Cost-effectiveness analysis of oral drug treatment for non-obese outpatients diagnosed with type-2 diabetes: a probabilistic sensitivity analysis

BackgroundIn many low-income to middle-income countries where there are limited economic resources, willingness to spend money is low. There is a need for analysis of the effects of small changes in willingness to pay (WTP) for health care to inform decision making in drug selection. Our goal in this study was to analyse the cost-effectiveness of oral hypoglycaemic agents (OHAs) for the initial treatment of non-obese outpatients diagnosed with type 2 diabetes in a resource-poor setting.MethodsWe performed a probabilistic sensitivity analysis to analyse the cost-effectiveness of three OHAs: metformin, glibenclamide, and acarbose. We used specialised software for programming a Markov model designed for a horizon time of 1 year considering 6 months for monotherapy and 6 months for combined oral dual therapy. Meta-analysis techniques were performed for the study of effectiveness and surveys for costs estimations. For the probabilistic sensitivity analysis we used distributions of probabilities as beta distributions and monthly costs as lognormal distributions of therapeutic alternatives. We performed a Monte Carlo simulation for a cohort of 10 000 patients for each treatment option.FindingsThe results of the Monte Carlo simulations were very close for metformin and glibenclamide, showing their strong dominance over acarbose. In the acceptability curve generated, for a WTP equal to 0 the probabilities of the therapies to be cost effective were 49·5%, 43·0%, and 7·5% for glibenclamide, metformin, and acarbose, respectively. In the glibenclamide versus metformin incremental cost-effectiveness analysis, the change of the WTP from 0 to 1 gross domestic product per capita by quality-adjusted life-year, the proportion of cost-effective iterations favouring glibenclamide increased from 53·4% to 59·7%.InterpretationThe initial drug therapies with glibenclamide or metformin are dominant over therapy with acarbose. Glibenclamide is slightly more cost effective than metformin in the treatment of non-obese outpatients diagnosed with type 2 diabetes. This kind of analysis is important for drug selection in low-income to middle-income countries.FundingNational Commission for Science and Technology supported a grant for the first author.