Improving recruitment for clinical trials: the human touch

Personal contact is crucial for encouraging patients to participate

Starting recruitment is an important milestone in the life cycle of a randomised clinical trial. It often marks the end of months or even years of planning, and is an exciting time for the investigators. However, despite the best planning, recruitment for clinical trials is never easy, and can be difficult for a multitude of reasons. Some relate to the trial itself; the inclusion and exclusion criteria, although intellectually robust, may prevent any patients from ever being eligible. Recruitment may be difficult because of trouble accessing the required types of patients; if the trial is inadequately funded, investigators may be unwilling to recruit participants, or they may be inexperienced regarding the intervention or the population being investigated. Patient‐related factors also determine participation rates: do the patients understand the research question and protocol, is it onerous for them, what are the perceived risks? All of these factors work in a complex interplay to determine the success of trial recruitment strategies.1,2,3

In this issue of the Journal, the investigators of the Evaluating recruitment strategies in the Australian Study for the Prevention through Immunisation of Cardiovascular Events (AUSPICE)4 report their experience in recruiting patients for a large pragmatic trial that compared the effectiveness of pneumococcal vaccine with placebo. They found that the participation rate in response to a mail‐out was only 3%, far short of the expected 10%. With reminders and advertising in traditional and social media, they managed to improve this rate somewhat, but it remained low. Why was the final participation rate so low? A sufficiently large patient population was identified in electronic health records for targeted recruitment; the study design was simple and relatively easy to understand; the risk to participants was low, as was the burden of the study visits. All of these features are important factors in determining patient participation in randomised trials.1,2,3 Large pragmatic trials such as AUSPICE are regarded as good solutions for the problem of complexity in clinical trials with respect to both the burden on participants and cost.5 So if these design factors did not seem to be barriers, perhaps it was a patient‐related factor? Could it have been the method of patient contact?

Several studies have found that face‐to‐face contact is an important factor in recruitment.1,2 In an analysis of 87 protocols for clinical trials and studies in children, two‐thirds of which reached at least 80% of their target enrolment, one of the biggest determinants of successful recruitment was whether investigators approached prospective participants in person; 86% of the trials that reached their target had included personal approaches, compared with 70% of those that did not reach target enrolment.6 Study complexity and long duration and complex inclusion and exclusion criteria were the only significant predictors of not reaching the recruitment target.

Patients rate their experience of research more highly when they have had a chance to engage with it and discuss it with investigators.7 It is therefore unsurprising that the AUSPICE patients did not respond as well as expected to an invitation letter or to advertising. While these relatively inexpensive methods are attractive because they can be designed to target a large population, investigators should be aware that response rates will probably be low, although reminders can help, as found by the AUSPICE investigators.

If personal contact is better, are there any specific aspects that patients respond to? It seems that recruiter flexibility and building rapport are two key qualities.8 Patients are also more likely to respond if they already have a relationship with the study team or if the study team is introduced by a member of the patient's clinical team.2,9 This personal contact appears to be crucial in successful recruitment strategies. In contrast to non‐selective methods such as postal invitations, however, it must be remembered that in‐person recruitment may be subject to bias (both conscious and unconscious) that can influence both recruitment and the representativeness of the participants.

If we are to improve rates of recruitment for clinical trials, we cannot forget that at the centre of this effort is a patient who is being asked to entrust their health and potentially their life to an investigator and their team, often without expecting any direct personal benefit. This requires trust and confidence, both of which are difficult to cultivate from a distance. While large scale electronic datasets and traditional and social media have transformed many aspects of clinical trials, successful recruitment still requires the human touch.