The potential impact of different risk screening methods on cardiovascular disease in New Zealand: a simulation study of policy options

Abstract

Preface This thesis was inspired by a UK simulation study which concluded that everyone in the UK should take cardiovascular disease (CVD) preventive medications from age 55 years. The authors compared screening strategies that determined eligibility for preventive medications based on age-alone or on predicted CVD risk. They reported that age-alone screening was simpler and more cost-effective. Their conclusion was highly controversial as guidelines increasingly recommend more personalised CVD risk screening using multivariable prediction equations. The goal of this thesis was to undertake a New Zealand (NZ) version of the UK study, accounting for NZ-specific factors, and addressing a number of methodological issues and assumptions. This involved: (i) creating a synthetic NZ population based on the 2013 NZ population of 2,451,278 individuals 30-84 years, each with a CVD risk factor profile, which replaced the hypothetical UK study population; (ii) replacing the Framingham Study CVD risk prediction equation used in the UK study (and previously in NZ) with a newly developed NZ equation; and (iii) undertaking simulations to compare benefits, harms and cost-effectiveness of screening strategies based on age-alone versus predicted CVD risk. The candidate's simulations used NZ screening and medication costs, accounted for medication adverse effects, disutility and adherence not considered in the UK study, and used cost per quality adjusted life year gained instead of the UK study's cost per CVD-free life year outcome measure. Comparing the NZ population's CVD risk profile based on the Framingham and NZ risk equations demonstrated that the population eligible for preventive medications would fall by 80% using the NZ equations unless treatment thresholds were also lowered. Based on these findings, the Ministry of Health has now recommended lowering the treatment threshold. The thesis simulations broadly supported the UK study conclusions, if it was assumed that the screening strategy would not affect medication adherence. However, when adherence using the age-alone strategy was reduced by 5%, any potential benefits of this strategy were lost. As adherence is likely to be better when drugs are prescribed by GPs after comprehensive risk assessments and discussion, the candidate's findings call into question the UK study conclusions.