We solidified our innovation strategy during the first quarter of the year by kickstarting several innovations in our pipeline. Notably, Living Goods secured Agency Fund support to join the Precision Health Partnership—a collective of Living Goods, Reach Digital, and D-Tree working to use predictive analytics to address health challenges, with the potential to improve healthcare access and outcomes more cost effectively.

We are developing an algorithm that Living Goods and governments can leverage to
identify children and households at higher risk of defaulting on routine childhood immunization schedules. Phase 1 of the project focused on data collation and processing, model design and building an internal validation using program data from Kenya and Uganda. Findings suggest we have developed a predictive algorithm yielding accurate and relatively precise results. Religion, gender, wealth quintile, and maternity history were identified as predictors of defaulting on vaccines. The initial testing of the algorithm predicts
both defaulters and non-defaulters with an error rate range of 11-12% and predicts true defaulters with a precision rate of 70% to 80%.

The Precision Health Partnership is currently seeking co-funding to unlock further investment by the Agency Fund to proceed to the second phase. This will include field testing of the algorithms as well as evaluating interventions targeted at high-risk client segments that have been identified using advanced predictive analyses. In the meantime, we are further refining the model to add more variables, such as nutrition, and are conducting a scoping exercise to identify more use cases for predictive analytics to improve CHWs’ efficiency and develop our data science capacity.