Development and practical use of a risk-sensitive population segmentation model for healthcare service planning: application in England

Population segmentation can be a powerful tool in healthcare management, in helping to match interventions and resources to individuals of a common health state and condition. However, studies to date indicate a lack of alignment to risk stratification – another key tool in Population Health Management – and insufficient demonstration of how segmentation can be used in practice. In this study, we obtain a five-cohort segmentation derived through four incremental thresholds on the risk-based Cambridge Multimorbidity Score, which is calculated for each member of the 762,117 adult population in and around Bristol, England. Appropriately selecting the four thresholds – 0.09, 0.69, 1.59, 2.95 – yields a segmentation with the convenient property that, with increasing risk, segments halve in size and double in per-person spend. The segmentation has been used to support various planning and management activities within the Bristol healthcare system; two of which are detailed here as case studies in demonstrating the practical value of the segmentation model.