Presentation (CTI)

OBJECTIVES: To develop a cardiometabolic disease policy model capable of projecting long-term health, economic, and inequality outcomes in a UK setting.
METHODS: A cohort of 184,845 individuals from the Clinical Practice Research Datalink (CPRD) Aurum was analysed, with linkage to Hospital Episode Statistics (HES), mortality records, and a deprivation index. A multi-state parametric survival model with a semi-Markov structure was applied to estimate transition-specific hazards, state occupancy probabilities, and transition probabilities, capturing time-dependent risks and enabling extrapolation. Model diagnostics and performance evaluations were conducted to assess model validity.
RESULTS: The model comprises seven health states: disease-free, type 2 diabetes mellitus (T2DM), myocardial infarction (MI), post-MI, stroke, post-stroke, and death, linked by 13 clinically relevant transitions. Standard parametric models produced more abrupt transitions with earlier exits from the disease-free state, offering computational simplicity. In comparison, flexible parametric models yielded smoother transitions and delayed onset of T2DM, MI, stroke, and death, better reflecting real-world disease patterns.The model operates as a hybrid of cohort and microsimulation approaches. Like a cohort model, it estimates state occupancy over time and aggregates outcomes such as life expectancy and disease prevalence. Reflecting microsimulation features, it supports individual-level risk stratification using covariates (e.g., age, sex, biomarkers), allowing transition probabilities to vary across individuals. Hazards are converted into time-dependent transition probabilities, enabling discrete-time simulation of disease trajectories.
CONCLUSIONS: By leveraging UK real world complex and linked data as well as by applying robust multi-state parametric survival modelling, the model provides: a) a transparent, flexible platform for simulating cardiometabolic disease progression, and b) evaluation of long-term health, health economics, and impact on inequalities of preventative interventions.