Abstract
BACKGROUND: A criticism of Randomized Controlled Trials (RCTs) in primary care is that they lack external validity, participants being unrepresentative of the wider population. Our aim was to determine whether published primary care-based RCTs report information about how the study sample is assembled, and whether this is associated with RCT characteristics. METHODS: We reviewed RCTs published in four primary care journals in the years 2001-2004. Main outcomes were: (1) eligibility fraction (proportion eligible of those screened), (2) enrolment fraction (proportion randomised of those eligible), (3) recruitment fraction (proportion of potential participants actually randomised), and (4) number of patients needed to be screened (NNS) in order to randomize one participant. RESULTS: A total of 148 RCTs were reviewed. One hundred and three trials (70%) reported the number of individuals assessed by investigators for eligibility, 119 (80%) reported the number eligible for participation, and all reported the actual number recruited. The median eligibility fraction was 83% (IQR 40% to 100%), and the median enrolment fraction was 74% (IQR 49% to 92%). The median NNS was 2.43, with some trials reportedly recruiting every patient or practice screened for eligibility, and one trial screening 484 for each patient recruited. We found no association between NNS and journal, trial size, multi- or single-centre, funding source or type of intervention. There may be associations between provision of sufficient recruitment data for the calculation of NNS and funding source and type of intervention. CONCLUSION: RCTs reporting recruitment data in primary care suggest that once screened for eligibility and found to match inclusion criteria patients are likely to be randomized. This finding needs to be treated with caution as it may represent inadequate identification or reporting of the eligible population. A substantial minority of RCTs did not provide sufficient information about the patient recruitment process.