External validity of randomized controlled trials of glycaemic control and vascular disease: how representative are participants?

Authors


Correspondence to: Sarah Wild. E-mail: Sarah.Wild@ed.ac.uk

Abstract

Aims

To describe the proportion of people with Type 2 diabetes living in Scotland who meet eligibility criteria for inclusion in several large randomized controlled trials of glycaemic control to inform physicians and guideline developers about the generalizibility of trial results.

Methods

A literature review was performed to identify large trials assessing the impact of glycaemic control on risk of macrovascular disease. Inclusion and exclusion criteria from each trial were applied to data on the population of people with a diagnosis of Type 2 diabetes living in Scotland in 2008 (n = 180 590) in a population-based cross-sectional study and the number and proportion of people eligible for each trial was determined.

Results

Seven trials were identified. The proportion of people with Type 2 diabetes who met the eligibility criteria for the trials ranged from 3.5 to 50.7%. Trial participants were younger at age of diagnosis of diabetes and at time of trial recruitment than in the Scottish study population. The application of upper age criteria excluded the largest proportion of patients, with up to 39% of people with Type 2 diabetes ineligible for a trial with the most stringent criteria based on age alone.

Conclusions

We found that many of the large trials of glycaemic control among people with Type 2 diabetes have limited external validity when applied to a population-based cohort of people with Type 2 diabetes. In particular, the age distribution of trial participants often does not reflect that of people with Type 2 diabetes in a contemporary British population.

Ancillary