The cost-effectiveness of testing strategies for type 2 diabetes: a modelling study

Abstract

BackgroundAn estimated 850,000 people have diabetes without knowing it and as many as 7 million more are at high risk of developing it. Within the NHS Health Checks programme, blood glucose testing can be undertaken using a fasting plasma glucose (FPG) or a glycated haemoglobin (HbA1c) test but the relative cost-effectiveness of these is unknown.ObjectivesTo estimate and compare the cost-effectiveness of screening for type 2 diabetes using a HbA1ctest versus a FPG test. In addition, to compare the use of a random capillary glucose (RCG) test versus a non-invasive risk score to prioritise individuals who should undertake a HbA1cor FPG test.DesignCost-effectiveness analysis using the Sheffield Type 2 Diabetes Model to model lifetime incidence of complications, costs and health benefits of screening.SettingEngland; population in the 40–74-years age range eligible for a NHS health check.Data sourcesThe Leicester Ethnic Atherosclerosis and Diabetes Risk (LEADER) data set was used to analyse prevalence and screening outcomes for a multiethnic population. Alternative prevalence rates were obtained from the literature or through personal communication.Methods(1) Modelling of screening pathways to determine the cost per case detected followed by long-term modelling of glucose progression and complications associated with hyperglycaemia; and (2) calculation of the costs and health-related quality of life arising from complications and calculation of overall cost per quality-adjusted life-year (QALY), net monetary benefit and the likelihood of cost-effectiveness.ResultsBased on the LEADER data set from a multiethnic population, the results indicate that screening using a HbA1ctest is more cost-effective than using a FPG. For National Institute for Health and Care Excellence (NICE)-recommended screening strategies, HbA1cleads to a cost saving of £12 and a QALY gain of 0.0220 per person when a risk score is used as a prescreen. With no prescreen, the cost saving is £30 with a QALY gain of 0.0224. Probabilistic sensitivity analysis indicates that the likelihood of HbA1cbeing more cost-effective than FPG is 98% and 95% with and without a risk score, respectively. One-way sensitivity analyses indicate that the results based on prevalence in the LEADER data set are insensitive to a variety of alternative assumptions. However, where a region of the country has a very different joint HbA1cand FPG distribution from the LEADER data set such that a FPG test yields a much higher prevalence of high-risk cases relative to HbA1c, FPG may be more cost-effective. The degree to which the FPG-based prevalence would have to be higher depends very much on the uncertain relative uptake rates of the two tests. Using a risk score such as the Leicester Practice Database Score (LPDS) appears to be more cost-effective than using a RCG test to identify individuals with the highest risk of diabetes who should undergo blood testing.LimitationsWe did not include rescreening because there was an absence of required relevant evidence.ConclusionsBased on the multiethnic LEADER population, among individuals currently attending NHS Health Checks, it is more cost-effective to screen for diabetes using a HbA1ctest than using a FPG test. However, in some localities, the prevalence of diabetes and high risk of diabetes may be higher for FPG relative to HbA1cthan in the LEADER cohort. In such cases, whether or not it still holds that HbA1cis likely to be more cost-effective than FPG depends on the relative uptake rates for HbA1cand FPG. Use of the LPDS appears to be more cost-effective than a RCG test for prescreening.FundingThe National Institute for Health Research Health Technology Assessment programme.