Lessons Learned in Electronic Immunization Registry Development
By Ola Hodne Titlestad, Health Information Systems Programme, University of Oslo, and Dawn Seymour, Deputy Director, BID Initiative
Nov 23, 2016
Posted in Products
Mulenga Malata of Simoonga Clinic explains to Markus Bekken of UiO what immunization data nurses capture in the child health card. Photo: Fred Njobvu.
In partnership with the Zambia Ministry of Health (MOH), we’re committed to strengthening Zambia’s immunization services through better data collection, quality and use. From the start, we knew that technology alone couldn’t achieve our goals which is why we chose a holistic investment in information system products, data management policies and practices, and the people who will use the data for decision-making.
As one of BID’s core principles is to design innovative solutions, the Zambia MOH and the BID Initiative partnered with the University of Oslo (UiO) to develop a standards-based electronic immunization registry (EIR) to complement BID’s broader set of interventions which focus on strengthening data management policies and practices.
The EIR being developed was a patient-level Android application for use in health facilities. This application was built on the generic DHIS 2 platform, which is already implemented in Zambia as an online health management information system for aggregate health indicator reporting. While there has been significant progress on the EIR development, more work must be done to ensure the system is secure, adaptable, and scalable.
Our partnership provided a deep, hands-on opportunity to fully understand the requirements needed for this cutting edge application. However, because the time needed for development to meet these requirements exceeds the BID Initiative’s timeline, we are identifying alternative EIR solutions in partnership with the MOH to use in the suite of data quality and use interventions for Zambia. The requirements for the EIR and lessons learned from the BID and UiO collaboration will continue to inform UiO’s ongoing development of the EIR which will be available to many countries for future adoption and implementation. This process has been a valuable research exercise and as part of the BID Initiative’s commitment to sharing achievements, challenges, and lessons; below is a snapshot of our lessons learned to help other countries and partners interested in developing digital health products, such as an EIR.
Lessons Learned
- When developing software for cutting edge, innovative solutions a thorough hands-on research phase is needed to ensure the proposed system will meet the needed requirements. While the web-based version of DHIS2 Tracker offers functionality for patient-level management, it was later clear that having this functionality on an Android device, especially in an offline environment, was more challenging than first envisioned. As the developers talked through the issues in detail with the subject matter experts (SMEs) and interacted with the BID Zambia team through prototypes, more detail to existing requirements resulted in unpacking new challenges that had not been catered for in the initial design. High-level reviews are simply not enough for software development of this size and complexity.
- Android devices and apps have not typically been designed for multi-user environments where data are shared offline between users. The requirement to support multiple users accessing the same “database” (metadata and data stored on the device) without connection back to the central server, and doing so with their own personal access control and auditing features, had not been catered for in the thin client architecture for the DHIS2 Android apps. The ability to tailor and audit patient data access for each individual user is important in keeping in line with best practice standards for maintaining client level medical records, and this requires a complete re-design of the DHIS2 Android app.
- All participants should have a full and clear understanding of requirements and expectations from the start. For example, developers should be involved in the full process of implementation and deployment planning, including spending enough time in facilities to ensure they understand the functionality needed and internalize the context in which the technology will be deployed (e.g. how many users, what is the workflow that needs to be supported). These are all key questions that should feed into technology design and therefore need to be understood by the developers in the beginning of the development cycle.
- Documenting decisions and developing a shared roadmap is essential. A good tracking system and shared roadmap kept in one place is needed to reflect requirement discussions and decisions as a way to ensure direct visibility by all parties into the complete process and development progress. Additionally, strong project management and communication across all teams allows questions and issues to be raised as they happen and enable more direct, timely communication to keep development on schedule.
- Critical requirements should be prioritized rather than prioritizing requirements by phases. A more efficient route would be to front load the items deemed to be key and technically hard to implement, such as maintaining an audit log of the changes and history of the data, and the ability to allow multiple users to work offline with data synchronization during connection with the server. Rather than moving incomplete requirements to future iterations, there should be greater focus on completing a higher percentage before moving to the next iteration to keep the workload more balanced.
- Ensure key MOH staff at the national level are in place to participate in the entire process. Engagement with the national level of the MOH throughout all stages can help identify gaps in the system and process at earlier stages.
- The depth and strength of skills sets needed on all sides during the process need to be agreed upon and guaranteed. Both teams saw the need for additional capacity and investing in a local, external expert and an independent software expert with Android experience would have helped with identifying solutions earlier on.
While the BID Initiative continues working with the MOH to identify alternative EIR solutions and UiO continues its own research and development, we’re moving forward to implement interventions to promote data use such as targeted supportive supervision, data use guides, and data visualization. Creating this culture of data use is critical to enabling health workers to make better decisions on service delivery and ways to improve care. The additional tools selected will complement other BID solutions and ease data collection and access.
Thank you for continuing to follow along on our journey and allowing us to share our successes as well as the challenges we face along the way. Stay tuned for more updates on BID’s rollout in Zambia.