From Cases to Consumption: Evaluating CHAP/DHIS2 Model Portability for Health Supply Chains
From Cases to Consumption: Evaluating CHAP/DHIS2 Model Portability for Health Supply Chains
Webinar: DHIS2 Climate and Health Tools for Planning and Monitoring Immunization Programmes.
Context
At the Webinar: DHIS2 Climate and Health Tools for Planning and Monitoring Immunization Programmes, we address a critical gap between disease forecasting and supply chain logistics.
Platforms like DHIS2 and CHAP are increasingly powerful at forecasting disease cases, often leveraging climate data to predict outbreaks. However, forecasting the product consumption needed to respond (e.g., vaccines, antimalarials) remains a persistent challenge. Consumption data is notoriously incomplete, inconsistent, and distorted by stockouts, making it difficult to plan effectively.
This presentation investigates model portability: can the robust, high-performing models built for cases be effectively repurposed to forecast consumption? We evaluate when to reuse, when to fine-tune (e.g., with climate or campaign data as external regressors), and when to retrain models from scratch.
Our work provides a practical framework that links statistical forecasts to inventory decisions. By understanding which model to trust, health systems can move from reactive ordering to predictive logistics, ensuring life-saving commodities are available to meet climate-driven health needs.