Clinical trial data is used to promote evidence based management. Ideally, clinical trial participants (CTP) should be representative of patients in clinical practice but rarely has this been tested. In this study, we examined demographic, clinical and case fatality data of all Diabetes Centre CTP (n = 444). Overall, the proportion of males who took part in clinical trials was higher than the Diabetes Centre population (68% vs 56%; X² = 25.3, p<0.0001). We matched CTP 1:5 with non clinical trial participants (NCTP) (n = 2220) for age, gender, type and duration of diabetes according to the year the trial was conducted. CTP were more likely to be anglo-celtic, less likely to smoke, have lower case fatality and die at an older age (Table 1).
To overcome the possibility that CTP had a less severe form of diabetes, we further matched a subgroup of CTP with comprehensive clinical data (n = 247) with NCTP (n = 1720) for glycaemic control, microvascular and macrovascular complications. Whilst case fatality rates were not significantly different (10.1 vs 14.3%; p=0.07), CTP lived longer (73.4 vs 68.3 yrs; p=0.01). In these cohorts, CTP had similar blood pressure and higher cholesterol than NCTP but were on less blood pressure and lipid treatment. Whether this is due to a perceived lower cardiovascular risk by the treating health care professional because of non-smoking status is unclear.
Our findings indicate that the applicability of clinical trial results to the diabetic population at large could be altered if the efficacy of the treatment modality being tested is affected by gender, ethnicity or smoking status, which are not normally considered as criteria in medication eligibility. These subtle differences of the clinical trial population may distort the calculation of cost benefit ratio in making a drug available to the general community.