On June 9, the USAID Country Health Information Systems and Data Use (CHISU) Project hosted their third webinar in the technical leadership series titled, “Digital Systems and Data Quality: Will we, or have we, gone beyond data quality conversations with digital systems?” Derek Kunaka, HIS Director, CHISU Project, JSI, moderated the webinar with panelists: Amanda BenDor, Deputy Director, Partnerships for Digital Square, PATH, Lungi Okoko, Senior Technical Advisor, PMI, USAID, Jeni Stockman, Senior Program Manager, Macro-Eyes (CHISU Partner), and Anton Zamora, Digital Health Project Lead and Researcher, Board of Directors of RECAINSA. Derek introduced the WHO’s classification of digital health interventions, which notes challenges of health systems and describes a health system taxonomy that is widely used, helping to set the stage for our panel to explore how to ensure data quality within digital systems.
Data-users at all levels need access to quality, actionable, and sometimes triangulated data in formats that are intuitive and interpretable. Too often indicators, software, and dashboards are designed without the perspective of those who will need to use the data to make decisions. Amanda includes “We also want to make sure that the tools are designed with the user in mind, this includes multi-language support, strong user documentation, and access to communities of users through communities of practice.”
However, it’s not enough to focus on building the right system and tools. Anton comments that often those working in digital systems are “thinking about the software, the solution, and all the fancy things, but the last thing is what are the data outcomes, what are the data analytics, how will this data be used?” Health information data users can vary from policy makers, academia, to health providers but how much time, effort, and resources is being spent to get that data, and are they using it?
Country health information systems (HIS) often can fall short in terms of routine data verification and data quality assurance (DQA), as evidenced by their own reports after measurements of data quality, and enormous amounts of funding goes to the quest for better data. However, Lungi notes that in a “comparison between the data that countries had shared with us initially and validated data, [they] found that corrections were in many cases, minimal, which leads to another barrier that perfection being the enemy of good.“
Leveraging technological advances and innovative tools could lessen the burden of effort to improve data quality and free up health officials to spend their time on analysis and interpretation, especially in low resource settings and hard to reach locations. In fact, artificial intelligence (AI) has the potential to use the imperfections, gaps, and errors in data to provide key insights. Jeni adds that the “opportunity for AI to correct human efforts and to start collecting data on its own, even data we never would have considered of value.”
Health information systems and the increasing emphasis on digitalization, health program-driven innovations and technologies, and user-centered design are promising for the future of data. However, we must consider various aspects of our design when building a digital system, from the system to the end-user, and find that balance between the quest for better data quality with seeing what data is at hand, how will it be used, and what can we learn from our sometimes imperfect and messy data.