Frail older people have been systematically excluded from randomised controlled trials (RCT). We aim to recruit older, frail hospitalised patients in an RCT and evaluate the frailty index (FI) as a measure to describe the types of people included in the study. We recruited 315 hospitalised patients aged 65 years; age ranged from 60 to 102 years. Baseline assessment scores ranged as follow: Mini-Mental Status Examination from 7 to 30, Barthel index from 5 to 100 and FI from 2 to 24. Total deaths were 20 (6%). We demonstrated that it is feasible to recruit frail older people into RCTs. The FI does not show any 'floor' or 'ceiling' effects. We can measure frailty in an RCT cohort, and we believe that clinical trials should include more frail older people and that the use of an FI can facilitate such trials and generate reliable data to guide future medical practice in a rapidly ageing society.