The OpenMRS story: A FOSS Medical Records System for Africa and Beyond

Facilitated by Ben Wolfe, OpenMRS

An overview of the what/when/why/how of developing a free medical record system for environments that need it most.

Open MRS

Notes, questions, and points

First install was three years ago in Kenya. Supports 70k patients

Project does not support installations. Uses the parent company to facilitate this.

Open MRS is built to be generic and modified for the specific clinical needs.

One does not need to know Java to customize forms

It is funded by parent companies who get grants, use OpenMRS to deploy, and pay for customization by the parent company. For example, a scheduling grant for WHO.

What does a local install look like without internet?

This is a patient based MRS; it does not do billing or prescriptions. It tracks patients to satisfy government needs for data and service metrics. It is used mostly in developing countries, but now some clinics serving the poor in the US are using.

Servlet container with java, drop in a mysql database, and load the openmrs java.

Open standards vs. privacy. HL7 is the standard.

Privacy is about what questions and answers are part of the patient records, and where they are stored.

Misconceptions exist from Kenya, for instance, about Proprietary companies taking advantage of the hosts' goodwill by using them as R&D or metrics accumulation.

What is the goal and impact of this effort?

To reduce the duplication of similar solutions that previously stayed in a 'silo' Now a granting agency can request a specific solution, and they keep the solution for further projects.

Many nations use paper forms, still, and the doctors are interfacing with paper due to usability concerns.

Usability through the software is still being tested and developed with Kayak.

To address the mobile point of entry options there are emerging options.

To deploy to domestic clinics, one would need to work with those not billing and simply needing comprehensive and usable records.

Many developers are project specific, and only one is full time OpenMRS developer. There are 7-10 regular committers and 3 voluntary committers.

To handle customization, they use Spring.

3 Key foundations for the project

1. Data layer and data model well described at website

2. API service layer on top of it. The jar file can go anywhere.

3. The common front-end is a tomcat war file and servlet.

Metadata is gathered for customization.

Key Takeaways

1. Patient based system not running with internet that tracks patients

2. Building a trust factor with African clinics wary of outside exploitation

3. Operates with the emerging hybrid model whereby the software stack is FLOSS, and the independents can deploy, re-submit the code thus perpetuating the project through more diversified deployments