Probabilistic Forecast Reconciliation in Global Health Supply Chain Operations

Within the domain of Global Health Supply Chain Operations, the accuracy of demand forecasts emerges as a critical element. These forecasts are of utmost importance for planning on national, regional, and sub-regional levels. Typically, these forecasts are generated independently by various teams, giving rise to inconsistencies at various levels of supply chain. This misalignment can result in conflicting decisions and a lack of synchronized coordination throughout the supply chain. To address this issue, we harness the inherent hierarchical structure embedded in demand time series data. By implementing advanced techniques in forecast reconciliation, we generate both point and probabilistic forecasts that are not only coherent but also capitalise on the spectrum of available time series patterns across the hierarchy, which can lead to forecast performance improvement. We apply these methods to real-world data containing two distinct global health supply chains, broken down into regions, counties, and sub-county segments. Employing an ensemble of forecasting models, we demonstrate that the resulting forecasts outperform the performance of individual models. Rigorous testing through time-series cross-validation validates the effectiveness of forecast reconciliation. By integrating forecast reconciliation principles with global health supply chain dynamics, our research presents a promising avenue for refining forecasting across the supply chain and promoting coordination.

Bahman is a Reader (Associate Professor) in Management Science at Cardiff Business School, Cardiff University, UK. He serves as the director of the Data Lab for Social Good Research Group at Cardiff University and is also the inaugural Chair of the Forecasting for Social Good committee within the International Institute of Forecasters.  Bahman specialises in the development and application of forecasting and management science tools and techniques. His focus lies in leveraging these methodologies to provide informed insights for planning & decision-making in sectors contributing to social good, including healthcare operations, global health and humanitarian supply chains, social sustainability, and governmental policy. His collaborative efforts have spanned a multitude of organisations, including notable bodies such as the National Health Service (NHS), Welsh Ambulance Service Trusts, United States Agency for International Developments, the International Committee of the Red Cross, JSI, and the Ethiopian Pharmaceutical Supply Service. A remarkable highlight of his contributions is his pivotal role in disseminating forecasting knowledge. Through his pioneering "democratizing forecasting" initiative, sponsored by the International Institute of Forecasters, Bahman has conducted impactful training sessions that have benefited over 600 learners across more than 13 low and lower-middle income countries and in the UK since its beginning in 2018.