the Project Open Data of the Gov- Knowledge Network, were Taiwan, Aus- ness of public health data. For example, ernment of the United States of America tralia, Great Britain, , Finland, it has been established that disease out- states that open data must be managed Canada, and Norway. These seven coun- breaks such as happened during the after being released (7). This principle tries were ranked from one to five, with Ebola and Zika virus epidemics are in- applies to academic research projects, two or more countries occupying the dicative of the need for countries to de- but they are usually only required to same position. Taiwan occupied the first velop a framework that will provide keep the data for a few years after the position, followed by Australia and guidance for managing public health end of the project. Great Britain sharing the second posi- data. Such frameworks should ensure A widely used definition of open data tion. France is fourth, while Finland, that data collected during public health is , which states that Canada, and Norway all shared the fifth emergencies are accessible to the appro- “open data and content can be freely position (11). priate authorities and in a form that can used, modified, and shared by anyone Some of the data areas on which the help with timely decision-making dur- for any purpose” (8). It is important to GODI assesses openness are: national ing such public health crises. Carney note that this definition is not applicable statistics, government budget, legisla- and Weber (3) have stated that “today’s to government and public health data. It tion, procurement tenders, elections re- public health crises, as exemplified by is crucial that public health data shared sults, national maps, weather forecasts, the Ebola outbreak, lead to dramatic by the government should not be modifi- air quality, company registers, location calls to action that typically include im- able. Thus, in this paper, our definition of datasets, water quality, land ownership, proved electronic monitoring systems to open data is a slight variation of the and government spending. Public health better prepare for, and respond to, simi- Open Definition. We define open data as data is currently not included in this list. lar occurrences in the future.” Having data that can be freely used and shared It is noteworthy that among the top 20 data-driven public health intelli- by anyone for any purpose, under strict countries with high OGD scores, based gence-gathering is only possible if the principles of privacy and confidentiality, on the 2016 Global Open Data Index required data are available and are be- when appropriate. (GODI), 7 of them are in the Americas. ing shared. These 7 countries are Canada, Brazil, Vest and Issel (4) have asserted that Open government data Mexico, United States, Colombia, Argen- there is a need for improved information tina, and Uruguay. system capabilities in public health agen- Around the world, governments and Available studies already show the cies and that these capabilities may be quasi-governmental entities collect mul- kinds of benefits that accrue when citi- the key to those agencies remaining rele- tiple categories of information, including zens utilize openly available government vant in evolving health care systems. on people, events, institutions, and data. For example, Bătăgan demon- There are also increased calls for inte- the environment. With new information strated how open government data could grating public health surveillance data and communication technologies—espe- enhance development in urban areas within health departments. The cost cially the Internet—it is now easy to store (12). In a similar study, a group of Finn- ­efficiency of such an approach to the this information in online databases. Ac- ish researchers looked at the Helsinki Re- ­coordination of care can be striking (5). cess to most of these databases was often gion Infoshare open data service and In managing public health data, various initially restricted to certain agencies of concluded that “open data can signifi- factors across different agencies have the government involved in collecting cantly contribute to open government been considered. According to Nguyen the data (9). However, governments and transparency, citizen’s engagement, (6), one of these factors in relation to pub- around the world are now realizing the better-informed decisions, innovative lic health data on child maltreatment is need to make this information available applications, better services, [and] new “interoperability,” which could be a syn- to the broader public for various reasons. businesses and jobs” (13). onym for openness (6). Some of the reasons for making govern- Government datasets are increasingly In the remainder of this literature re- ment data available, according to the available on the Web, with open data view, we provide an overview of the in- ­information available on the Open Gov- standards over different platforms. In tersection of open data and public ernment Data website (10), are (a) transpa­ most cases, the datasets are in a nonpro- health by first defining and discussing rency, which ensures that a government’s prietary format and are free of charge. open government data (OGD), public activities are made known to the citi- Although such datasets can provide un- health data, and other key concepts and zenry; (b) to release social and commer- limited opportunities for decisionmak- terminologies. cial value by encouraging innovative use ers, institutions, and researchers, their

2 Rev Panam Salud Publica 42, 2018 D’Agostino et al. • Open data and public health Review quality and usability are yet to be deter- city, state, or country in order to help our Scopus (http://www.scopus.com/). mined (14). decision-making in public health? How These articles focused on public health do we manage public health data among data openly released by governments Principles of “open” government countries, especially ones sharing a geo- and international organizations, such as data graphical boundary? At what level do the World Health Organization (WHO). we have access to these datasets, and Some of these openly released govern- On 7-8 December 2007, 30 open gov- what are the requirements needed for us- ment data are accessible online through ernment advocates met to develop a set ing these data? Public health data that government websites, including the of principles for open government data. are available in the right format at the websites of international organizations Their objective was to create a more ro- right time can help in making decisions such as WHO. We did not include schol- bust understanding of what “open” that would prevent health-related ca- arly articles about individual health means and why government data is es- tastrophes. Given that public health was data and those using data gathered by sential to democracy (15). They agreed not listed as part of the areas of open nongovernment entities. Our search pri- that government information would be government data initiatives indicated in oritized journal articles and conference considered open if it complies with the the Global Open Data Index (11), we papers. We used the following search principles listed in Table 1. wonder if there any issues that prevent terms: [1] “open data” AND “public making public health data open in some health” and [2] “open data” AND Public health data countries, or whether it was just an over- “health” AND “government.” sight by the team behind the initiative. We found 27 relevant scholarly articles As opposed to individual or personal- in our search. Predominantly, these arti- ized health, public health concentrates on METHODS cles used open data from either the na- the promotion and protection of the tional government or local governments. health of the entire population within a Literature search Countries whose data were used in some country, state, county, or city. Govern- of the articles include Australia, Brazil, ments promote healthy lifestyles and India, Malaysia, Romania, South Korea, This literature review is driven by the conduct research on preventing injuries and the United States. One study used following research questions: and on detecting and controlling infec- data from the World Health Organiza- tious diseases. At times, governments in 1. What are the themes of publications tion. These countries, along with others different countries—especially ones shar- about government open public health listed in Table 2, have extensive open ing a border—will agree and cooperate data? data guidelines published online. on public health efforts that promote the 2. What principles are considered im- well-being of all their citizens (16-18). portant for open government public DISCUSSION We often distinguish between public health data? and personal health, especially in rela- 3. What factors get in the way of open One important element stressed in a tion to data sharing. However, it is im- government public health data? few of the research articles is the concept portant to note that public health data 4. How is government-released public of “.” Linked data is a term are usually obtained from individual or health data being used in research coined by Tim Berners-Lee (20) to de- personal health data that have been ano- studies? scribe best practices to publish and link nymized in a required manner (19). 5. How many publications are about structured data to other data available on Therefore, for effective public health data or used government-released public the Web. Tilahun et al. (21) have asserted collection and management, there must health data from countries in the that publishing data as linked data will be adequate and effective personal health Americas? allow it to be understood by machines data collection and management, espe- and discovered by search engines. cially for diseases that can result in major We accessed scholarly articles from Linked data is not a new term in the outbreaks. How do we effectively man- two databases: PubMed (http://www. health community. The Government of age health data generated from a county, ncbi.nlm.nih.gov/pubmed/) and the United States has published some of

TABLE 1. Principles of open government data as formulated by a group of open government advocates in 2007

Principle Description Complete All public data is made available. Public data is data that is not subject to valid privacy, security, or privilege limitations. Primary Data is as collected at the source, with the highest possible level of granularity, not in aggregate or modified forms. Timely Data is made available as quickly as necessary to preserve the value of the data. Accessible Data is available to the widest range of users for the widest range of purposes. Machine processable Data is reasonably structured to allow automated processing. Nondiscriminatory Data is available to anyone, with no requirement of registration. Nonproprietary Data is available in a format over which no entity has exclusive control. License-free Data is not subject to any copyright, patent, trademark, or trade secret regulation. Reasonable privacy, security, and privilege restrictions may be allowed. Source: Fiorenza (15).

Rev Panam Salud Publica 42, 2018 3 Review D’Agostino et al. • Open data and public health

TABLE 2. Some country-specific open data guidelines effective collection of demographic and malaria data in western Kenya. Country Open data platform(s) In an example relating to policy, Gas- Australia Australian Bureau of Statistics (http://abs.gov.au/) ner et al. (5) presented a case study of Data SA (https://data.sa.gov.au/) data sharing between public health pro- PHIDU (http://publichealth.gov.au/data) grams in New York City. They examined Queensland Health (https://www.health.qld.gov.au/) public health laws and agency policies Brazil Brazilian Government Open Data (http://dados.gov.br/) for data sharing across surveillance pro- Interagency Health Information Network (http://www.ripsa.org.br/) grams for HIV, sexually transmitted dis- Dados Recife (http://dados.recife.pe.gov.br/) eases, tuberculosis, and viral hepatitis. Rio Open Data (http://data.rio/) The authors reported that recent changes Data Warehouse Web (http://www.saude.mt.gov.br/sistemas) to state laws provide greater opportuni- Canada Open Data Canada (http://open.canada.ca/en/open-data/) ties for data sharing but that agency pol- Chile Open Data Portal (http://datos.gob.cl/) icies must be updated to allow increased Colombia Open Data Colombia (https://datos.gov.co) data integration. This finding, for one Jamaica Jamaica Open Data portal (http://data.gov.jm/) city, would likely be even more impor- Malaysia Malaysian Government Open Data (http://www.data.gov.my/) tant when considering entities spread Mexico Open Data Mexico (https://datos.gob.mx/) across larger geographical areas. Romania Romanian Government Open Data (http://data.gov.ro/) South Korea Korean Government Open Data (https://www.data.go.kr/) Key policy issues United States U.S. Open Government Data (https://www.data.gov/) HealthData.gov (http://www.healthdata.gov/) Our discussion so far has presented National Center for Health Statistics (https://www.cdc.gov/nchs/) evidence supporting the open data OpenFDA (https://open.fda.gov/) movement around the world. However, OpenColorado (http://data.opencolorado.org/) to have greater benefits from open access Health Data NY (https://health.data.ny.gov/) within the public health sector will re- NYC OpenData (https://nycopendata.socrata.com/) quire conscientious efforts addressing Uruguay Open data catalog (https://catalogodatos.gub.uy/) many of the issues that are pertinent and Source: Prepared by the authors from their study data. peculiar to that sector. As shown in ­Figure 1, effective policy formulation for open public health data must deal with its health datasets as linked data, accord- products, scientists can utilize such data four fundamental concerns: benefits, ing to Hendler et al. (22). in their research work, and nongovern- risks, examples, and attributes. These In fact, according to those authors, the mental agencies can study such data to four issues are key concerns for us after most extensive use of linked data from know more about areas of public health. having considered some of the concepts Government of United States datasets There are a few questions about openly discussed in the extant literature on open has been by the health community. In Ro- available public health data. For exam- health data. We note that it is important mania, Rinciog and Posea, two faculty ple, what are the factors that may en- to articulate the potential benefits of members at the University Politehnica of hance or inhibit openly available public open data, potential risks associated with Bucharest, have transformed the Roma- health data? How do we position our- it, examples of data that should be made nian public health data, originally pub- selves to solve problems in such areas as open, and the attributes of such data. lished in XLS format, to linked data (23). policy and technical capability? In terms of potential benefits, it is es- Linked data will enable quick access to Vest and Issel (4) conducted a cross- sential to highlight the importance of similar and complementary data, espe- sectional analysis of organizational fac- sharing health data within a public pol- cially during a public health emergency. tors associated with gaps in data sharing icy supporting open public health data. between state health agencies and local The public needs to be aware, without Open data and public health health departments. They found that the any ambiguity, of the potential benefits proportion of local health departments as they relate to subgroups, and to the There are many advantages or benefits experiencing a data sharing gap was overall population within a given city, associated with open data. These include 34.0% for childhood immunizations, state, or country, or even across coun- transparency and democratic control, par- 69.8% for vital records, and 81.8% for re- tries. Those benefits can give credence to ticipation, self-empowerment, improved portable conditions. They also reported the content of such public policy. or new products and services, innovation, that increased state health agency size Another issue of potential concern is and new knowledge from combined data and technological capacity reduced the risks associated with data sharing. It is sources and patterns in large data vol- likelihood of gaps. important as part of a public policy to in- umes (24). Openly available public health In a similar study on the importance of dicate potential risks that may be associ- data can bring about these benefits as adequate technical capability with re- ated with making any type of public well. For example, government entities gard to collecting and managing public health data open. It is also important to can use public health data for better health data, Homan et al. (25) concluded identify ways to limit such risks. Possible decision-making. Private entities can ap- that adopting technological innovations benefits need to be weighed against any ply such data to create health-related was important in enabling the quick and potential risks.

4 Rev Panam Salud Publica 42, 2018 D’Agostino et al. • Open data and public health Review

FIGURE 1. Key policy issues with open public health data two questions the authors of this paper and other scholars may want to investi- gate are: 1) How open is health data Potential that could help make effective public benefits of openness health decisions in the United States? 2) Is the current policy framework strong enough to cover public health data shar- ing, especially during public health Attributes of Open public Examples emergencies? data health data of data Another panelist at that April 2014 ses- sion was Dominique Babini, the open ac- cess program coordinator of the Latin American Council on Social Sciences Potential (CLACSO). She pointed to a number of risks of Latin American countries whose legisla- openness tures have approved open access policies that are similar to those in the United States, requiring the creation of open ac- Source: Prepared by the authors from previous studies, including Crouch and Wilson (26). cess digital repositories for government- funded research results. Brazil did that in Additionally, mention must be made enhance our understanding of the 2011, Argentina and Peru did so in 2013, of the examples of the specific types of world. These benefits are only possible and Mexico followed in 2014. data that would constitute public health when there is access to and utilization According to a report from the United data and thereby qualify to be made of the results of this scientific research. Nations Educational, Scientific and Cul- open to a larger, public audience. Awareness of that great potential has tural Organization (UNESCO) (28), The attributes of such examples of led the United States Government to de- among Latin American countries, Brazil health data must also be adequately velop open access policies contained in has the largest number of institutional stated. We know that it is the collection such documents as the Office of Man- open access policies (in total, 16), trailed of individual health data that is aggre- agement and Budget (OMB) Circular by Peru (6), Argentina (4), Venezuela (4), gated as public health data. There must No. A-130 (1). Colombia (3), Mexico (3), and Bolivia (1). be guidelines that will prevent the re- In that April 2014 panel, Joseph also In addition to the efforts on open ac- lease of data attributes that can be used said that, in conscientious efforts over cess policies being made in individual to directly link publicly available health the years, the Congress of the United Latin American countries, there are vari- data to specific individuals. States has endorsed policy statements ous major regional projects and initia- supporting open access to government- tives. According to UNESCO (28), one of Open data in the Americas funded research work at different levels. these is the Scientific Electronic Library Over time, the language in those docu- Online (SciELO). SciELO is a coopera- Any discussion of open data in the ments has shifted from encouraging schol- tive, decentralized platform for elec- Americas needs to consider the issue of ars to deposit their government-funded tronic publishing of scientific journals, open access. In an April 2014 panel dis- results (articles plus data) on open plat- with national focal points in 15 countries, cussion at Columbia University in New forms within a given period to mandating mainly in Latin America. Other regional York City, open access initiatives across them to do so as a condition for financial initiatives include the Red de Revistas the Americas were examined (27). One support. Científicas de América Latina y el Caribe, of the panelists was Heather Joseph, the Apart from the OMB Circular No. España y Portugal (Redalyc); Regional executive director of the Scholarly Pub- A-130, some of the documents embody- Cooperative Online Information System lishing and Academic Resources Coali- ing support for open access in the for Scholarly Journals from Latin Amer- tion (SPARC). She provided an overview United States include the Copyright Act ica, the Caribbean, Spain and Portugal of the policy framework in the United (17 U.S.C. 105), the Freedom of Informa- (Latindex); Biblioteca Virtual en Salud States regarding open access to schol- tion Act (FOIA), the Paperwork Reduc- (Virtual Health Library) (BVS); Latin-​ arly articles and data that are the output tion Act, the Electronic Freedom of American Citations in Social Sciences of government-sponsored projects, Information Act Amendments of 1996, and Humanities (CLASE); Index of Lat- through such agencies as the National and the Government Paperwork Elimi- in-American Journals in Science (PERI- Institutes of Health (NIH). She said that nation Act. ODICA); Agricultural Information and the drivers for open access policies in Joseph also noted that the first agency Documentation Service of the Americas the United States include the fact that, of the United States Government to (SIDALC); and Digital Library of the Ca- on a yearly basis, the federal Govern- ­implement an open access policy was ribbean (dLOC). ment’s investment in basic and applied the NIH. The research community could Given these many regional and subre- scientific research is around US$ 60 bil- benefit from understanding how the gional efforts on open access, sharing lion. Spending such huge public sums NIH push for openness has affected public health data in the Americas should on research ought to help generate public health data sharing at large. be less of a challenge than some persons new ideas, make new discoveries, and Based on the NIH open access policy, might think. This optimism is buttressed

Rev Panam Salud Publica 42, 2018 5 Review D’Agostino et al. • Open data and public health by initiatives around the world that mea- that collect official data. However, there data in the Americas. Any national data sure government performance on open is a lack of standardization and interop- governance strategy must address po- government data, which show that many erability among countries and among tential benefits, potential risks, examples countries in the Americas have adopted portals and repositories. Further, while of data that could be shared, and the at- open access policies (11). our study found examples of exception- tributes of such data. The central challenge in relation to ally well-organized national open data We hope that the information pro- open data on public health in the Ameri- portals, fewer than one-third of the vided in this article will support coun- cas is that of developing regional frame- world’s countries had principled, orga- tries and health-related organizations as works that can be adopted or adapted by nized portals set up to systematically they develop national strategies on inter- each country through national or subna- share open data. connected and interoperable open health tional policies. Such frameworks must be data portals, especially within the con- easy to apply and to fit into the needs of Limitations text of national public health strategies each country, as well as those of the re- and in line with other open government gion overall. To help achieve this, teams This review is based on the available initiatives. of experts must work with individual secondary information. A further analy- countries in sharing ideas and lessons sis of primary data related to current ini- Acknowledgments. We thank the stu- learned and in focusing on common at- tiatives that have not been made public dents of the University of Illinois at tributes that can help resolve complex may provide further evidence of policies Urbana-Champaign who participated in decision-making issues. that are not publicly available. the Global Health Informatics class in the spring of 2016. We presented initial ideas Open data in the world CONCLUSIONS there that we later expanded on in this paper, and the students’ comments and Government data repositories are un- Open data measurement is still a new questions helped us clarify our views. equally spread around the world, and and experimental field. A national data We are also grateful to the anonymous the health information in them is not al- governance strategy could be a critical reviewers for questions and observations ways open and available online, accord- tool to improve health analysis and to that assisted us in refining the paper. ing to a 2016 observational study (29). forecast possible scenarios (30). In this While there are international mandates paper, we have identified some current, Conflicts of interest. None declared. and commitments on collecting health functional open data policies in many data, current open health data portals countries, including in North America Disclaimer. Authors hold sole respon- use a variety of formats. These portals and South America. Our analysis has fur- sibility for the views expressed in the mainly contain data on new diseases, di- ther shown that there are currently no manuscript, which may not necessarily rectories of health institutions, and de- articulated policy guidelines for the col- reflect the opinion or policy of the RPSP/ tailed descriptions of regulatory agencies lection and management of public health PAJPH or PAHO.

REFERENCES

1. Stebbins M. Expanding public access 6. Nguyen LH. A public health response to from the Helsinki Metropolitan Area. Case to the results of federally funded re- data interoperability to prevent child mal- Helsinki Region Infoshare www.hri.fi. Stat search. Available from: https://www. treatment. Am J Public Health. 2014;104(11): J IAOS. 2015;31(1):117–22. http://doi. whitehouse.gov/blog/2013/02/22/ex- 2043–8. org/10.3233/SJI-150873 panding-public-access-results-federally- 7. Project Open Data. Principles. Available 14. Martin EG, Law J, Ran W, Helbig N, funded-research Accessed on 1 August from: https://project-open-data.cio.gov/ Birkhead GS. Evaluating the quality 2016. principles/ Accessed on 22 August 2016. and usability of open data for public 2. Sheehan J. Increasing access to the results 8. Open Definition. The Open Definition. health research: a systematic review of of federally funded science. Available Available from: http://opendefinition. data offerings on 3 open data platforms. from: https://obamawhitehouse.archives. org/ Accessed on 10 August 2016. J Public Health Manag Pract. 2017;23(4): gov/blog/2016/02/22/increasing-ac- 9. Samuel NO. Research support and open e5–e13. cess-results-federally-funded-science access: notes from Nigeria. Afr J Lib Arch 15. Fiorenza P. A brief history of open data Accessed on 28 February 2018. Inf Sci. 2016;26(1):17–29. and GIS. Available from https://www. 3. Carney TJ, Weber DJ. Public health intelli- 10. Open Government Data. Welcome to govloop.com/a-brief-history-of-open- gence: learning from the Ebola crisis. Am J Open Government Data. Available from: data-and-gis/ Accessed on 28 February Public Health. 2015;105(9):1740-4. doi: http://opengovernmentdata.org/ 2018. 10.2105/AJPH.2015.302771 Accessed on 6 July 2016. 16. American Public Health Association. 4. Vest JR, Issel LM. Factors related to public 11. Open Knowledge International. Global What is public health? Available from: health data sharing between local Open Data Index; 2015. Available from: https://www.apha.org/what-is-public- and state health departments. BMC http://2015.index.okfn.org/place/ health Accessed on 6 July 2016. Health Serv Res. 2014;49(1pt2):373–91. Accessed on 29 January 2018. 17. CDC Foundation. What is public health? doi: 10.1111/1475-6773.12138 12. Bătăgan L. The role of open government Available from http://www.cdcfounda- 5. Gasner MR, Fuld J, Ann D, Jay KV. Legal data in urban areas development. Inform tion.org/content/what-public-health and policy barriers to sharing data be- Econ. 2014;18(2):80–7. doi.org/10.12948/ Accessed on 6 July 2016. tween public health programs in New issn14531305/18.2.2014.08 18. HealthData.gov. Welcome to HealthData. York City: a case study. Am J Public 13. Jaakola A, Kekkonen H, Lahti T, Manninen gov. Available from http://www.health- Health. 2014;104(6):993–7. A. Open data, open cities: experiences data.gov/ Accessed on 6 July 2016.

6 Rev Panam Salud Publica 42, 2018 D’Agostino et al. • Open data and public health Review

19. Richardson V, Milam S, Chrysler D. Is shar- Paper presented at: 2015 E-Health and Access Portal. Latin America and the ing de-identified data legal? The state of Bioengineering Conference (EHB); 2015 Caribbean. Available from: http://www. public health confidentiality laws and their Oct 19-21, Iaúi; Romania. unesco.org/new/en/communication- interplay with statistical disclosure limita- 24. Open Knowledge International. Why and-information/portals-and-platforms/ tion techniques. J Law Med Ethics. open data? Available from: http://open- goap/access-by-region/latin-america- 2015;43:83–6. http://doi.org/10.1111/jlme. datahandbook.org/guide/en/why-open- and-the-caribbean/ Accessed on 1 March 12224. data/ Accessed 12 July 2016. 2018. 20. Berners-Lee T. Linked data. Available 25. Homan T, Di Pasquale A, Kiche I, Onoka 29. Greenberg CJ, Narang S. World Health from: https://www.w3.org/DesignIssues/ K, Hiscox A, Mweresa C, et al. Innovative Organization member states and open LinkedData.html Accessed on 19 September tools and OpenHDS for health and demo- health data: an observational study. 2016. graphic surveillance on Rusinga Island, Epidemiol Biostat Public Health. 2016; 21. Tilahun B, Kauppinen T, Keßler C, Fritz, F. Kenya. BMC Res Notes. 2015;8(1):397. 13(3). Design and development of a linked open http://doi.org/10.1186/s13104-015- 30. Global Open Data Index. Open Data Index data-based health information representa- 1373-8 2015: civil society audit of the state of open tion and visualization system: potentials 26. Crouch EAC, Wilson R. Risk/benefit anal- data. Available from: http://2015.index. and preliminary evaluation. J Med ysis. Available from: https://www.osti. okfn.org/about/ Accessed 15 February Internet Res. 2014;16(10):e31. http://dx. gov/biblio/5557457 Accessed on 28 2018. doi.org/10.2196/medinform.3531 February 2018. 22. Hendler J, Holm J, Musialek C, Thomas G. 27. Joseph H, Sinatra M, Babini D, Graham US government linked open data: Semantic. PM. Open access in the Americas [mp4 re- data.gov. IEEE Intell Syst. 2012;27(3):25-31. cording]. Available from: http://dx.doi. http://dx.doi.org/10.1109/MIS. org/10.7916/D8736P0S Accessed 28 2012.27 February 2018. 23. Rinciog O, Posea V. Publishing Romanian 28. United Nations Educational, Scientific Manuscript received on 30 September 2016. Revised public health data as Linked Open Data. and Cultural Organization, Global Open version accepted for publication on 8 February 2017.

RESUMEN Este artículo ofrece un panorama de la intersección entre los datos abiertos y la salud pública al definir en primer lugar qué son los datos gubernamentales abiertos, los datos sobre salud pública y otros conceptos fundamentales y términos pertinentes. Datos abiertos y Hay percepciones dispares sobre la premura y la importancia de la apertura de los salud pública datos sobre salud pública. Se ha establecido que los brotes de ciertas enfermedades, como las epidemias por los virus del Ébola y del Zika, demuestran la necesidad de que los países elaboren un marco que oriente la gestión de los datos sobre salud pública. Dicho marco debe garantizar que los datos recopilados durante las emergencias de salud pública sean accesibles para las autoridades competentes y en una forma que contribuya a la toma oportuna de decisiones durante estas crisis de salud pública. En este artículo, destacamos las políticas de datos abiertos existentes en diversos países, incluidos varios de la Región de las Américas. Nuestro análisis muestra que actual- mente en muchos países no hay directrices articuladas de políticas públicas para la recopilación y gestión de los datos sobre salud pública, en especial en América Latina. Proponemos que toda estrategia nacional de gobernanza relativa a los datos debe abordar los posibles riesgos y beneficios, ejemplos de los datos que podrían compar- tirse y los atributos de tales datos. Por último, subrayamos que el interés fundamental en la Región de las Américas debe ser la creación de marcos regionales para datos abiertos sobre salud pública que cada país pueda adoptar o adaptar mediante las políticas y estrategias nacionales o subnacionales apropiadas.

Palabras clave Sistemas de información; fuentes de datos; salud pública; sistemas de salud; confidencialidad.

Rev Panam Salud Publica 42, 2018 7 Review D’Agostino et al. • Open data and public health

RESUMO Este artigo expõe um panorama da interseção entre dados abertos e saúde pública, começando por definir o que são dados abertos do governo, dados de saúde pública e outros conceitos fundamentais e terminologias relevantes. Existem distintas percep- Dados abertos e ções quanto à premência e à importância da abertura de dados de saúde pública. saúde pública Reconhecidamente, os surtos de doenças, como os ocorridos nas epidemias do vírus Ebola e vírus zika, apontam para a necessidade de os países desenvolverem uma estrutura para direcionar o gerenciamento dos dados de saúde pública. Esta estrutura deve servir para garantir que os dados coletados nas emergências de saúde pública estejam acessíveis às autoridades cabíveis em uma forma que possa subsidiar a tomada de decisão oportuna durante tais crises de saúde pública. No artigo, destacam-se as políticas de dados abertos de diversos muitos países, inclusive dos países na Região das Américas. Nossa análise demonstra que vários países, sobretudo na América Latina, não possuem diretrizes claramente definidas de políticas para a coleta e o gerenciamento de dados de saúde pública. Recomendamos que qualquer estratégia nacional de governança de dados nacionais precisa contemplar os possíveis benefícios e riscos, explicitar os dados a ser compartilhados assim como descrever os atributos de tais dados. Por fim, salientamos que a principal preocupação nas Américas deve ser o desenvolvimento de estruturas regionais para dados abertos em saúde pública que possam ser postas em prática ou adaptadas por cada país como parte de estratégias e políticas nacionais ou subnacionais adequadas.

Palavras-chave Sistemas de informação; fontes de dados; saúde pública; sistemas de saúde; confidencialidade.

8 Rev Panam Salud Publica 42, 2018