Dementia is an increasing problem in today’s ageing society, and meeting future demand for care is a major concern for policy-makers and planners. This paper presents a novel hybrid simulation model that simultaneously takes population-level and patient-level perspectives to calculate the numbers of patients at different stages of disease severity over time, and their associated care costs. System Dynamics is used at population level to capture ageing, dementia onset, and all-cause mortality, whereas disease progression is modelled at individual patient level using Agent-Based methods. This enables the model to account for variability between patients in the rate of cognitive decline, dementia-related mortality and response to treatment interventions. Using epidemiological data from the medical literature, disease progression is modelled via a longitudinal clustering method to identify progression type, followed by mixed-effects regression to reflect each individual’s rate of cognitive decline. Results are presented for population data from the south of England, and show that the currently available interventions have only modest effects at population level.