Abstract
Providing accurate utilization forecasts is key to maintaining optimal vaccine stocks in any health facility. Current approaches to vaccine utilization forecasting are based on often outdated population census data, and rely on weak, low-dimensional demand forecasting models. Further, these models provide very little insights into factors that influence vaccine utilization. Here, we built a state-of-the-art, machine learning model using novel, temporally and regionally relevant vaccine utilization data. This highly multidimensional machine learning approach accurately predicted bi-weekly vaccine utilization at the individual health facility level. Specifically, we achieved a forecasting fraction error of less than two for about 45% of regional health facilities in both the Tanzania regions analyzed. Our “random forest regressor” had an average forecasting fraction error that was almost 18 times less compared to the existing system. Importantly, using our model, we gleaned several key insights into factors underlying utilization forecasts. This work serves as an important starting point to reimagining predictive health systems in the developing world by leveraging the power of Artificial Intelligence and big data.
Highlights
Vaccines have been touted as the “single most life-saving healthcare innovation ever” (Orenstein and Ahmed, 2017)
The majority of existing vaccine utilization forecasting systems fall into one of two broad categories: 1) Routine data collection such as data on immunization and/or stock level changes entered by health workers (Logistimo, 2011) and past trends detected from immunization and/or stock level change data extrapolated to forecast future utilization; and 2) Population level data using population level survey data on pregnant women and child births
Facility refers to a health facility. 1 indicates the data for the feature was obtained through PATH and Tanzania MoH. 2 indicates the data for the feature was obtained from other sources
Summary
Vaccines have been touted as the “single most life-saving healthcare innovation ever” (Orenstein and Ahmed, 2017). A recent study on 94 low- and middle-income countries estimated that a $34 billion investment in immunization programs resulted in savings of $1.53 trillion in broad illness-related economic benefits (Ozawa et al, 2016). Maximizing immunization coverage for any population is an important public health goal for all countries and 194 Member States of the World Health Assembly in May 2012 agree, having developed a framework to prevent millions of deaths by 2020 through more equitable access to existing vaccines for people in all communities (WHO, Global Vaccine Action Plan 2012–2020). One of the challenges that countries need to overcome to move closer to this goal is accurate forecasting of vaccine utilization (Meuller et al, 2016). The majority of existing vaccine utilization forecasting systems fall into one of two broad categories: 1) Routine data collection such as data on immunization and/or stock level changes entered by health workers (Logistimo, 2011) and past trends detected from immunization and/or stock level change data extrapolated to forecast future utilization; and 2) Population level data using population level survey data on pregnant women and child births
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have