scroll

Creating a predictive supply chain to enable proactive vaccine delivery

November, 2021

For immunization programs, accurate utilization forecasts are key to maintaining an optimal supply chain for vaccines; accurate forecasts save lives.

Challenge:

For every 100 vials of vaccine used in Tanzania, the existing supply chain delivers between 43 too many and 43 too few vials. When too few vials were delivered, lives are at risk. When too many doses are delivered, wastage occurs. These significant inaccuracies stemmed from incorrect data in the system and from traditional forecasting methods incapable of addressing volatility and uncertainty. Forecasts were developed from averages of prior consumption or based on outdated population estimates. Most significantly, legacy forecasts did not take into account factors that would impact vaccine demand. The legacy system assumes that the size of the catchment population translates directly to units of demand.

Solution:

To address wastage and the missed opportunities for immunization that inaccurate forecasts create, STRIATA-Forecast was deployed to predict vaccine utilization at all levels of the health system in two regions of Tanzania. STRIATA-Forecast learns from government-owned data sources as well as public data including satellite imagery, demographic and health surveys, and World Bank education and economic data.

Why does STRIATA use public data?

Vaccine utilization at any given health facility is driven by factors apart from demographics. Factors predictive of demand may include the physical characteristics of the facility and how its placement in the broader geography restricts or eases access, type of facility (private or public), altitude, and public sentiment regarding the site and its services. STRIATA-Forecast uses vaccine utilization data and data about the health facility and population to provide accurate vaccine utilization forecasts and insights into the data itself.

Results:

STRIATA-Forecast cut the rate of error from 43 vials per 100 to between 1 and 2 vials for every 100, improving forecast accuracy of essential vaccines by 96% compared to the prior forecasting methods. STRIATA-Forecast is being implemented to power the supply chain to reduce vaccine stock-outs, eliminate missed opportunities for immunization, and cut expensive wastage by predicting consumption and making automated recommendations for supply replenishment.

Currently, STRIATA-Forecast is being integrated into the Tanzania Vaccine Immunization Management System (VIMs) to be used by the District Immunization Vaccination Officials (DIVOs) who have final decision-making power over vaccine supply allocation to facilities. STRIATA-Forecast much more clearly predicts the future, enabling the government to much better prepare.

For more information on STRIATA-Forecast in Tanzania, read the recently published peer -reviewed paper here.