Recommendation for Electronic Health Records in Haiti more

Recommendation for Electronic Health Records in Haiti Travis Horsley & Patrick Linton Georgia Institute of Technology trhorsley@ga.tech.edu, patrick.linton@gatech.edu Abstract: Haiti is the poorest country in the Western Hemisphere. After the 2010 earthquake, much of the infrastructure in the capital area was destroyed, including many hospitals and clinics. The paper-based patient records stored within were destroyed. Many tent- hospitals staffed by international volunteers had to treat trauma patients without any sort of medical record. An electronic health records (EHR) system could have maintained records for use in this crisis. This paper analyzes several EHR systems that have been implemented in the past in countries with HDI similar to Haiti, develops a typology for analyzing these systems, and proposes a path toward implementing a new EHR system in Haiti. 1. Introduction In 2010, and earthquake measuring 7.0 in magnitude on the Richter scale struck Haiti. Although the epicenter was centered in the town of Leogane, destruction was widespread across the southern region of the country, affecting regions such as the capital city of Port-au-Prince, Jacmel, and Petit-Goave. Global media coverage and on-the-ground accounts have given a broad assessment of damage in the region, including the impacts on the people, infrastructure, and political stability in the area. The earthquake and its subsequent aftershocks caused building structures to collapse. Likewise, death tolls from the natural disaster are inconclusive. Some estimates from European groups such as Radio Netherlands Haiti showed estimates under 100,000 whereas other conservative estimates gauge the toll around 230,000 persons. Hundreds of thousands of other Haitians were displaced by the earthquake and of those, many were injured and needed medical attention. However, as hospitals collapsed, and the remaining ones overflowed, many people were sent to tent hospitals run by aid groups and had to be treated without any report of medical history and without the ability to add these treatments to their medical record. 1.1 The Problem The problem is that Haiti lacks a secure, stable system to store health records. Haiti also lacks the infrastructure to create an electronic health records system. We intend to propose an path to implementation of an electronic health records system that will work in the context of Haiti (which lacks infrastructure in many places and may not have full access to a network or stable electricity in all places). Adoption of EHR for medical professionals in Haiti is necessary and crucial, especially in a post-natural disaster context. A lack of knowledge exists in the literature concerning native adoption (that is to say, adoption by in-country physicians in Haiti) of EHR systems in a post-natural disaster scenario. A wealth of literature is available concerning the use of EHR by outsiders coming into Haiti for disaster relief, yet little analysis focuses on the training of individuals and organizations within Haiti to use these systems years ahead. The purpose of this study is to provide an accurate synopsis of organizational conditions necessary to implement an EHR system in Haiti. This study conducts a survey of relevant literature on EHR analyses from countries that are within a similar development bracket as Haiti. Our goal is to provide a typology of EHR methods that have been tested and successfully implemented in medium-development economies. This typology can be utilized by NGOs, governments (such as Haiti), and other intergovernmental institutions to have a metric for successful strategies of implementation of an EHR 1 system, bot in micro-community level settings, or macro-state health system contexts. A comparative policy analysis of project implementation techniques is the main school of thought that is used in this study. The theoretical framework for this study grounds our meso-organizational analysis in a comparison of secondary data on EHR implementation in other countries, found in scholarly articles in databases, as to provide a basis for EHR program implementation policies in Haiti. This theoretical framework will be further explored in Section 1.3: Why Comparative Analysis? 1.2. Introduction to this Study This study is both an environmental scan and project implementation plan for Haiti. Thus, our analysis will discuss usage and factors that inhibit adoption of e-health record systems in regions with similar Human Development Index scores. The questions we aim to answer with this study are: Which factors at the individual and organizational level predict m-EHR adoption by medical professionals? What are system component level (mobile handset, database) features that contribute to m-EHR adoption? Are the individual/organizational level factors and the system component factors to m-EHR adoption interrelated? The research design proposed in this study focuses on a systems study of countries similar to Haiti in development as ranked by the Human Development Index (HDI). This research posits that similar patterns of EHR implementation exist at the meso-level in developing countries within the ten countries surrounding Haiti on the United Nations Human Development Index. Further, this research posits that program complexity (of EHR initiatives) is a byproduct of support internal to the program host country, both financially and organizationally. The dependent variable in this research is the organizational structure of EHR project implementation and the independent variable is the barrier of implementation of a EHR system in a medium development country on the HDI Index. Our level of analysis is 1-2 implementations per country in Human Development Index strata around Haiti. 1.3 Why Comparative Policy Analysis? Techniques of comparative policy analysis for this research are grounded in literature surrounding comparative policy analysis for developing countries and comparative policy analysis of healthcare system structures in different countries. Comparative policy analysis on the international level emerged as a discipline of study in the 1970s, after a focus on comparative state policy analysis emerged (focusing on similarities of policy solutions between states). "Public policy and problem solving are extraordinarily imitative arts. [...] Rarely do policy makers embark upon entirely new courses of action; rather they borrow heavily from an apparently finite, existing repertoire of policy solutions" (Leichter) The Leichter text goes on to indicate different variables that influence, to some degree, health policy in countries: the type of political regime (a structural variable), a country's process of policy making and policy implementation, political culture and ideology, economic wealth or development of a nation, and the process of turning budget expenditures in to health system developments. Because this study pulls from secondary data sources to formulate a basis for the implementation of a EHR system in Haiti, theory about the efficacy for such a study must be grounded in the literature as well. Sarah Boslaugh's text, Secondary Data Sources for Public Health: A Practical Guide, indicates that the use of secondary data is a common practice in understanding the functions of health systems. As is to be expected in this research, access problems prevail. "The process of locating appropriate secondary data is not always straightforward" (Boslaugh). A comprehensive approach must be taken to acquire data relevant to the field of study; in this case, an ontology of search terms applied across multiple databases will reveal relevant data by country. 2. Electronic Health Records According to Marion Ball of Johns Hopkins University, e-health is "the use of information and communication technologies such as physician web pages, e-mails, and healthcare portals to improve or enable stronger and more effective connections among 2 patients, doctors, hospitals, laboratories, pharmacies, and supplies" (Omary, Lupiana, Mtenzi, & Wu). One important subset of e-health is electronic health records. Electronic health records are an electronic set of patient records that contain information on a patient's medical history, demographics, laboratory data, medication, and other important medical information. Paper records are bulky, retrieving them is labour-intensive, and cannot be utilized by multiple departments at the same time (Williams). Computer- based records take up as much hospital space as the device that accesses them, search is almost instantaneous, and can be accessed by multiple departments at the same time. Computerised systems help reduce medical errors: Automatic range checks will ask for confirmation if a number entered into a field is outside of an expected range; Systems can check prescriptions against patient drug allergies, dangerous drug combinations, and misunderstandings caused by illegible hand-written prescriptions; They can also identify abnormal results that could be signs of disease. EHR is also helpful in gathering statistical data, in order to track the health of a population and follow patterns of epidemics. Because the data is stored in a server and backed up offsite, health records still exist in the case of disaster. This is especially important in cases such as the earthquake that occured in Haiti in early January 2010. Hospitals were destroyed and thousands of injured Hatians had to be treated without any records in tent hospitals that popped up around the island. However, there are some criticisms of EMR. James Cimino of Columbia University warns of risks associated with EMR, such as cognitive overload, disorientation, and blind acceptance of information and recommendations (Cimino, Teich, Patel, & Zhang). Another concern is security, which is especially sensitive because of doctor-patient confidentiality. Breached information about a patient's health can often be a cause for embarrassment. 2.1 Electronic Health Records in Haiti EHR systems have been implemented in Haiti in the past. In 1996, Partners in Health, a Boston- based health-advocacy nonprofit, and Haiti-based Zanmi Lasante implemented their 'PIH-EMR' system in seven health clinics in rural Haiti. Each clinic was provided with a satellite internet connection. The web- based system was backed by an Oracle database (open- source). An offline component was even developed to overcome unreliable internet. According to Fraser et aL, This system showed "the feasibility of implementing a medical record system in remote clinics in a remote area with virtually no infrastructure and limited technical expertise." In the aftermath of the January 2010 earthquake, smaller mobile electronic health records projects, such as the iChart application, popped up. The iChart application was an application for the Apple iPhone that foreign earthquake-relief volunteers could use to store records about patients they admitted in their tent hospitals. It was a quick, temporary solution to the problem of health records in Haiti. 3. The Research In this study, a comparison was conducted of implementations of EHR initiatives in countries with a similar Human Development Index (According to the UN Development Programme, Haiti had an HDI of .532 in 2009, and has probably significantly lowered since then). In our study, we compare implementations of EHR in Haiti as well as in ten countries with similar HDIs: Nepal, Madagascar, Bangladesh, Kenya, Papua New Guinea, Sudan, Tanzania, Ghana, Cameroon, and Mauritania. In addition to HDI as a metric for choosing these countries, 8 of the 10 countries have low teledensity ratings from the ITU (per 100 inhabitants). Moreover, to make our selection of these countries more rigorous, the World Health Organization ranking of world health systems (2000) was observed to see if the countries selected from the HDI fell into a similar category. All countries (when available) fell below 135 on a ranking of 190 countries, with Haiti positioned at 138. Implementation methodologies were sourced from a series of databases available through the Georgia Institute of Technology. Databases that were searched included some of the following, per recommendation of reference librarians with understanding of the literature: InSpec Engineering Village, CSA Illumina - PAIS International, the 3 Columbia International Affairs Online database, Medline EBSCOhost, Web of Science, Social Science Index, IEEE Explore, scholarly blogs on health information systems issues, government documents from countries being studied, and Google Scholar. Terms searched in each of these databases fall along an ontology of search terms relating to the following search subjects: m-health, mobile health, electronic health record, electronic medical record, computer- based patient record, knowledge-based medical record, computerised medical record, patient-carried medical record, personal health record, digital medical record, and the list of countries being researched. This ontology is broad-scoping in order to receive a wide net of literature with EHR implementation analyses. Data regarding implementation analysis were divided by host country, with the prospect of obtaining ten analyses of EHR implementation models per country. We then compared recurring themes found in each source, quantifying them based on their successes and failures, and retain the most important themes. Also, themes that have a negative impact on the success of a project were rejected. Toward the conclusion of our research, only one to two analyses per country were found that met the academic rigor that was needed for a project implementation study. 3.1 Results The trends discovered through this research show that partnerships are the preferred form of organizational response to the policy problem of a lack of Electronic Health Records. Partnerships formed between institutional units (government-to- government, academia-to-academia, government-to- NGO, and NGO-to-NGO) as a basis to problem solving allow for different ideas to flourish between institutions. Another recurring trend that resulted from an analysis of this literature was that lack of electricity was a recurring theme (and was even a surprise to many researchers). Urban areas were often completely electrified, but still had many regular blackouts and many hospitals could not quickly switch to backup generators. Other environmental factors that appeared within the literature included insufficient data connections and lack of computer training in other areas. Lack of capital internal to the host state of an implemented EHR program would add to the complicated implementation of such initiatives. The text indicated a clear dichotomy between the implementation of systems that used MS Access or SQL-based databases, citing the pros and cons of the closed and open nature of each code structure respectively. Moreover, dependent upon the system being implemented, different web interfaces were utilized for displaying the e-health information. A lack of interoperability between existing EHR systems was cited in the literature. Typically, conclusions led to a call for standards in EHR. However, some also noted that maintaining EHR standards is difficult due to the complexity and different needs of each unit and department with the healthcare industry. Other antecedent variables that were not included in our dataset, but were tagged as miscellaneous (but important antecedent variables) included patients lack of a unique identifies in one country. This lack of national ID lead to issues with transferring secure documents over electronic mediums. 3.2 Findings A typology that was developed for the study of implementation analysis papers involved structuring the solution complexity into a tiered system. Implemented and proposed-implementation E-Health projects were then assigned a tier related to the complexity of the individual program. The complexity is stratified into Level 1 (simple solution), Level 2 (medium-complexity solution), and Level 3 (complex-comprehensive solution). ■Email-only ■Email and photo-based consultancy ■Telephone only •Only computers in use •Lack of system sustainability cited in literature •Computers in use •Data back-up (internal server, paper based, or mix) •Data stored at children and parent node computers Figure 1: The program complexity typology. 4 They typology of the program complexity described one side of the policy process model that the researchers desired to express. The expression of the policy process was a compilation of variables that were qualitatively assessed based on a reading of the literature. The literature was coded in such a way that four elements were extrapolated from the literature: organizational implementation, barriers to entry, finance, and system components. Each country will be described based on the data found. r I Power I [internet] $ - Funding from Infl Loan, Iridic Countries, & Slate Support } Level 1: Simple Solution Figure 2: The policy implementation process model for Papua New Guinea. | Power 1 [inlernell |Financej * d U H * L Figure 3: The policy implementation process model for Haiti. $ - Funding from Int'l Funds & Charily Level 2: Medium- Complexity Solution Figure 4: The policy implementation process model for Kenya. Level 1 and Level 2 program implementation were implemented in countries that showed two or more organizational barriers to entry. Level 1 and Level 2 program implementations were always financed from institutions outside of the host country, sometimes having minimal internal financial support from the host country. Level 3 program implementation was observed in two countries, both sharing the same characteristic of only having financial support internal to the host country. Likewise, Kenya, which was one of the countries with, had an embedded NGO that was created as an offspring of an academic partnership between to educational institutions and two "open doors" were cited in the assessment instead of barriers to entry. LEVEL 1 - Papua New Guinea Papua New Guinea's internal support was coded as an NGO, which was technically a church acting as a stand alone institution for implmenting the EHR solution listed in the literature. Low Internet penetration and a monopoly on the mobile phone industry in the country, coupled with unreliable power as cited in the literature were two barriers to entry. Support for this initiative was jointly funded by the state and the church internally. Externally, financial support was given from the governments of Japan, New Zealand, and the United States of America. A Level 1 solution was implemented that mostly relied on email or telephone based consultancy as an offering of e- health. Data on the "EHR" side was often stored remotely on the consultant's computers, leaving the recorded data inaccessible within the host-country until the consultant released their diagnosis or advice. LEVEL 2 - Haiti A partnership between a foreign NGO (Partners in Health) and a local NGO (Zanmi Lasante) was formed. Amidst lack of electricity in rural areas and low penetration of electrical infrastructure in urban regions, Haiti also has a infrastructure deficiency in Internet penetration. With funding for programmatic assistance in the realm of EHR coming from the Global Fund for AIDS, an external financial stream, the program's likelihood of being self-sustainable is bleak. Although this solution had components of a system that would be tagged as Level 3 complex-comprehensive (more than two S - Funding from State . & internal philantrophy Level 3: Comprehensive- Cornple* Solution J 5 computers, server, etc.), it was coded as Level 2 medium-complexity based on the literature's conclusion that such a system would need more internal support to survive long term. LEVEL 3 - Kenya Kenya was the only country that was observed in this assessment that organizationally structured itself differently, in such that a collaboration between a foreign University (Indiana University) and a state university (Moi University in Kenya). The interesting anomaly in this assessment was that out of this collaboration was formed an embedded NGO that was self-sustaining, the Moi Teaching and Referral Hospital. Likewise, internal policy was supportive of quality healthcare, with political support backing the use of Health Information Technology. Financial support, also according to the literature, came internally from Kenyan government programs as well as internal charitable efforts, the literature cites. The solution that was implemented is the only Level 3 complex-comprehensive solution assessed in this review. It used a broad mix of computer systems, servers, as well as backup both in paper form and digital form. The paper form of the health record could also be transferred in copy form to the patient so that he or she has a copy of medical records. The sustainability of this system appears strong based on internal support of the host country. 4. Recommendations Because the Level 3 (complex-comprehensive) solutions produced the best results, we recommend the following: 1. An institution should be formed internally within Haiti that is a byproduct of an international collaboration between Haiti and a developed state. This would mean that a new, permanent NGO would have to be formed to help to develop and initially fund the system, continually train staff on how to use the system, and help to maintain the system. 2. Decision-makers should buy into the necessity for policy mechanisms that are pro-e-Health. This would mean that the NGO should lobby to the legislature for policies that allow for systems like this, make medical records mandatory (not necessarily electronic), to have national ID numbers (in order to have a key by which to identify patients, and other policies. 3. Electricity and internet connectivity are an important step toward infrastructure that allows for the implementation of electronic health records and other health information systems. This mean that a stable electrical system must be implemented in Haiti. A development like this will not only be useful for developing an electronic health records system and other ICTs, but also for reviving the general economy in Haiti. 5- Conclusion Financial, political, and infrastructure problems abound presently in Haiti. Although financial mechanisms exist within Haiti for government taxation, sustainable business models exist in urban areas, funding for an EHR program would be unlikely from such mechanisms. With the amount of international aid in Haiti, and existing programs of level-2 implementation, partnerships between institutions in foreign countries and Haiti could readily be formed. The sustainability of such programs, though, is questionable. We propose that entrance into Haiti be a collaborative effort. It will require participation on behalf of international organisations, the Haitian government, and the local clinics and hospital. 5.1 Limitations Due to the limited data available on electronic medical records, it was difficult to find sufficient data of EMR implementations in each country. Particularly difficult was our original goal. We had initially planned on tracking systems of mobile electronic electronic health records with the intention of procuring a mobile solution to be implemented in Haiti. No scholarly articles were found relating to m- EHR in the context of these countries. Furthermore, we had initially set out to find ten sources for each country, for a total of one-hundred ten articles. The reality we encountered was that it was difficult to find even a few scholarly articles related to EHR in the selected countries, especially Madagascar. However, with fewer articles, we were able to better analyse each article. 6 5.2 Future Research As part of a future study, we would like to try to analyse implementations of mobile electronic health records in various countries, not just countries with similar HDR rankings to Haiti. That sort of study would find the factors that lead to the successful implementation of an m-EHR system. That, in conjunction with this study could lead to the possible implementation of a successful m-EHR system in Haiti. 6. References Au, L. (2009). Assessing the potential needs for telehealth in Papua New Guinea. Canterbury: University of Canterbury. Bagayoko, C, Miiller, H., & Geissbuhler, A. (2006). Assessment of internet-based telemedicine in Africa (the RAFT project). Comput Med Imaging Graph , 30, 407-416. Bitzer, J. (2004). Health Information System in Southern Sudan. Stuttgart: Diakonie Emergency Aid. Blaya, J. , Fraser, H., & Holt, B. (2010). E-Health Technologies Show Promise in Developing Countries. Health Affairs, 29(2), 244-251. Boslaugh, S. (2007). Secondary data sources for public health: a practical guide. Cambridge: Cambridge University Press. Cimino, J. J., Teich, J.M., Patel, V.L., & Zhang, J. (1999). What's wrong with EMR. Proceedings of 1999 American Medical Informatics Association Annual Meeting. Deleon, P. & Resnick-Terry, P. (1999). Comparative Policy Analysis: Deja vu All Over Again?. Journal of Comparative Policy Analysis, 1(1), 9-22. Evans, D.B., & Murray, C.J.L. (2001). Comparative efficiency of national health systems: cross national econometric analysis. British Medical Journal, 323, 307-310. Fraser, H., Biondich, P., Moodley, D., Choi, S., Mamlin, B., & Szolovits, P. (2005). Implementing electronic medical record systems in developing countries. Informatics in Primary Care, 13, 83-95. Fraser, H., Jazayeri, D., Nevil, P., Karocaoglu, Y., Farmer, P., Lyon, E., et al. (2004). An information system and medical record to support HIV treatment in rural Haiti. British Medical Journal, 329, 1142- 1146. Graham, E., Flynn, P., Cooke, S., & Patterson, V. (2001). The interdisciplinary management of cerebral haemorrhage using telemedicine - a case report from Nepal. Journal of Telemedicine and Telecare, 7 (5), 304-306. Iluyemi, A., Briggs, J., & Fitch, T. (2007). Electronic Health Records in Developing Countries, Integrating with Mobile Technology and Legacy Systems for Community Based Health Workers: Organisational and End-Users' Issues. In ECIME 2007: The European Conference on Information Management and Evaluation. Montpelier, France: Academic Conferences Limited. Kamadjeu, R, Tapang, E., & Moluh, R (2005). Designing and implementing an electronic health record system in primary care practice in sub-Saharan Africa: a case study from Cameroon. Informatics in Primary Care, 13, 179-186. Leichter, H.M. (1979). A comparative approach to policy analysis: Healthcare problems in four nations. Cambridge: Cambridge University Press. Mechael, P. (2009). The case for mhealth in developing countries. Innovations: Technology, Governance, Globalization, 4 (1), 103-118. Nessa, A., Al Ameen, M., Ullah, S., & Kwak, K. (2010). Applicability of telemedicine in Bangladesh: Current status and future prospects. The Arab Journal of Information Technology, 7(2), 138-145. 7 Omary, Z., Lupiana, D., Mtenzi, F., & Wu, B. (2009). Analysis of the challenges affecting e- healthcare adoption in developing countries: A case study of Tanzania. First International Conference on Networked Digital Technologies (pp. 201-209). Ostrava, Czech Republic: IEEE. Siika, A., Rotich, J., Simiyu, C, Kigotho, E., Smith, F., Sidle, J., et al. (2005). An electronic medical record system for ambulatory care of HIV-infected patients in Kenya. International Journal of Medical Informatics, 74 (5), 345-355. Swinfen, R., & Swinfen, P. (2002). Low-cost telemedicine in the developing world. Journal of Telemedicine and Telecare, 8 (2), 63-65. Tierney, W.M., Rotich, J.K., Smith, F.E., Bii, J., Einterz, R.M., & Hannan, T.J. (2002) Crossing the "digital divide:" implementing an electronic medical record system in a rural Kenyan health center to support clinical care and research. In Proceedings from the American Medical Informatics Association, 792-795. San Antonio, TX: Hanley & Belfus. United Nations Development Programme. (2009). Human Development Report 2009. New York: UNDP. Walt, C, & Gilson, L. (1994). Reforming the health sector in developing countries: the central role of policy analysis. Health Policy and Planning, 9(4), 353- 370. Williams, F. (2009). The role of electronic medical record in nation care delivery, development: case study on Ghana. Columbia, Missouri: University of Missouri- Columbia. 8