October 2017 Global Digital Health Network Meeting – A discussion on Interoperability: Where have we come from, where are we, and where do we still need to go
Dr. Paul Biondich from Regenstrief kicked off the meeting by giving an overview of interoperability and architecture. When considering how to integrate and compare data from different environments, it is possible to do point-to-point integration, but what is now becoming more common is establishing national or regional level standards of language. In Kenya, for example, they have established a unique set of facility codes across the whole enterprise.
The need for establishing standards became an opportunity to launch a community of practice: Open Health Information Exchange. This platform facilitates agreement and good governance around data, semantics, syntax, and workflows and shows how these can be used with established systems and an interoperability layer. Anyone can join, and content is free and available, including standards, workflows, and implementation guides. The primary work around interoperability is getting countries to recognize that they need appropriate governance, capacity, and robust processes to support and maintain the health enterprise. This is a multi-year journey.
Paul made the key point that countries that use interoperability architecture to solve a health problem are often most successful. Implementing fully realized technology upfront may be challenging if it is not proven against specific health needs. Expanding to implement more use cases should be done iteratively. For example, Tanzania spent 2-2.5 years to establish processes and good governance around a national facility registry. This is just one piece of standardization, but, now that this is in place, they have a sustainable model to guide future standardization.
The second speaker, Sheel Shah from Dimagi, spoke about Dimagi’s role as one of the systems in the operability space and changes in their perspectives and approaches over time. Dimagi develops software and provides services support for mobile solutions for health workers. CommCare is the primary tool, which supports mobile application development for data collection, service delivery, and more. Dimagi had previously built customized software for each project and client, but it became very difficult to support all of these unique systems. They wanted a standard system, and this led to the development of CommCare. Dimagi often integrates CommCare with common systems, such as DHIS2 and OpenMRS and has worked with Grameen on Motech to create a user-configurable product for integrating across common platforms that non-technical staff can use.
Dimagi sees 3 key types of demands and needs for integration:
DHIS2 integration was a common need and relatively easy to achieve. Complex workflows don’t lend well for user configurability. A lot of integration work is custom and very time consuming. There wasn’t a market fit for user configurable integration product. However, integration is still important and Dimagi wants to support it. It is worth building an end user configurable interface for integration with DHIS2 and refocusing Motech on building scaffolding to enable integration.
The final speaker, Michael Stelmach of Bowlink Technologies discussed how Health-e-Link and JSI are working to build out the HIE for Tanzania. A conceptual model shows system inputs and how they interact via an interoperability layer with the HMIS (health data registry). The Health Information Mediators or interoperability layer is provided by Health-e-Link. The Tanzania use cases accept data from hospitals’ EMR and provides information to ministry in the form of a dashboard. Different hospital senders may use different types of transport and send different kinds of messages. Each hospital has its own configuration in the interoperability layer, which translates into full normalized data into the health data repository. Michael then provided a demonstration to share the processes and value of Health-e-link, including how to view the dashboards and see key indicators by hospital.
During the discussion, Michael stressed that this is a never-ending journey because of the scope and complexity to change workflow and health care delivery. There are three important considerations:
- Data reporting: Collect data in one system and need to push it to another system, often DHIS2. There is a clear pattern for the integration, but it is generally necessary to modify both DHIS2 and CommCare.
- Workflow: Need to exchange information between systems to support service delivery, often OpenMRS. Every OpenMRS project is unique, and the workflows are complex, so the goal of having non-technical people configure integration was not realistic.
- Tactical/Political: Desired integration with another system, often a custom platform, but less clear objectives or value. Typically these systems were not designed for integration. These projects generally wanted to be hosted in country and required ongoing maintenance and support.
- Governance and Consistent leadership and ownership
- Technology – reliance on standards alone is insufficient; technology will continue to evolve
- Sustainability – it is important to consider how to support or compel health providers to participate. Networks need to be paid for by organizations in charge of the health care system, like ministries of health or states. The sustainability question remains unanswered.
Thanks to JSI for hosting this meeting. This is our last monthly meeting of the year as we prepare for the Global Digital Health Forum Dec 4-6th.