Economic evaluation of type 2 diabetes prevention programmes: Markov model of low- and high-intensity lifestyle programmes and metformin in participants with different categories of intermediate hyperglycaemia.
Roberts S., Craig D., Adler A., McPherson K., Greenhalgh T.
BACKGROUND: National guidance on preventing type 2 diabetes mellitus (T2DM) in the UK recommends low-intensity lifestyle interventions for individuals with intermediate categories of hyperglycaemia defined in terms of impaired fasting glucose (IFG) or 'at-risk' levels of HbA1c. In a recent systematic review of economic evaluations of such interventions, most studies had evaluated intensive trial-based lifestyle programmes in participants with impaired glucose tolerance (IGT). This study examines the costs and effects of different intensity lifestyle programmes and metformin in participants with different categories of intermediate hyperglycaemia. METHODS: We developed a decision tree and Markov model (50-year horizon) to compare four approaches, namely (1) a low-intensity lifestyle programme based on current NICE guidance, (2) a high-intensity lifestyle programme based on the US Diabetes Prevention Program, (3) metformin, and (4) no intervention, modelled for three different types of intermediate hyperglycaemia (IFG, IGT and HbA1c). A health system perspective was adopted and incremental analysis undertaken at an individual and population-wide level, taking England as a case study. RESULTS: Low-intensity lifestyle programmes were the most cost-effective (£44/QALY, £195/QALY and £186/QALY compared to no intervention in IGT, IFG and HbA1c, respectively). Intensive lifestyle interventions were also cost-effective compared to no intervention (£2775/QALY, £6820/QALY and £7376/QALY, respectively, in IGT, IFG and HbA1c). Metformin was cost-effective relative to no intervention (£5224/QALY, £6842/QALY and £372/QALY in IGT, IFG and HbA1c, respectively), but was only cost-effective relative to other treatments in participants identified with HbA1c. At a willingness-to-pay threshold of £20,000/QALY, low- and high-intensity lifestyle programmes were cost-effective 98%, 99% and 98% and 81%, 81% and 71% of the time in IGT, IFG and HbA1c, respectively. An England-wide programme for 50-59 year olds could reduce T2DM incidence by < 3.5% over 50 years and would cost 0.2-5.2% of the current diabetes budget for 2-9 years. DISCUSSION: This analysis suggests that current English national policy of low-intensity lifestyle programmes in participants with IFG or HbA1c will be cost-effective and have the most favourable budget impact, but will prevent only a fraction of cases of T2DM. Additional approaches to prevention need to be investigated urgently.