Report on the Global OpenMRS Community Meeting in Malawi - Towards Evidence Based Health Service Delivery and Interoperability

“eHealth to promote evidence based health service delivery and interoperability”

One hundred seventy five members of the worldwide OpenMRS community–representing 20 countries–met in Malawi this past December for the 2017 OpenMRS Implementers’ Conference. This event was the second consecutive year a national government sponsored this global meetup, with Uganda hosting and sponsoring this meeting the previous year. The December conference was hosted by Malawi’s Ministry of Health and key-noted by ministry officials and leaders such Maganizo Monawe, Senior HIS Technical Advisor; and Anthony Muyepa, Director General at National Commission for Science and Technology.

Attendees of the 2017 International OpenMRS Implementers’ Conference pose for a group photo. Credit-OpenMRS

The OpenMRS meeting in Uganda was such an extraordinary success that a large number of countries applied to host the 2017 conference. There were so many applications that the OpenMRS organization had to create a set of guidelines and process to make the decision. Malawi was chosen as the conference host as described in this article. 

This years’ theme focused on how eHealth could help improve delivery of health service at the patient level or point of care (support decision and data processes) and how eHealth could support planning for health service delivery. The key components within this scope are:

• eHealth: the ICT technologies
• Data use: how eHealth supports the collection and utilization of data at the patient level.
• Interoperability to strengthen data exchange between systems.

OpenMRS has helped revolutionise management of patients and patient level data at the point of care in low income countries through a range of tools in its architecture. Due to the program based approach that has been followed in improving their health systems, many countries now have large data sets in disparate electronic data sets. This has made it difficult to conduct cross analysis of related data to support decision making. OpenMRS has particularly been vital in facilitating health service delivery by incorporating standards and protocols to enable interoperability with other systems. The interoperability framework that OpenMRS has adopted is helping the delivery of health service at the patient level and decision making by making it easier to share data with these systems.

The objectives of the 2017 Implementers’ Conference were: to demonstrate how Malawi is building a national implementation of OpenMRS; to expand the OpenMRS community practices to support quality assurance processes; to identify ways to measure a “successful high quality implementation” of OpenMRS; and to understand how OpenMRS outputs can be used to positively impact clinical care.

alawi is one of the furthest destinations for an implementers’ conference, nevertheless large numbers of OpenMRS developers arrived from all over the world to assist and observe the early phase of Malawi’s Ministry of Health’s implementation of OpenMRS. The government of Malawi and the and Baobab Health Trust, a national organization, have been using an earlier version of OpenMRS as the backbone of their health system,

Founded in 2001 by Dr. Thuy Bui and Dr. Gerald P. Douglas as a partnership, and later transformed into a Malawian Trust in 2008, Baobab Health Trust (BHT) uses tech innovation to improve response to health crises in Malawi and other developing countries. BHT’s goals are to collaborate with government and healthcare workers to build, deploy, and maintain innovative, robust and sustainable healthcare information systems suitable for the developing world. Malawi is using OpenMRS to treat HIV and tuberculosis (TB).

OpenMRS Community leaders gathered prior to the start of the conference to discuss the 2017 operational plan and strategic goals, developed annually, and finalize any remaining details with the OMRS17 Conference Planning Committee. This leadership meeting was a unique experience for the OpenMRS community leaders to meet face-to-face and bypass the weekly challenges of internet bandwidth and timezone conversions. Some of the agenda items discussed were the OMRS17 conference, organizational structures and operations challenges, role of the BOD and its members, fundraising, and the management of Bahmni.

Photos of the event can be viewed and downloaded from our Google Drive.

Opening Ceremony

The 2017 OpenMRS Implementers’ Conference kicked off with the opening remarks and keynote presentation by Anthony Muyepa, Director, National Committee Science and Technology, and discussed why Malawi chose to use OpenMRS as their EHR platform; the key components of the OMRS meeting; how healthcare centers deliver services and make decisions on healthcare delivery; their infrastructure needs and commitment to improving their existing infrastructure to support facilities in rural communities depending on reliable internet and electricity.

After the keynote presentation, Burke Mamlin, co-founder of OpenMRS, voiced the unification of our software and community, and how reiterated how developers have the unique opportunity to save lives through writing code. In closing, Mamlin expressed the community’s gratitude for our Malawi community members for hosting this year’s conference

After the opening remarks it was time for the unconference planning to begin, in true OpenMRS style, by writing topic suggestions on colorful pages and asking the attending delegates to vote on the sessions they wanted to participate in. Year after year, participants rank the unconference planning session as one of their favorite experiences because they feel it makes the conference a very personal experience for them.

Lightning Talks

A lightning talk is a very short (5 minute maximum, strictly enforced) presentation. Conference attendees gave short presentations about active projects and also shared new ideas or told a story about their interaction and/or usage of OpenMRS. Lightning talks are often one of the favorite sessions at OpenMRS Conferences, since the audience gets to hear about many amazing projects or ideas in a short time, while watching presenters trying to beat the clock.

Lightning talk presentations covered many amazing projects and ideas, including Data Warehousing, new features in PIH EMR, What’s new in Bahmni 0.90, OpenSRP, REST- WS TODOs, HEARTH, FACES, and mUzima. Many of the topics presented on included how OpenMRS is being used or what types of features they’re working on based on the needs of their local healthcare facilities. Notes and presentation documents can be found on OpenMRS’ Google Drive.

Unconference Sessions

Individual sessions at the conference addressed a wide range of topics. Some of these session topics included:

• Cost-Benefit Measures for EMRs
• OpenMRS’ Use in Cancer Care
• Clinician Usability & Adherence
• Increasing Collaboration in the Web Application layer
• Point of Care (PoC) with OpenMRS
• OCL and Concept Management
• OpenHIE
• OpenMRS ETL & Data Warehousing
• Collaborating on OpenMRS Android Apps

These sessions showed the value of the unconference format; allowing bidirectional sharing of information between multiple participants. Below you’ll find more detailed information regarding the topics discussed and what the next steps are.

Cost-Benefit Measures for EMRs

Presented by Tyler Smith

A recent systematic review of eHealth interventions concluded that “some benefits in the quality of care” have been identified, however no “measurable impact on cost effectiveness in a hospital setting” has been found to date. Electronic medical record (EMR) systems being deployed in low-resource settings are thought to improve the quality of patient care and efficiency of service delivery, but little compelling evidence has been offered to support these claims. One reason identified is lack of data and standard system evaluation methods, particularly with respect to cost and efficiency. Further investment and expansion of EMRs will require better evidence describing their impact.

This session aimed to have participants:

1. Share their experiences with measuring system performance and impact in various country contexts
2. Outline metrics that should be used to measure cost and quality resulting from EMR introduction
3. Highlight challenges with data collection or interpretation and communicating findings
4. Propose methods and standards for performance and cost monitoring that would improve evidence of impact

OMRS in Cancer Care

Presented by Johnblack Kabukye

With impressive strides made in dealing with infectious diseases in developing countries, non-communicable diseases are are now becoming the leading public health challenge in low and middle income countries (LMICs). Cancer is now a major cause of mortality and morbidity, responsible for more deaths than HIV/AIDS, TB and Malaria combined. There is therefore a need to shift focus (or widen it) to cancer in terms of improvement in health systems, research, and importantly in harnessing health IT oncology. This was the motivation for this session, in addition to other discussions on the topic that have been going on over the past year.

The aim was to gather key people in the OpenMRS community who are interested in the topic to explore, brainstorm and discuss key issues for developing and improving oncology support in OpenMRS, including the following:

• Current use of OpenMRS in cancer care: Members shared experiences of current use of OpenMRS in cancer care in Haiti, Rwanda and Kenya. There was general agreement that better oncology support is needed, and that currently the system is being used for part of the cancer care timeline and for relatively simple tasks e.g. for screening clinic at AMPATH Oncology in Kenya, Initial encounter (intake) in Haiti, and for clinical notes in Rwanda. Some work on order sets is ongoing in Rwanda by PIH.

• The oncology workflow: we discussed some uniquenesses in the workflows involved in cancer care that might require special attention and functionality in OpenMRS such as
  • Multidisciplinary and multi-modality care – chemotherapy, radiotherapy, surgery, palliative care, cancer screening, etc which require improved coordination and have different information needs or documentation styles (different forms)
  • complexity of chemotherapy prescription and management i.e. several drugs in a protocol/combination, tightly controlled doses calculated according to body surface area and often adjusted for individual patients (e.g. in case of HIV positive cancer patients)
  • Oncology specific documentation such as tumor descriptions (size, location, metastases, laterality, as well as tracking of these over time), cancer staging (FIGO, AJCC, Risk stratification for Leukemias), Pathology reports and other special investigations
• Reporting and data sharing: we discussed the cancer registration process and the CanReg software currently in use internationally as well as looking at the form for data collection. Thoughts were shared of incorporating such forms in the OpenMRS or mapping data from different parts of the system and exporting to CanReg.
• Advanced functionalities – CDSS: We discussed the possibility of adding Clinical Decision Support System (CDSS) for challenging oncology tasks such as staging, treatment choices, chemo safety checks and monitoring, and order or order sets for chemo, investigations, etc. This would call for changes in the documentation – structuring the data, adding concepts into the concept dictionary, and maybe even changing the core data model e.g. to better represent chemotherapy drugs (cyclic, combinations, dose modifications, pre-medications, etc). We also talked about the opportunity to incorporate the recently released National Comprehensive Cancer Network (NCCN) Harmonized Guidelines for Sub-Saharan Africa, which IBM, American Cancer Society and Cancer centers and oncologists in Africa are interested in digitizing.

Next Steps

Members agreed that there is extensive work that needs to be done to have OpenMRS fully support oncology. In the next steps to realize this, the following were suggested:

• Share this information and vision with the community and get interest and engagement of key people – subject matter experts, developers, etc. There currently is an Oncology working group where existing resources can be shared and discussion of ideas done. We should also engage other people who are not in the community.
• Look for funding: to make a brief write/concept note which can be used to pitch the idea to small private donors initially until we gain momentum

Clinician Usability & Adherence

Presented by Paul Park

Significance

Whether the data is entered in the EMR by data officers or clinicians at point of care, the downstream data quality is significantly driven by the clinicians’ accurate and complete documentation of forms. Thus, clinician buy-in and support of OpenMRS implementation is a key driver for feedback and improvement of the UI and variables collected.

Challenges (by Outputs)

The following were discussed with examples from the field:

• Clinicians do not fill out forms (incomplete)
• Clinicians fill forms incorrectly
• Clinicians do not give feedback
• Desired data by clinicians is not included; undesired data by clinicians is included
• Requirement of paper version by MOH

Implementation Gaps

The following gaps were identified from the implementer’s perspective:

• Limited training: how to train clinician users?
  • Large scale training vs in-service
  • This training should be part of the initial onboarding of clinicians
• Limited initial input from clinicians on UI design
  • This leads to non-user friendly forms (tick boxes, order or “flow” of form, etc)
• Limited feedback loop: to both clinicians and EMR team
  • Clinicians can be too hierarchical to receive criticism from the EMR or M&E teams
• Required double entry of POC and paper during initial rollout
  • Some Governments require paper versions of all forms. Perhaps there remains some lack of confidence in the EMR system

Closing the Gaps

The following items were discussed as possible solutions to close the implementation gaps:

• Cultivate Buy-in: this is YOUR data; convince clinicians that it will increase efficiency
  • Invite to engage in fruits of data, like research
• Senior officer or other MD provides training and communication on behalf of EMR team
• Monitoring and feedback system:
  • “Public shaming” – share the quality of form filling of clinicians individually in front of their peers on a regular basis
  • Identify point person/champion of the feedback/QI system (MD if possible)
  • Institute immediate turnaround time for informing and correcting errors or mistakes
• POC: guidance alerts and “hard stops” to fill completely
  • Short help videos within the page/form
• Paper requirements by MOH
  • Bare minimum filling out of paper form
  • Print out all EMR forms at end of day to put in chart
  • Data officer transcribes from EMR to paper
  • Real time data quality checks

Action Items

• Initiate OpenMRS Talk discussion
  • Identifying easier avenue to modify/customize OMRS
  • Strategies for OpenMRS Training: need a standardized training guideline, inclusive of clinicians
  • Merge with Training discussion
  • Cross-site sharing forum
• Create OMRS Category

Point of Care (PoC) with OpenMRS

Presented by Ellen Ball

Encore presentation from Global Health Digital Forum (GHDF) in December 2017. A panel exploring real-world experiences implementing OpenMRS as point of care (POC) in Kenya, Mozambique, and Haiti; including discussion of the approaches for both the technical and the human factors of the projects, implementation challenges, and strategies for long-term sustainability. OpenMRS is a collaboratively-developed free and open EMR platform currently used worldwide across 64 countries, in ~2000 sites, serving ~6 million patients. Traditionally used in a retrospective data entry workflow, there is rapidly growing interest in implementing OpenMRS in a point of care workflow. The panel presentations will emphasize lessons learned from actual challenges of four high impact projects: eSaude (Mozambique), KenyaEMR, AMPATH (Kenya), and Partners In Health EMR (Haiti). Click here to view the slideshow on PoC with OpenMRS.

eSaude is a Mozambique based community focused on the development and implementation of a Mozambican specific configuration of the OpenMRS medical record software and the integration into a national eHealth architecture.

The KenyaEMR is a tailored distribution of OpenMRS which meets the requirements laid out in the Kenya Ministry of Health document: 2011 Kenya EMR Standards and Guidelines. It’s currently being developed by I-TECH and has been deployed to almost 140 sites in Kenya.

AMPATH (Kenya) is a Moi University, Moi Teaching and Referral Hospital, and a consortium of North American academic health centers led by Indiana University working in partnership with the Government of Kenya.

The Partners In Health (PIH) program in Haiti, known locally as Zanmi Lasante (ZL), operates clinics and hospitals at 12 sites across the Central Plateau and the lower Artibonite, two of the country’s poorest regions.OCL and Concept Management

Presented by Andrew Kanter

This session continued previous discussions about the potential to use the development of the Open Concept Lab (OCL) to supplement or even replace the native OpenMRS concept management screens. Most of the current concept work is being performed by the Columbia International eHealth Laboratory (CIEL) in an older version of OpenMRS (to ensure legacy effectiveness). This requires quite a bit of overhead. The benefit of having a trusted, curated source of the concept dictionary to OpenMRS is clear. Moving the community and the code base to support both the publishing/subscribing method as well as the curation of the concepts requires further attention. OCL is looking for early adopters to work with the developers and content curators (CIEL) to clarify requirements and be a dedicated user community to provide feedback.

OpenHIE

Presented by Carl Leitner

Open Health Information Exchange (OpenHIE) is a global community working to help resource-constrained environments better leverage their healthcare data. This session began with a brief overview of the OpenHIE architecture. We reviewed the key components such as the Client Registry, Interoperability Layer and Shared Health Record, that a point of service application, such as OpenMRS, would interact with. We reviewed how OpenHIE operates as a community of community, its role in identifying data exchange standards and discussed reference software implementations of OpenHIE components.

After this overview, we discussed practical examples on how OpenMRS could fit into the OpenHIE architecture as well as gave some examples of where OpenHIE has already been deployed or adopted as a framework for a national health information exchange. We compared the use of a Shared Health Record to facilitate continuity of clinical data across multiple types of EMRs to the Sync 2.0 module which provides synchronization of data across OpenMRS instances.

OpenMRS ETL & Data Warehousing

Presented by Willa Mhawila

The aim of the topic was to explore ways to improve data accessibility and decision support systems. Some participants presented the existing techniques which in general involves extracting OpenMRS data into flatten tables (usually using MySQL tables). Once data is extracted different teams use different tools for report generation/data consumption. Some teams have built REST API framework, some are using existing BI products to run queries.

Collaboration between teams was discussed but it seemed very challenging given the fact that, data is usually mapped differently in different implementations. Different teams use different concepts to record information. A solution was proposed to define a starter metadata set that can be used as a baseline for building a tool that could be shared between implementations.

There’s also a common need of denormalizing encounters and observations into separate tables to make reporting easier and faster. Multiple attempts have been made to do so, and there’s a need to create a generic process which periodically does this job.

Collaborating on OpenMRS Android Apps

Presented by Owais Ahmed

The idea was discussed between teams developing different apps, that instead of replicating our efforts we should focus on building a generic mobile app for OpenMRS. At present, the set of functionality offered by Aao TB Mitao, mUzima and other apps is:

• 100% Native Android app
• Offline operation for health workers working in remote areas with none or low connectivity
• Right-to-Left language and layout support
• Mobile UI to provide experience to user consistent with web interface, for example Patient creation widget
• Customizable forms like in Open Data Kit (ODK)
• Android 4.4 and onwards compatibility
• Using REST-WS for interaction with OpenMRS
• Detailed preferences to choose defaults

All of the above have been provided in separate apps, but not in a single application. Teams developing OpenMRS Mobile clients should collaborate with other client developers to eventually come up with one solution that works for many implementations.

One of the challenges discussed was regarding the capacity of low-end devices to hold data. Once a device exceeds data of 1000 patients, it tends to slow down. There are alternatives to SqliteDB, like realm which addresses this issue.

Site Visits

The recipe for a successful OpenMRS conference is one part unconference, one part hackathon, and one part site visit. Each year the host country organizes the opportunity for visiting community members to see firsthand how the local Ministry of Health is addressing their public health needs. The Ministry of Health of Malawi built the foundation for their EMR platform using OpenMRS and called it Baobob Health Trust (BHT). The implementation of this system has translated into over 1.5 million patients being given unique patient identifiers since 2001. These unique identifiers enable patient follow-up across the continuum of care. Additionally, 30% of the national patients enrolled and treated by the ART program are managed using these EMRs that have been deployed in all regions of the country. The MOH plans to further develop and implement other modules in a phased-approach. These modules include management of chronic diseases; antenatal, maternity and under-5 services; radiology information system; management of lad specimen; and TB-ART management for co-infected patients.

In order to encourage knowledge sharing and continuous development of the EMR, BHT implemented free and open source software (FOSS) standards. BHT innovation includes technology that can operate in challenging environments afflicted by by power failures; dusty conditions; and healthcare workers unfamiliar with the use of these technologies. The result is BHT using low power-consumption devices, relying on renewable energy sources off the national electricity grid; solid state hard drives; and intuitive touch screen user interfaces to reduce the learning curve and and empower healthcare workers to rapidly adapt the new technologies.

Community Voices

The distance from Indianapolis, IN to Lilongwe, Malawi is 24 hours excluding delays, layovers, and lost luggage; where many experienced all of these travel woes. Regardless, the warm-welcome and cheery smiles of the OpenMRS community serves as the best remedy to jet-lag, and the incredibly helpful planning committee helped to acquire any lost baggage along the journey.

In all, the single to multi-day-long journey is worth it for the many that seek to bring OpenMRS to the furthest corners of the earth. Here’s what they have to say:

“OpenMRS has helped me personally and professionally by encouraging me to work more collaboratively. I’ve also made a lot of friends! The hospitality of the community encourages me to contribute and share as much as possible. I think I’ve fallen in love with OpenMRS! XD.” – Chine Zoheir, Algeria

“It was a great experience being in the conference. For the first time, I saw how big OpenMRS is and how it grew from the beginning and it made me feel very honored to be a part of it. It was fascinating to see the Baobab system in Malawi and how it improves the workflow of treatment and treatment of the patient. I hope that Mozambique will someday achieve this level or better.” – Carina Matimbe, Mozambique

“The OpenMRS 2017 conference was a clear demonstration of the direction OpenMRS is headed in helping to solve electronic patient data needs at a national level.” – Stephen S. Musoke, Uganda

“Healthcare systems can be fragmented and, as a user, we are concerned with the data interoperability between the mobile based care and facility based care. Sessions during the conference gave us the clear picture on the importance of data exchange between different platforms. OpenMRS is growing and is easy to implement in any facility with minimal resources. The beauty of OpenMRS is that there are no other dependencies with the service level setup. The community is growing and these conferences helps us share our experience and bring all the Health informatics/HealthIT people to one common platform. I would say its the best platform for all the clinicians and those group of people who wants to see the real impact through the data driven health care model to create smile on the ground.” – Sanjaya Poudel, Nepal

Closing Ceremony

The closing ceremony of the 2017 implementers’ meeting gave ample time to the community members to share what their expectations were heading into this conference and how they were met, or weren’t met. This feedback is a critically important process the OpenMRS meetings include each year in order to improve the following year’s meet up. Additionally, delegates shared their ideas for how to improve the meetings in the future.

Looking Forward

OpenMRS concludes 2017 and looks toward 2018 and beyond to fulfill their mission and support the community’s efforts toward improving healthcare delivery and healthIT. Recently, OpenMRS has been awarded a 1 million USD donation from an anonymous cryptocurrency philanthropist known as ‘Pine.’ Pine started the Pineapple Fund to as a way to support organizations working to improve the world we’re in. OpenMRS will use these funds in accordance with a strategy that is still in development. Attention will be paid to the strategic goals developed by the leadership team as well as the goals established by the community at the 2017 Implementers’ Conference. These strategic goals were agreed upon at the end of the unconference sessions and is a unique way for the community to get more involved at an administrative level.

As of 2017, the OpenMRS platform is being used by over 1,845 sites in over 64 countries, and impacting more than 6.3 million patients. Continuing into 2018, the community’s ambitions, combined with the recent and incredibly generous contribution from the Pineapple Fund, ensure that these facilities that depend on OpenMRS, and systems like it, will receive the additional technical support they need to effectively treat and care for their patients.

To keep up to date with OpenMRS’ strategic plans, visit their Community Strategic Goals Wiki page.