Calibrating Cardiovascular Disease Policy Model Using Large Cohort Data
Author(s)
Wu R1, Williams C2, Schlackow I2, Zhou J2, Emberson J2, Reith C2, Keech A3, Robson J1, Wilkinson K4, Armitage J2, Collins R2, Gray A2, Simes J3, Baigent C2, Mihaylova B2
1Queen Mary University of London, LONDON, LON, UK, 2University of Oxford, Oxford, UK, 3University of Sydney, Sydney, NSW, Australia, 4Oxfordshire, Oxford, UK
Presentation Documents
OBJECTIVES : Declining cardiovascular disease (CVD) morbidity and mortality over the last decades suggest that policy models developed using historic data likely overestimate current risks. We illustrate an approach to calibrating a micro-simulation CVD model, developed using historic clinical trials data, using the 2006-2020 UK Biobank cohort. METHODS : A micro-simulation model, developed using individual participant data from the Cholesterol Treatment Trialists’ collaboration (CTT: 118,000 participants), was calibrated in the UK Biobank cohort (UKB: 502,000 participants). Original parametric survival models estimated risks of key endpoints (myocardial infarction, stroke, coronary revascularisation, incident cancer and vascular and nonvascular death) using CTT participants’ sociodemographic and clinical characteristics at entry and incidents of the key endpoints during follow-up (significant at p<0.01). Model calibration using UKB data involved re-fitting the intercept and linear predictors from these equations; excluding/de-novo estimating factors with substantially different effects; and adding factors and endpoint/s to enrich the model. Standard approaches to risk equation modelling and model validation were employed. Both original and calibrated models were used to predict endpoints in UKB. RESULTS : We demonstrate the feasibility of calibrating a detailed decision analytic model using individual participant data. However, it is a data intensive and cumbersome process likely to lead to modified strength of relationships between disease endpoints. The process required all equations to be calibrated with few factors de-novo estimated (e.g. smoking) but enabled new factors (e.g. cancer duration) and endpoint (incident diabetes) to be included. The incidence of endpoints predicted by the calibrated model, unlike in the original model, corresponded well to their observed incidence in the UKB, overall and in categories of participants. CONCLUSIONS : A new calibrated lifetime CVD model accurately predicts morbidity and mortality in contemporary UK populations. It will be made available to provide individualised projections of expected lifetime health outcomes and benefits of treatment.
Conference/Value in Health Info
2021-11, ISPOR Europe 2021, Copenhagen, Denmark
Value in Health, Volume 24, Issue 12, S2 (December 2021)
Acceptance Code
P70
Topic
Economic Evaluation, Epidemiology & Public Health, Methodological & Statistical Research
Topic Subcategory
Cost-comparison, Effectiveness, Utility, Benefit Analysis
Disease
Cardiovascular Disorders