Does Paying Subjects to Participate in Research Improve Recruitment and Increase the Proportion of Socially Deprived and Elderly Participants?

In the UK, it is already common practice to reimburse out-of-pocket expenses eg travelling expenses and parking charges to patients participating in clinical research studies, but not to give any other incentives. This contrasts with healthy volunteer studies where financial payments are often made to encourage participation in studies.

There are few data on how payments to patients in the UK would influence the recruitment and demographic structure of patients recruited to clinical research studies.

The single biggest cost of clinical trials in the UK is the cost of recruiting suitable subjects. The Academy of Medical Sciences report on the UK clinical trials environment, "A new pathway for the regulation and governance of health research", provided a comprehensive review of the barriers faced in running clinical trials . However, other barriers to recruitment exist after these hurdles have been overcome. Importantly, in our experience, of 100 suitable subjects identified as suitable to be entered into a study only about 14% are actually randomised. Most of the identified suitable subjects do not reply to invitation letters from their doctor and some patients who do reply decline the invitation to participate. Campaigns to engage the public on the importance of research have had modest effects on improving the awareness about trials but no measurable effect on willingness to participate .

An important finding is that those subjects who do participate in clinical trials usually have much lower rates of events than predicted from observational data. This is probably because trials attract subjects who care about their own health and the health of others and these subjects exhibit good 'health behaviour'. Those subjects with less good health behaviour and who constitute the subjects who suffer more consequences from their disorders in the real world, do not participate in trials as readily. The result is that trial participants tend to be middle class, middle aged, low risk subjects. These subjects are unrepresentative of the population at large or the population who will eventually get most of the prescribing (lower class, elderly and high risk). In addition, including only low risk subjects may result in studies being abandoned because the low event rates make achieving a statistically reliable result futile.

How can we address this issue within an ethical framework?