Renowned fictional detective Sherlock Holmes, who first appeared in print long before artificial intelligence came along, knew the critical role of data for solving problems. He famously said: “It’s a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” Sherlock Holmes’ insight is spot on. But how can we hope to strengthen health systems when the data we rely on is often incomplete, riddled with missing values, or poorly digitised? In this post, we explore innovative strategies to transform flawed health data into actionable insights. Keep reading to discover our solutions.
Our strategies
Here are six key strategies to transform complex or incomplete health data into actionable insights.
Close collaboration: We always work in very close collaboration with local partners to better understand their routines and bottlenecks. We can then ascertain which information is most useful to them, as well as helping come up with a solution where data collection is less than ideal.
Epidemiological knowledge: Based on our knowledge of epidemiology – i.e. how diseases are distributed and the factors that influence their distribution in populations – we can assess how our data can best be used, and what assumptions we can fairly make.
Open source data: We like to use relevant, contextual data to compensate for any missing or incomplete data, or where we only have facility-based data. Open source data could include environmental data, as well as relevant local data including culture, income and the political situation.
Advanced technology: Utilising AI and machine learning approaches, we can process vast amounts of data quickly to produce predictive models for infectious disease control, or alternative public health decision making. We also incorporate real-time data collection so the models are as accurate and up to date as possible. To help with communication, we use simple-to-read portals and dashboards.
Pilot phases: We launch small pilot phases in select regions. We can then address any challenges or bottlenecks in data acquisition. Successful pilots are then scaled up to cover larger regions.
Capacity building: Ideally, we want local teams to use our platforms independently. We work to empower regional local teamsworkers to upload and manage their data so that they can gain autonomy over the process. Simple tools, such as dashboards, make data usable at all levels, ensuring that even community workers can capture and utilise crucial information that might otherwise be overlooked.