# Closing the Gap: Refining Vaccine Forecasting and Resupply Calculations in Mozambique Through Advanced Routine Data Analysis

## TL;DR

This study improves vaccine forecasting in Mozambique by analyzing routine data to reduce stockouts and improve resupply decisions.

## Contribution

The study proposes using facility-level data and tailored wastage rates instead of population-based forecasts for better vaccine resupply.

## Key findings

- Population-based forecasting undersupplies facilities by 20% on average.
- Using BCG doses as a proxy leads to tighter variance in resupply forecasts.
- Facility-level wastage rates differ significantly from global standards.

## Abstract

Background: Vaccines save lives, but only if they are available at health facilities for administration. Stockouts occur for various reasons, including inaccurate forecasting and resupply calculations. Population-based forecasts are typically used for immunization programs, yet they are often based on inaccurate population estimates. This retrospective study analyzed available routine facility-level data from two districts in Mozambique to provide insights for improved supply chain management, including resupply decisions, at the facility level. Methodology: Data from August to October 2023 were collected and analyzed for wastage rate, session cohort, and forecast accuracy. Results: The results show that district-level wastage rates are nominally different from globally acceptable standards, while being significantly different at the facility level. Analysis also showed divergence of vaccination doses provided to a session cohort during the month-long periods that appear to be correlated with periods of stockout. Using population-based forecasting for resupply methodology consistently undersupplied facilities by 20% (ranging from 5 to 41% across 16 facilities), while using the number of doses of administered BCG as a proxy for the population oversupplied by 12% (ranging from 1% underforecast and 28% overforecast), with tighter variance. Conclusions: Despite limitations due to the availability and quality of data, the results suggest an opportunity to shift from a traditional population-based approach to forecasting and resupply decisions, leveraging existing data systems, applying tailored wastage rates, and adjusting inventory management policies to ensure vaccine availability.