Edinburgh Napier University

  Background
Recruitment to randomised controlled trials (RCTs) is a perennial problem. Calls have been made for trialists to make recruitment performance publicly available. This article presents our experience of recruiting to a pilot RCT of cardiac rehabilitation for patients with bowel cancer with an embedded process evaluation.

Methods
Recruitment took place at three UK hospitals. Recruitment figures were based on the following: i) estimated number of patient admissions, ii) number of patients likely to meet inclusion criteria from clinician input and iii) recruitment rates in previous studies. The following recruitment procedure was used:
1.
Nurse assessed patients for eligibility.

2.
Patients signed a screening form indicating interest in and agreement to be approached by a researcher about the study.

3.
An appointment was made at which the patient signed a consent form and was randomised to the intervention or control group.

Information about all patients considered for the study and subsequently included or excluded at each stage of the recruitment process and reasons given were recorded.
Results
There were variations in the time taken to award Research Management approval to run the study at the three sites (45–359 days). Sixty-two percent of the original recruitment estimate was reached. The main reason for under-recruitment was due to over-estimation of the number of patient admissions; other reasons were i) not assessing all patients for eligibility, ii) not completing a screening form for eligible patients and iii) patients who signed a screening form being lost to the study before consenting and randomisation.

Conclusions
Pilot trials should not simply aim to improve recruitment estimates but should also identify factors likely to influence recruitment performance in a future trial and inform the development of that trial’s recruitment strategies. Pilot trials are a crucial part of RCT design. Nevertheless, pilot trials are likely to be small scale, involving only a small number of sites, and contextual differences between sites are likely to impact recruitment performance in any future trial. This means that ongoing monitoring and evaluation in trials are likely to be required.