The validity of cost-effectiveness analyses of tight glycemic control. A systematic survey of economic evaluations of pharmacological interventions in patients with type 2 diabetes

Purpose Currently available randomized trial evidence has shown no reductions in type 2 diabetes (T2D) complications important to patients with tight glycemic control. Yet, economic analyses consistently find tight glycemic control to be cost-effective. To understand this apparent paradox, we systematically identified and appraised economic analyses of tight glycemic control for T2D. Methods We searched multiple databases from January 2016 to January 2018 for cost-effectiveness or cost-utility analyses of any glucose-lowering treatments for adults with T2D using simulations with long-40 years to lifetime-time horizons. Reviewers selected and appraised each study independently and in duplicate with good reproducibility. Results We found 30 analyses, most comparing the glycemic impact of glucose-lowering drugs and applying their impact on HbA1c to model (most commonly IMS CORE or Cardiff T2DM) their impact on the incidence of diabetes-related complication. Models drew from observational evidence of the correlation of HbA1c levels and diabetes-related complication rates; none used estimates of the effect of lowering HbA1c on these outcomes from systematic reviews of randomized trials. Sensitivity analyses, when conducted, demonstrate substantial loss of cost-effectiveness as simulations approach the results seen in these trials. Conclusions Reliance on the association between glycemic control and diabetes-related complications evident in observational studies but not apparent in randomized trial bias the estimates of the cost-effectiveness of interventions to improve glycemic control.

Automated summary

The current randomized trial evidence does not show significant reductions in type 2 diabetes complications with tight glycemic control, but economic analyses suggest it is cost-effective. The reliance on observational studies linking glycemic control to diabetes-related complications, without incorporating systematic review estimates from randomized trials, may bias the cost-effectiveness estimates of interventions to improve glycemic control.