Interim Technical Report

Final Technical Report – OASIS ProjectChris Seebregts, 11 April 2010

Final Technical ReportDeveloping Open Architecture, Standards and Information

Systems for Healthcare in Africa (OASIS Project)

IDRC Grant Number 104508-001

Reporting Period: 1 October 2008 – 31 December 2009

Report Date: 11 April 2010

Research InstitutionSouth African Medical Research Council

Research CountriesSouth Africa, Mozambique, Zimbabwe, Rwanda

Research TeamName Title

Chris Seebregts Senior Manager, Biomedical Informatics Research Division, eHealth Research and Innovation Platform, Medical Research Council, Cape Town, South Africa

Jose Leopoldo Nhampossa Registrar and Senior Lecturer, Department of Computer Science, Universidade Eduardo Mondlane, Maputo, Mozambique

Tunga Simbini Department of Health Informatics, University of Zimbabwe, Harare, Zimbabwe

Deshen Moodley Head, Department of Computer Science, University of KwaZulu-Natal, South Africa

Carl Fourie General Manager, Jembi, Cape Town South Africa

Andrew Kanter Director: Medical Informatics, Millennium Villages Project, Columbia University, New York, USA

1


Table of ContentsResearch Institution......................................................................................................... 1

Research Countries........................................................................................................... 1

Research Team................................................................................................................... 1

Table of Contents............................................................................................................... 2

Synthesis.............................................................................................................................. 5Outcomes....................................................................................................................................... 5

Capacity Development...........................................................................................................................5Open Source Software............................................................................................................................6Open Architecture....................................................................................................................................6Sustainability..............................................................................................................................................6

Strategic Lessons......................................................................................................................... 6Local Capacity Development...............................................................................................................7Alignment with In-Country Government Initiatives.................................................................7Importance of Open Architecture.....................................................................................................8

Research Problem............................................................................................................. 8

Research Methodology.................................................................................................... 9Action Research........................................................................................................................... 9Outcome Mapping.................................................................................................................... 10Implementation Science........................................................................................................ 10

Research Findings.......................................................................................................... 10Objective 1. Support and Expand the OpenMRS Implementers Network.............10

2008 Implementers Meeting............................................................................................................112009 Implementers Meeting............................................................................................................11OASIS Meetings.......................................................................................................................................14Reinforce Existing Implementations in South Africa, Mozambique and Zimbabwe14Adoption of OpenMRS and Open Technologies........................................................................18

Objective 2. Develop the OpenMRS Internship Program............................................19Objective 3. Interoperability and Data Integration......................................................20Objective Four. Data Integration and Open Architectures........................................20Objective 5. Developing a Sustainability Model.............................................................21Objective 6. Promotion...........................................................................................................21

Project Implementation and Management.............................................................21

Project Outputs and Dissemination..........................................................................21OpenMRS Implementers Meetings.................................................................................................21Publications..............................................................................................................................................22Conference Presentations..................................................................................................................22

Capacity Building............................................................................................................ 23Institutional reinforcement...............................................................................................................23Increased Research and Administrative Skills.........................................................................23Capacity Development of Women and Marginalized Groups.............................................24

Impact................................................................................................................................. 24

2


Recommendations.......................................................................................................... 24

Appendices........................................................................................................................ 25Appendix One – Members of the OASIS Project..............................................................25Appendix Two – Jembi Status Report Q2, 2009..............................................................26Appendix Three – Jembi Status Report Q3, 2009..........................................................26Appendix Four – Jembi Status Report Q1, 2010.............................................................26Appendix Five – Mozambique OASIS Final technical Report.....................................26Appendix Six – MoU Between MoH Mozambique and M-OASIS................................26Appendix Seven – Zimbabwe OASIS Final Technical report......................................26Appendix Eight – Nairobi MVP-OASIS Research Meeting Report.............................26

3


Abbreviations

Abbreviation Description

CCD Continuity of Care Document

CCR Continuity of Care Record

CDA Clinical Document Architecture

DHIS District Health Information System

EMR Electronic Medical Record

FOSSIL Free and Open Source Software and System Interoperability Laboratory

GSoC Google Summer of Code

HISP Health Information Systems Program

IXF Indicator Exchange Format

MoH Ministry of Health

NGO Non-governmental Organizatiion

OASIS Open Architectures, Standards and Information Systems

OIP OpenMRS Internship Program

SDMX Statistical Data and Metadata Exchange

UEM Universidade Eduardo Mondlane

UKZN University of KwaZulu-Natal

UZ University of Zimbabwe

4


SynthesisOpen technologies have significant potential for improving the adoption and use of eHealth and health information systems in developing countries. In the Open Architectures, Standards and Information Systems (OASIS project) we evolved a general model for health information systems strengthening involving the following three actions: (i) initiate in–country capacity development and training programs focusing on open technologies applied to health; (ii) initiate health information system development projects in partnership with the Ministry of Health and other donors, universities and NGOs working in-country using local capacity, and; (iii) create a sustainable working environment in which appropriate projects are supported and maintained by local capacity. The fundamental research question was to evaluate the potential to build capacity and improve access to eHealth and strengthen health information systems in Africa using open technologies.

In meeting our overall objectives and model, stated above, we have established health informatics (OASIS) nodes – centered around free and open source software and system interoperability laboratory (FOSSIL) nodes - at three African universities in South Africa, Mozambique and Zimbabwe. We also established a collaboration with a training and capacity development program in a fourth country (Rwanda). The approach adopted and tested in this project involves addressing different actors and levels within the respective health information systems and measuring effect based on changes of behavior at various levels. We have developed partnerships between OASIS nodes and the respective country Ministries of Health as well as local universities, NGOs, private and donor-funded health information system projects. In addition, we have strengthened access to open source health software in these countries by providing training, workshops and an internship program. We have received recognition and endorsement from the Ministries of Health in all four countries and have also developed a regional network and developed links between different applications and linked this to other regional networks of open source

Outcomes

Capacity DevelopmentThe OASIS project demonstrated that African capacity can be developed by means of community-driven open source projects and that this contribution to development should be considered when software choices are evaluated in a health setting. Particular outcomes include

1. South African OASIS node developers have excelled in OpenMRS development and are recognized by the community, eg Carl Fourie, the original OASIS recruit has graduated to become the General Manager of Jembi Health Systems (below) and also the head of the two FOSSIL developer labs (Cape Town and Durban). Ryan Crichton one of the original two OASIS recruits from UKZN has become an international expert in SDMX-HD, an emerging data aggregation standard, developed by

5


the WHO. Daniel Futerman was selected as a mentor for the 2009 Google Summer of Code and Pascal Brandt will be starting a project with the core developer group, programming the OpenMRS logic service. As a whole, the group has been developing code under contract to other members of the OpenMRS consortium, particularly in the are of functionality pronting interoperability between OpenMRS and other open source applications.

2. The Mozambican OASIS node developers have become recognized as the seat of health informatics expertise in Mozambique. The Ministry of Health has recognized this achievement by entering into a Memorandum of Understanding with the M-OASIS group to continue providing the informatics requirements for the Ministry.

Open Source Software1. The OASIS project has demonstrated the importance both of the

community aspect of open source software development and also that networking and meetings are an important component of creating community-driven open source software. Through the OASIS project, IDRC has supported the OpenMRS Implementers meeting for the past four years. This has played a significant role in making OpenMRS, probably the most important example of community-driven open source software in the health domain.

Open Architecture1. The OASIS project has highlighted the fundamental importance of

open architecture in strengthening health information systems in developing countries. The open architectures component of the OASIS project has raised awareness of the importance of architecture at a global level and has also become a clarion call for broader work from IDRC and other funders on Enterprise Architecture in eHealth. In particular, the project has resulted in significant co-funding from Rockefeller Foundation and also resulted in the Health Informatics Public Private Partnership (HI-PPP) project, funded by PEPFAR, which also includes an important donor harmonization component.

Sustainability1. The OASIS project has provided some positive examples of how the

difficult problem of sustainability might be addressed the future. One of the outputs from the OASIS project is Jembi Health Systems, a non-profit company registered in South Africa that has taken over much of the implementation work started by the OASIS project. Jembi is presently has grown from strength to strength, largely on the basis of the capacity and projects developed as part of the OASIS project and is now adopting a leadership position in the application of enterprise architecture at a global level. Several of the future projects will be channeled through Jembi, testing and proving its strategic model and contributing to long-term sustainability.

Strategic LessonsThe following main strategic lessons resulted from this project:

6


Local Capacity Development1. Perhaps the single most important lesson learned was re-emphasizing

the importance of local capacity development for several reasons:

a. Solutions that do not involve long-term local capacity development are not sustainable and, at best, not useful and possibly even harmful to the system.

b. Building capacity is a complex undertaking and we have learned that an interdependent ecosystem approach needs to be followed where educational opportunities are developed alongside career opportunities providing opportunities and challenging projects. Strengthening one component in isolation can be negative as when trained people contribute to the brain drain.

c. The link to universities is critical and the ability to have inputs into the training process and then also attract students into a direction that is attractive (open source, challenging, moral etc) and provide projects with a small stipend is very productive. On the other hand, offering stipends to students in Africa without sufficient support is not successful.

d. We have learned that, in most African settings, it is not sufficient to think only of one objective in isolation, eg improving health outcomes. Interventions also need to simultaneously address wider objectives, such as unemployment, poverty and development.

Alignment with In-Country Government Initiatives1. It is critically important both to remain aligned with government

initiatives and to ensure that one’s actions are strengthening the in-country systems. Usually, this requires buy-in at the highest level of the Ministry. In Mozambique, M-OASIS was the only organization that survived a purge by the Minister of Health and the incoming Director of Health Information Systems when it was found that many NGO’s and projects were not aligned with Ministry policy. In Zimbabwe, OASIS was forced to reduce its activities as the government started appropriating foreign funds from the central bank. Yet, with the new power-sharing government, it is now assisting again with strategic projects, in-country. Alignment with the MoH has also contributed to the OASIS nodes being sought by other partners, including the CDC and OGAC/PEPFAR as well as universities, such as Vanderbilt, Columbia and the University of Washington and NGO’s for collaborations in-country.

2. It is important to develop and nurture strategic connections to further the work of OASIS beyond its current scope (ie the PIH training in Rwanda, working with OASIS-RHEIN, working with MVP and others to explore a network of networks – OASIS II – in this field.

7


Importance of Open Architecture1. Through the OASIS project, we have learned the importance of an open

enterprise architecture framework for developing countries. There is a window of opportunity to try and leapfrog the painful lessons learned in other countries by not paying enough attention to architecture and critical components such as standards and interoperability. If this fails, developing countries like those in Africa will fail to realize the benefits of eHealth.

In the first year of the project, we focused our effort on developing and implementing selected open software applications in public health systems and also developed systems promoting interoperability between open source health software applications important within an African context. In the second year of the project, we initiated a successful NGO to create long term sustainability for the actions initiated in this project and developed a strategy to adopt an open enterprise architecture approach to health systems strengthening which is proving highly popular to other groups.

While there often are significant political and technical challenges, we have found substantial acceptance of the nodes by respective Ministries of Health and computer science departments within universities. The OASIS nodes and the model have proven successful in all four countries, interestingly, often with different characteristics, and to different degrees. Predictably, we have found that political considerations are the single most important determinant of OASIS node adoption. And success.

In conclusion, the project has met all of its original objectives and has served as a platform for the development of a number of ongoing projects to elaborate the basic concepts, including the OASIS II project, funded by IDRC. It is likely that the actions initiated in the OASIS project will live on and will be recognized for the catalytic and pivotal role they have played in the development of these larger initiatives that have potential for substantial impact on health systems strengthening in developing countries, particularly Africa.

Research ProblemThe main research problem we addressed was: “what are the factors that can contribute to the successful and long-term sustainability of health information systems in Africa, particularly using open technologies?” In order to address this research problem, the project had six main objectives:

1. Continue to support and expand the OpenMRS implementers network, reinforce existing implementations in South Africa, Mozambique and Zimbabwe and investigate the principles underlying cooperative open source software development, implementation and maintenance in developing countries;

2. Build the capacity of African developers through a `Southern Summer of Code’ focused on OpenMRS and evaluate the potential for growing the

8


OpenMRS developers network to include open source developers from low- and medium-income African;

3. Better understand the challenges and opportunities for systems integration at the application level by examining other health information systems currently being deployed in other developing countries;

4. Evaluate data integration, open data access models and data sharing across heterogeneous health information systems and the challenges in developing integrated and aggregated information systems;

5. Explore sustainable models for supporting OpenMRS and other open source health software applications in a commercial competitive market;

6. Integrate and evaluate all the lessons learned and formulate best practices, policy briefs and scientific/technical publications elaborating the application of open source software in healthcare and disseminate the findings of the research.

Research MethodologyThe OASIS project used action research and outcome mapping methodologies as part of its core research methodology.

Action ResearchAs part of the action research (AR) methodology, we adopted an iterative approach to development with regular meetings and feedback sessions to reflect on the findings and identify what was working and what was not working. In general, this approach was very effective and enabled us to evolve a methodology (described above) that is now fundamental to the way we approach in-country HIS development. We also note that the methodology we use is different in many respects from that used by many other projects that often are less successful than we have been. A notable example is the Health Information Systems Project (HISP) that develops and maintains the District Health information System (DHIS) software and with whom we have a close working relationship.

We also found that the iterative approach inherent in the AR approach is closely analogous to the best practice way that software is developed (agile development). Both methodologies promote a close relationship between the researcher/developer and the subject/user and an iterative approach with joint reflection in order to converge on a solution.

One of the key influences that have resulted from the application of the AR approach in the OASIS project has been the realization that the bottom-up approach of experiential learning that is inherent in this approach is fundamental to understanding and developing useful systems in a developing country context. While many people talk about involving representatives from developing countries when solutions targeting developing countries are designed, it is probably not enough to simply involve these people. One needs to go one step further and ensure that developing country representatives are inherently involved as partners in each step, and preferably in a leadership role.

9


Outcome MappingWe also incorporate an outcome mapping approach into the general OASIS research methodology and found it to be particularly useful in this context. In particular, we found that focusing on outcomes such as influencing behavioral change in boundary partners to be a highly useful way to understand the difficult question of impact. For example, when we started the OASIS project in Mozambique, the Ministry of Health had adopted a policy of not supporting relatively highly-paid external consultants and, instead, was trying to develop capacity in HIS internally within the Ministry. However, this is unsustainable strategy because system development is not the core business of the Ministry and it cannot provide the kind of environment that is conducive to retaining computer and systems development professionals. In addition, it also pays low wages that are uncompetitive with regards to the private sector so it struggles to keep staff. M-OASIS demonstrated that an alternative approach could be successful, i.e. an independent, non-profit, Mozambican organization that can develop capacity to build and support systems for the Ministry. This has resulted in a fundamental change in thinking within the Ministry, right up to the level of the Minister of Health who has personally signed the MoU entrenching the relationship with M-OASIS. Impact is difficult to measure in these settings but this proxy of impact, defined by the OM methodology has given us a useful indicator of impact.

Implementation Science“Implementation research is the scientific study of methods to promote the systematic uptake of research findings and other evidence-based practices into routine practice, and, hence, to improve the quality and effectiveness of health services. It includes the study of influences on healthcare professional and organisational behaviour”1. In future iterations of the OASIS project and its spinoff projects, we will likely look deeper into this newer field for methodologies to assist with ongoing endeavor to incorporate operational research and evaluation into development and implementation projects so that it becomes part of the fundamental methodology.

Research Findings

Objective 1. Support and Expand the OpenMRS Implementers NetworkOpenMRS has been the flagship open source software application supported and elaborated through the OASIS project. Two OpenMRS Annual Implementers meetings were supported directly by the OASIS project, one in 2008 and one in 2009. Altogether, four meetings have been held and have contributed significantly to the growth of the OpenMRS community. Nurturing this community has contributed to a stronger and more active southern contingent, working hand-in-hand with northern and other southern developers, implementers and users. It has also become apparent that community-

1 Martin P Eccles and Brian S Mittman. Welcome to Implementation Science. Implemtation Science. 2006; 1:1.

10


development is almost certainly key to the long-term success and sustainability of any open source project, as elegantly demonstrated in the OASIS project.

2008 Implementers MeetingThe 2008 OpenMRS Implementers meeting was held in association with the biannual Health Informatics in Africa (HISA) conference, hosted by the South African Health Informatics Association (SAHIA) in Durban. In this meeting, we tried out a new concept of holding the meeting in association with another meeting. In addition, we had a room devoted to a developer hackathon and interoperability among open source projects. In general, the concept of holding the meeting in association with HISA was not successful but the developer hackathon was successful and has influenced future meetings.

There was a significant cultural difference between the HISA program and community and the OpenMRS program and community. The HISA program is highly structured with a strong academic focus, based on traditional research methods that requires the submission and evaluation of papers long in advance. OpenMRS, on the other hand, embraces an unconferencing style that evolves the agenda in response to attendee interests and needs. It also rather emphasizes experiential learning and conclusions. The original idea was to have some complementarity which, to some extent, did work, although it was sometimes difficult to adopt the correct format for a particular session. It was also the case that the OpenMRS leadership group felt that the community is large enough to warrant a dedicated conference and, while the conference with HISA was useful, it should not replace the dedicated Implementers meeting. Nevertheless, the conference was successful although not as successful as the 2009 conference that benefitted from the lessons learned in 2008.

2009 Implementers MeetingThe 2009 OpenMRS Implementers Meeting was held as a standalone meeting at Monkey Valley Beach Resort in Cape Town and was arguably the most successful meeting to date. The meetings have generally followed a progression in the application of an un-conferencing style and the 2009 meeting began with the setting of an agenda with evaluation and daily adjustment. In addition, there was more focus on development issues.

11


Figure 1. 4th OpenMRS Implementers Meeting, September 2009, Cape Town, South Africa

The OpenMRS Implementers meetings have made a definitive contribution to the growth of OpenMRS, particularly the community-driven aspect. Following an expansive phase of increasing number attending the OpenMRS Implementers meeting, we have begun to restrict the numbers attending the meeting. At the moment, the optimal number is probably between 120 and 140 that has been possible with the existing levels of funding and conferencing facilities. If the future demand for the meeting continues to increase, we will need to investigate additional funding mechanisms and either a different style of meeting or consider having two meetings, focusing on different constituencies.

12


Figure 2. OpenMRS Implementers Meeting Attendees from 2006 to 2009.

1 2 3 40

50

100

150

200

250

67

167

200

135

OpenMRS Meeting Attendees

Attendees

Years 2006 - 2009

Num

ber

Perhaps the closest other model to OpenMRS is the District Health Information System (DHIS) / Health Information Systems Program (HISP) model. The DHIS is largely developed and maintained by postgraduate students from the University of Oslo and the HISP programs in South Africa and India. In contrast, OpenMRS is substantially community-driven and comprises a very large network of developers and implementers across the world, evidenced by the fact that OpenMRS has the highest number of applicants registered in the annual Google Summer of Code (GSoC) program. Community-driven is a key feature of mature open source software projects that leads to long-term sustainability and growth of the software. The OpenMRS Implementers meetings have played a significant role in the growth of the OpenMRS community by providing an annual event at which people are able to meet and an occasion for new developers and implementers to become part of the community.

13


Figure 3. Known OpenMRS Implementations

OASIS MeetingsTwo OASIS meetings were also held during this period and focused mainly on research methodology and implementation. The meetings build directly on the interim results of this project. The results of these meetings have been reported separately in the introductory sections of the OASIS II funding proposal in which the research question was posed2:

Reinforce Existing Implementations in South Africa, Mozambique and ZimbabweA central objective of the OASIS project was to develop centers of capacity and competency in open source health information systems development and implementation and we have successfully established OASIS nodes at three Southern African universities, University of KwaZulu-Natal (UKZN), South Africa; Universidade Eduardo Mondlane (UEM), Mozambique and University of Zimbabwe (UZ). We are also collaborating with a training program supported by Partners in health (PIH) and a new eHealth Center of Excellence in Rwanda.

South Africa OASIS Node

The original South African OASIS node was started at the South African Medical Research Council (MRC), which also acted as the host institution for the Project Lead, Dr Chris Seebregts. This node was responsible for overall project coordination across the three nodes, and for providing technical and project

2 Open Architecture Standards and Information Systems (OASIS) II – Developing Capacity, Sharing Knowledge and Good Principles Across eHealth Projects in Africa. Proposal to IDRC. August 2009.

14


leadership. This model was accepted among the nodes and the South African node was successful at coordinating and administering the various implementation, research and administrative aspects of the project. Although these functions were housed at the MRC, the main South African node for capacity development was instituted at the School of Computer Science at the University of KwaZulu-Natal, where a significant concentration of students exists and the project lead has an appointment as an Honorary Associate Professor. Significant success was achieved in attracting and absorbing graduates from UKZN-SCS and four graduates were absorbed into the OASIS project during the course of the project.

The South African node was also responsible for starting a separate not-for-profit company (Jembi Health Systems), which falls within the fifth objective of this project, and transitioning the project coordination and OASIS node functions into Jembi. The first three quarterly reports are attached as Appendices one to three.

At present, Jembi has two developers based at the OASIS node at UKZN-SCS in Durban and two at its offices in Cape Town. Jembi has also appointed its first trainee from the University of Cape Town and hosts an epidemiologist, seconded from CDC Columbia. The growth of this new organization illustrates the respect it has from established donors and researchers in the field of HIS in southern Africa. The current organogram is shown below.

Jembi Organogram

Achievements of the South Africa OASIS over this period include the following:

1. The group hosted two OpenMRS Implementers meeting, including the first two OpenMRS hackathons in developing countries. The hackathons have played a significant role in growing the implementers community in Africa.

15


2. Carl Fourie, one of the original two UKZN graduates, has been appointed as the General Manager of Jembi Health Systems and is coordinating many of the OASIS South Africa node projects.

3. Ryan Crichton, one of the original two people recruited to Jembi has become the main developer on the SDMX-HD standard, a new emerging standard developed by the WHO for managing aggregate health data.

4. Daniel Futerman was the first OASIS South Africa developer to be invited to act as a mentor on the OpenMRS Google Summer of Code project where he participated in the supervision and development of the core HTML forms module

5. Pascal Brandt has developed core reporting modules for the Partners in Health group and has been invited to spend time with the core OpenMRS developer team in Indianapolis to work on the logic service module as part of the core of OpenMRS.

6. Wayne Naidoo, the latest recruit from UKZN, graduated summa cum Laude for his Masters degree in Computer Science and has completed several successful projects for the node.

Mozambique OASIS Node

The Mozambique OASIS node (M-OASIS) is located at the Universidade Eduardo Mondlane and currently has nine staff, including a country Principal Investigator, a country Director, a fulltime project coordinator, three fulltime developers, three part-time developers and an implementer.

M-OASIS Organogram

Although the South African OASIS node is the most accomplished in terms of software development and OpenMRS implementations, the Mozambique node is the most well-developed node in terms of general health information systems implementation. The M-OASIS group has been involved in a significant number of different projects in Mozambique, e.g. the development, implementation of a mortality system, a review of EMR applications, selection of hospital information system for the Maputo Central Hospital and maintenance of Modulo Basico, the main national data warehouse supporting health management information

16


systems in Mozambique. The M-OASIS Country Report is attached as Appendix Four and a copy of the MoU with the Ministry of Health as Appendix Five. Highlights of the M-OASIS node include the following:

1. The M-OASIS node concluded a Memorandum of Understanding with the Ministry of Health in terms of which M-OASIS will provide ongoing informatics support for the Ministry and specific projects for an initial period of five years (optionally renewable).

2. The M-OASIS node has implemented a mortality register for Mozambique that has been approved by the Ministry of Health. The application is currently being implemented in ten hospitals across the country in all provinces and plans are underway to elaborate it as a general register for recording deaths in Mozambique. A timeline and schedule has been developed and included in the combined strategic plans of M-OASIS and the Department of Information Systems (DIS) of the Ministry of Health.

3. The node successfully completed an EMR documentation project under contract to the University of California at San Francisco in which structured information was collected concerning different EMR applications deployed in Mozambique and entered into a database.

4. The node has completed a strategic planning exercise with the Ministry of Health and prioritized a number of projects, including the development of en enterprise architecture for Mozambique.

Zimbabwe OASIS Node

The Zimbabwe node is located within the Center for Public Health Informatics (CEPHI) at the University of Zimbabwe and comprises one node manager and two developers. The node has experienced significant difficulties as a result of the political turmoil in Zimbabwe. The implementation of an OpenMRS HIV/AIDS application at Chitungwiza Hospital had to be discontinued as the Opportunistic Infection Clinic was mobilized to address the cholera outbreak in Zimbabwe. However, we have recently resurrected the OASIS node in Zimbabwe and, with new partners, including Pangaea Global AIDS Foundation (PGAF), the World Wide AIDS Coalition (WWAC) and InSTEDD as part of the global HI-PPP initiative, and have entered into discussions with the Ministry of Health to begin an enterprise architecture project in Zimbabwe. The Zimbabwe country report is attached as Appendix Six.

Rwanda OASIS Node

The Rwanda node was added to the project because of opportunities presented during the project. We participated in the planning of the RHEIN (Rwanda Health and Education Information Network) project and have also begun a large enterprise architecture project in collaboration with the Ministry of Health. In fact, the RHEIN project is called OASIS-RHEIN, to indicate it is a satellite member of the OASIS work. The OASIS node in Rwanda is also working closely with another project funded by IDRC, the Partners in Health Rwanda Training Program is training Rwandans in computer software development using OpenMRS as an example.

17


MVP-OASIS Project

As part of a build up to the OASIS II project, OASIS held joint discussions and events with the Millennium Villages Project (MVP). The alignment with MVP was not part of the original proposal and was added on when an opportunity was identified to proactively look at areas of convergence between MVP and OASIS that have highly similar goals related to health systems strengthening in developing countries. A central ethic that has evolved during the OASIS project is to attempt to reduce unnecessary variation which results in an exponential increase in costly complexity and which often makes systems difficult to understand and inherently fragile. Other advantages include economies of scale and convergence of best practices. The convergence of tools and techniques is also part of the thinking for the open architectures project where implementation lessons are generalized into an architectural framework that can be used as a starting point for design in other settings and, with appropriate customization and localization, rapidly develop local solutions. A report of the research meeting held jointly in Nairobi is attached as Appendix Seven

Adoption of OpenMRS and Open TechnologiesAnother objective was to support the adoption and use of OpenMRS (www.openmrs.org) through the OASIS nodes serving both to promote the adoption of electronic medical records (EMR) in appropriate settings and also test the adoption of a flagship community-driven open source health software application (Objective One). Results have varied between nodes, largely reflecting differences in staff background, training programs and in-country political and organizational dynamics.

South Africa OASIS Node

The South African OASIS developers have readily adopted OpenMRS, are well integrated into the OpenMRS developer network and are contributing at a high level both to module and core OpenMRS development. The group is regarded as one of the Centers of Excellence in OpenMRS (the other two being at Makerere University in Uganda and a commercial eHealth company started by Joaquin Blaya and colleagues in Chile) and also has the greatest concentration of OpenMRS developers outside of the USA. The group has developed a version of OpenMRS for HIV treatment failure clinics in the Free State province and Gauteng and working on integrating OpenMRS with RegaDB. The group also has competencies in SDMX and developed the main Java library for SDMX-HD. With the move into Jembi, the developers are working mostly on in-house or contract projects, particularly for Partners in Health and, in future, for the OpenMRS Foundation.

Mozambique OASIS Node

The M-OASIS node has been slow to adopt OpenMRS as a development platform although the group does have one or two pilot projects using OpenMRS. Instead, the group has successfully developed several public health applications using a mixed proprietary-open source software model (Visual Basic and MySQL). It is likely that this reflects the relatively lower proficiency of these developers in Java programming as well as the requirement in Mozambique for relatively simpler public health applications, such as simple registers that are more

18

http://www.openmrs.org/


accessible to simpler visual programming languages and less appropriate for the added infrastructural cost of OpenMRS. However, our group collaborates with a group from Vanderbilt University that has developed and implemented an OpenMRS application in Zambezia province and there are proposals to link some of the M-OASIS tasks with this project.

The choice of software development environment (Visual Studio) in the M-OASIS node is not conducive to open source development although the code is still openly available. We would prefer to move to an open source development environment but this will likely take some time as most of the programmers tend to be more proficient in the relatively easier visual programming tools, like Visual Studio, at this stage. We will need to implement actions similar to that which PIH has implemented in Rwanda if we want to get more movement I that direction. While this represents another potential area of convergence, similar to that mentioned in the context of MVP, above, it was beyond the scope of the current project to implement fundamental Java training programs. However, this may be necessary in future.

However, the approach of allowing the use of Visual Studio for development also reflects the shifting thinking in terms of priorities of open technologies. While we still advocate open source software development, for a variety of reasons, the emphasis has shifted more towards open architectures and standards. It is probably useful to have several different software applications, both open source and proprietary in a particular environment as it stimulates competition and improves quality. However, it is important to have consensus and standardization at the level of the overall country architecture (even regionally) and the adoption and implementation of standards.

Zimbabwe OASIS Node

The OASIS group in Zimbabwe has struggled to keep programmers and a competency in OpenMRS. As a result of the lack of staff and difficulties with maintaining technical people, we have shifted the focus to Enterprise Architecture and are working with implementation partners, such as the DHIS, in order to build systems in Zimbabwe. It is likely we will have more of an impact at the design level where we can help Zimbabwe elaborate its national eHealth strategic plan and take better advantage of the investments in the country from foreign donors.

Objective 2. Develop the OpenMRS Internship ProgramIn Objective Two, we proposed to stimulate open source software development by adopting the successful Google Summer of Code (http://code.google.com/soc/) model and applying to developing countries. Unfortunately, the OIP has not been as successful as originally hoped. We originally planned to award up to twenty stipends over the term of the project (two years), but only managed to award four final payments. Many interns registered but failed to complete the project. It is most likely a result of interns trying to take on the OIP as a way to make extra money while still holding down an additional job. While useful as a capacity development program, it seems that the majority of the OIP and GSoC projects do not result in code that is useable

19

http://code.google.com/soc/


within the normal OpenMRS context. Another contributory factor is often the lack of a paid job after the internship program is done. As mentioned above, it is likely that sustainable capacity development requires more of an ecosystem, approach to capacity development that develops training and skills at the same time that projects and jobs are developed and infrastructure and mentorship are improved. One way that the OIP will succeed is by recruiting students to work during the vacation while attached to a FOSSIL lab where they can receive mentorship and also understand the future growth possibilities.

Objective 3. Interoperability and Data IntegrationInteroperability and data integration have been critical success areas for the OASIS group and successful collaborations have been developed between OpenMRS and several other open source software developer groups. The most notable achievement has been with the DHIS group. OpenMRS has been integrated with version one and version two of the DHIS. The latter integration also required the development of an SDMX-HD Java library that has also been used to develop data integration between OpenMRS and TRACnet in Rwanda. Other integrations include the integration of OpenMRS with a pharmacy management application (iDART) and with mobile phone applications (JavaROSA). The OpenMRS DHIS 2 integration has been elaborated in a project in Sierra Leone and the TRACnet integration will be used in Rwanda as part of the large-scale OpenMRS implementation.

Objective Four. Data Integration and Open ArchitecturesThe development of open architecture has been the most significant development in the second year of the OASIS project. Much awareness has been created around the need for a design phase in the development of health information systems. Enterprise architecture addresses this need and also promotes the development of frameworks that can be customized and reused. The OASIS group has played a pioneering role in promoting an open architectural framework for developing country health information systems along with several other international institutions. Open architectures will play a much more significant role in the second phase of the OASIS project. The outcomes of this objective have formed the cornerstone of the next wave of projects, including the OASIS II project, and several other donor-funded HIS initiatives that are striving to strengthen health systems in-country. Much of the critical point of departure in these projects has been informed directly by the OASIS project, such as the following:

1. The importance of building an architectural framework as an implementation-driven activity (Action Research methodology).

2. Targeting behavior change in influential boundary partners, Minister of Health, who can make the real changes we seek (Outcome Mapping approach)

3. Open architectures and standards the two most fundamentally important open concepts in addition to open source software (Open Architectures project).

4. Simplicity and only building just enough to get to the next stage and then iteratively improving with much reflection to converge on an optimal

20


state is the preferred way to develop systems and software in this environment (Agile Development and Action Research, combined).

Objective 5. Developing a Sustainability ModelSustainability is fundamentally important to successful health information systems in developing countries. In Objective 5, we proposed to explore models to improve sustainability, particularly the formation of a non-governmental organization (NGO) to carry out implementation work in this area. In our previous interim report, we reported on the adoption of HealthNet South African and formation of Jembi as a registered no-profit company in South Africa. Jembi has grown significantly since inception and is serving as a role model for similar institutions that have grown up around OpenMRS.

Objective 6. PromotionThe sixth objective was to promote the lessons learned through appropriate channels. During this project, OASIS collaborators delivered a number of posters and presentations at international conferences and published one peer-reviewed publication (see Project Outputs and Dissemination, below).

Project Implementation and ManagementThe project has been managed through the South African Medical Research Council (MRC). The MRC has robust grant and financial administration systems, which, in general, has provided an excellent infrastructure for managing the project and funds. However, some challenges have been experienced in terms of international transactions. Despite best efforts by MRC administrative staff, it has taken substantial effort and time to transfer funds to organizations in other countries. This is to be expected since the MRC systems are primarily focused on obtaining donor income and expending against projects locally in South Africa. As a semi-government organization, the MRC is also subject to the Public Finance and Management Act and the consequent financial controls. This problem has largely been solved by forming Jembi that has a foreign exchange account that is able to efficiently perform foreign exchange payments.

Project Outputs and Dissemination

OpenMRS Implementers Meetings We held the 3rd Annual OpenMRS Implementers meeting in conjunction

with the HISA (Health Informatics in Africa) 2008 conference in July 2008. The 4th Annual OpenMRS Implementers meeting will take place during

September 2009.

Publications1) CJ Seebregts, BW Mamlin, PG Biondich, HSF Fraser, BA Wolfe,

DJazayeri, C Allen, J Miranda, A Kanter, N Lesh, E Baker, C Fourie, Y Singh, C Yiannoutsos, C Bailey & OpenMRS Implementers Group. The

21


OpenMRS Implementers Network. Int. J. Med. Inform. (2009), doi:10.1016/j.ijmedinf.2008.09.005

2) Mars M and Seebregts C. Review of e-Health in South Africa. 2008. Country-level report for the Rockefeller Foundation and the Bellagio series on e-Health. July/August 2008.

Conference Presentations1) Seebregts C., CFAR Asangansi I, Hanmer L, Kanter A. Building an

Integration Framework for eHealth. Health Informatics in Africa Conference. April 2009, Abidjan, Cote d’Ivoire.

2) Seebregts Spring AMIA. Open Enterprise Architecture for Health.3) Almeida E, Simbini T, Nhampossa JL, Fourie C, Manders E-J, Mars M

and Seebregts C. The OASIS Project: Developing Open Architecture Public Health Information Systems in South Africa, Mozambique and Zimbabwe. Poster presented at the Second Global Partners in Public Health Informatics, Seattle, September 2008.

4) Fraser HS, Seebregts C and Bailey C. Presentation. Implementing Integrated Systems with OpenMRS. Second Global Partners in Public Health Informatics. 18-19 September 2008, Seattle, USA.

5) Seebregts C. Chairperson. Workshop on Developing Architectural Frameworks for Developing Countries. Seattle, USA. 18 September 2008

6) Seebregts C. Organiser. OASIS and ICOHRTA workshop on integrated health systems. Harare, Zimbabwe. 7-9 July 2008.

7) Seebregts C. Organiser. OpenMRS Implementers Meeting and Health Informatics in South Africa 2008. Durban, South Africa, 17-20 June 2008.

8) Seebregts C. Organiser and Chair. OpenROSA partners meeting. Durban, South Africa. 16 June 2008.

9) Seebregts C. Developing Integrated Country-Level Health Information Systems. Rwanda Health and Education Information Network Funding Workshop. Kigali, Rwanda, 29-31 April 2008.

10) Seebregts C. HIV/TB Data Integration with OpenMRS. Presentation at the Health Systems Development under NRHM: Meeting the Challenge of Integration Conference, New Delhi India, 26-28 April 2008.

11) Seebregts C. Achieving Interoperability of Health Management Information Systems in South Africa. Session Chair. Conference of the Health Management Information Systems Group of the Global AIDS Programme, Centers for Disease Control. Gordon’s Bay, South Africa. 14-16 April 2008.

12) Seebregts C. Health Outcomes: The Role of ICT Applications, Standards and Practices. Invited Panelist, Third Global Knowledge Partnership Conference (GK3), Kuala Lumpur, Malaysia; 13 December 2007.

13) Seebregts C. Mobile Forms Development and Standards. WHO International Data Standards Meeting, Nairobi, Kenya, 6 December 2007

22


14) Seebregts C. OpenMRS and Open Source Software for Monitoring HIV/AIDS. Invited Presenter, Appraisal and Action: HIV/AIDS in Southern Africa. University of Michigan Advanced HIV/AIDS Research Group, University of Michigan, Ann Arbor, USA, 29 November 2007.

15) Seebregts C. Developing Integrated Health Information Systems in Mozambique. Organiser and Funder of a two-day workshop on behalf of the Ministry of Health, Maputo, Mozambique; 6-7 November 2007.

16) Seebregts C. Integrated HIV and TB Information Management. Invited presentation to the Global AIDS Program (President’s Emergency Plan for AIDS Relief) Health Management Information Systems Meeting, Lusaka, Zambia; October 2007.

Capacity BuildingCapacity building is a central part and objective of the OASIS project. The model we developed is to first develop in country capacity and then use this capacity to strengthen in–country health information systems (HIS). Each of the four main OASIS nodes (two in South Africa and one each in Mozambique and Zimbabwe) has strengthened local capacity at various levels.

Institutional reinforcementThe project has strengthened the ability of the three institutions involved in the project. In South Africa, we have developed a node at UKZN within the department of computer science that will become the seed for a medical informatics unit. In Mozambique, we have developed an informatics implementation group at the UEM that is working on several projects for the Ministry of Health and is strengthening the ability to implement HIS. It is likely that this node will serve as an important model. The Zimbabwe node is growing slower than the other nodes but is being strengthened and will hopefully become more productive in the next year or two.

Increased Research and Administrative SkillsThis project has afforded the collaborators much opportunity to implement a research methodology associated with the implementation of health information systems. There has been training in action research and outcome mapping and have proposed a consolidated approach to research methodology in the OASIS II project. We will also continue to explore implementation science as a useful paradigm for ongoing research into the development of systems.

Capacity Development of Women and Marginalized GroupsThe project has developed significant capacity among marginalized groups, mainly racial. Both in Mozambique and Zimbabwe, the project is involved in development of groups of indigenous African people and providing opportunities for personal and skills development. The project has relatively more men, reflecting the traditional demographics of the software development industry. We did hire a gifted young developer, Ms Phumzile Khumalo. However she was eventually restricted in working with OASIS on account of the fact that she had

23


previously been funded by the CSIR during her studies and needed to complete in-service work to repay her bursary. However, we will continue to try to identify and recruit people who can help even out the gender and race imbalances within OASIS.

ImpactThe OASIS project has already begun to have an impact on the countries in which it has been operating and has even attracted interest at an international level. The M-OASIS node is the best example of impact, considering that the Ministry of Health has written M-OASIS into its strategic plan and acknowledges the fact that OASIS plays a strategically important role within the Ministry. We are planning to elaborate the success in Mozambique in other countries.

The Open Architectures aspect of the project is strategically placed to have more impact than the open source software projects and is the general strategy we are following, going forward.

RecommendationsTo date, this research project has resulted in the following recommendations:

1. We should continue to strengthen the OASIS nodes in terms of technical capability and ability to become sustainable. The open source model is potentially important in Africa as it contributes to lowering costs and improving sustainability. However, there is a critical need for more robust, enterprise level applications. Focusing on an open enterprise architectural level will likely have more impact than focusing on individual software applications and standards.

2. The OpenMRS Internship Program was not entirely successful and should be changed to follow a more ecosystem approach where students have the required infrastructure and mentoring in order to have a better chance of succeeding. However, several other capacity development initiatives have begun and we are working with some of them to find other ways to build capacity that can dovetail with some of the project work in of our program. In particular, the AMIA Global Partnership and Health Informatics Building Blocks programs are of particular interest in this area.

3. OASIS nodes should be harmonized with similar capacity and systems development concepts created by other groups such as the HISP and MVP. This concept is important for developing countries and it will be important to better understand the lessons learned and evolve best practices and, in addition, achieve economies of scale in terms of funding and also reduce unnecessary variation and complexity.

4. OpenMRS needs to grow in a more enterprise direction in order to cater for a more general set of needs in typical developing countries. This should include interoperability with other applications that can be used with OpenMRS to develop health systems as well as smaller applications

24


that can be rapidly deployed. The OpenMRS Foundation that has recently been established has plans to grow OpenMRS in this direction.

5. Open architecture and patient-driven systems are emerging as dominant themes that will drive health information systems, going forward. OASIS needs to remain actively involved in these two fields and has developed strategic plans in both directions. The bottom-up approach informed by implementation experiences should drive the development of the future architectural framework so it does not suffer from the problems inherent on other frameworks which are more like theoretical constructs than practical products.

6. We should continue trying to attract more women and Africans to assist with these projects, rather than the current white, male-dominated scenarios.

Appendices

Appendix One – Members of the OASIS Project

OASIS South Africa Node

Dr Chris Seebregts Program Director Jembi Health Systems Dr Alessandro Campione Country Director Jembi Health Systems Carl Fourie General Manager Jembi Health Systems Ryan Crichton Developer Jembi Health Systems Daniel Futerman Developer Jembi Health Systems Pascal Brandt Developer Jembi Health Systems

OASIS Mozambique Node

Prof Leopoldo Nhampossa Program Director M-OASIS Dr Antonio Sitoi Program Coordinator M-OASIS Claudio Rocha Developer M-OASIS Joao Machiano Developer M-OASIS Leonel Machava Developer M-OASIS

OASIS Zimbabwe Node

Dr Tunga Simbini Principal Investigator OASIS Zimbabwe

25


Appendix Two – Jembi Status Report Q2, 2009

Appendix Three – Jembi Status Report Q3, 2009

Appendix Four – Jembi Status Report Q1, 2010

Appendix Five – Mozambique OASIS Final technical Report

Appendix Six – MoU Between MoH Mozambique and M-OASIS

Appendix Seven – Zimbabwe OASIS Final Technical report

Appendix Eight – Nairobi MVP-OASIS Research Meeting Report

26

Interim Technical Report

Health & Medicine

ehealth research

research problem

research methodology

action research

research findings

rwanda research team

university of zimbabwe

open architectures