In 2013, we met with Prof. Azad, Additional Director of the Directorate General of Health Services (DGHS) of Bangladesh, whose vision was to introduce a longitudinal electronic health record for the citizens of Bangladesh. He was looking for a partner and we knew that the use of technology to improve health care services in these resource-constrained environments could make this revolutionary.
We partnered with the DGHS, Department for International Development (DFID) and various organizations to enable this vision by building an ecosystem of inter-operable products to create a Shared Health Records (SHR) platform, making health information available across facilities. It would also allow for more efficient care and better allocation of hospital resources. There was no platform readily available, suitable for a large country like Bangladesh. Additionally, we had to integrate this platform with all existing operational systems in Bangladesh, making the scenario rather complex and challenging. We needed to build a system that was frugal in cost, both to build and maintain, that could also work in environments of moderate Internet connectivity and be conducive to scale.
Demonstration in a Lab Setting
The central Health Information Exchange (HIE), which includes the SHR, was built by a Thoughtworks India-based development team over a one-year period, prior to successfully demonstrating the ability to seamlessly share health information across facilities in a lab setting. The solution consisted of a central HIE and a reference electronic medical record (EMR) system, Bahmni (a distribution of OpenMRS), for hospitals and community clinics. Once the feasibility was assessed in a lab setting, the DGHS asked the Thoughtworks team to pilot this solution at two locations - to make it work in a real environment and to uncover challenges that would need to be addressed before implementing this solution nationwide.
Setup for the Implementation
The DGHS formed a working committee of various government bodies to set expectations on the ground. We, on the other hand, were dealing with a mixture of emotions - the excitement of doing something hugely impactful over the long-term, coupled with fear of making the SHR work in an environment that was relatively unknown. However, when we thought of the question “How often do we get to make a positive impact on millions of lives?”, it trumped any apprehension.
The Thoughtworks team did an initial assessment of the facilities where the pilot was being run - Kaliganj Upazila Health Complex (UHC) and Durbati and Chandaya community clinics, during which we identified and fixed gaps before our solution was rolled out.
The assessment showed us that the registration desk, that saw over 350 people daily, was swamped with people on all sides. The patients crowded around clerks, who were registering patients using paper. We realized that the lack of an actual counter could potentially risk the equipment that would be made available to facilitate the registration process.
Up until then, their approach had always been paper-based, with little to no exposure or experience with technology. We were aware of the challenge that we had to address - the need to take people on a journey, to not only learn to use computers, but to trust its abilities and value its benefits.
The most important aspect was to build local capability to manage this change. We knew that it would be wise to identify a few key people who would partner with us during the implementation, which would also help us manage scale. We formed an implementation team comprising of people from the DGHS, Thoughtworks, key hospital staff and the DGHS’ local partner ICDDR,B (International Centre for Diarrhoeal Disease Research, Bangladesh). We ran an extensive training programme on Bahmni for this team to prepare them for the implementation.
Over the next 2 months, we worked as a distributed team on each of our responsibilities - setting up the central HIE, sourcing and setting-up data, feature development for Bahmni and the SHR, configurations to suit the facilities, procurement and the setup of the required infrastructure. There was never a dull moment.
Even though additional requests kept coming our way, we went live with what we believed was a good minimum viable product (MVP).
It was an intense 3 weeks of implementation. Since the implementation team had been through an intensive product training, we encouraged them to train the staff at the Kaliganj UHC and community clinics on the features of the system. The training also included basic computer training, to cater to those who had never used a computer. To make the training interesting, we used innovative learning methods like games and typing tools.
During the course of the training, the team received a lot of valuable user feedback that was useful to make refinements to the system. Most of the feedback was related to data that had to be available on the system. This prepared us for the rollout by mitigating possible challenges that we would have otherwise faced.
We rolled out Bahmni in phases at the hospital to better manage change of this magnitude. During the first phase, we focused on the registration process for a subset of patients to ensure effectiveness and a smooth flow. At that time, both the original paper-based system and the digital process were operating simultaneously. In each phase, we focused solely on one department of the hospital. We found that this phased approach reduced disruption and allowed for the hospital staff to get comfortable with the system. Over time, we made sure that all processes and departments were introduced to Bahmni.
A simplified version of Bahmni was implemented in parallel at the Durbati and Chandaya community clinics. We made adjustments to the interface on the spot as we got feedback from the Community Health Workers (CHW), which was very effective.
On completion of the implementations, the systems were exchanging information through the central HIE. We were able to clearly demonstrate the capability of the HIE platform via a live environment, as information from the community clinics was flowing to the next referral facility, Kaliganj UHC, and vice versa. Very exciting!
The team balanced multiple threads, with the support of the India team - training during the day, sourcing final data, working on changes to the system and coordinating with our Bangalore and Hyderabad teams until late at night. The close collaboration and assistance from the entire team helped us progress efficiently. Every day, I summarized our on-ground experience through a Skype journal for the entire team involved, which provided much needed insights for people not on the ground. We did manage to take some time off to recharge ourselves - we took an evening boat ride on the river.
After the successful deployment at Kaliganj, in early 2016, we implemented Bahmni at Gazipur, which is the referral hospital for Kaliganj UHC.
While we were able to successfully demonstrate the solution, we have uncovered operational challenges that need to be addressed prior to the national rollout. It is imperative to have training modules for new clinicians and refresher programs. It is critical for the DGHS to have an IT focal point to enforce process changes at the hospital, to provide continued maintenance and support, as well as ensure adequate power backup. We need a solution to eliminate redundant manual work. Another problem to be addressed is the duplication of registration for returning patients.
We are addressing all of these operational challenges in collaboration with the DGHS. In order to scale nationally, we will need to foster partnerships, which will help build additional capability for implementations. This will also be made possible with the continued support of the Joint Donor Technical Assistance Fund (JDTAF) of the DFID.
We are numerous steps away from the national rollout, but once it’s complete, over 160 million Bangladeshi citizens will have an electronic health record. We would have contributed to a piece of Bangladesh’s medical history by building a strong foundation towards universal health coverage for all.
Disclaimer: The statements and opinions expressed in this article are those of the author(s) and do not necessarily reflect the positions of Thoughtworks.