Trials within trials? Researcher, funder and ethical perspectives on the practicality and acceptability of nesting trials of recruitment methods in existing primary care trials

Trials frequently encounter difficulties in recruitment, but evidence on effective recruitment methods in primary care is sparse. A robust test of recruitment methods involves comparing alternative methods using a randomized trial, 'nested' in an ongoing 'host' trial. There are potential scientific, logistical and ethical obstacles to such studies.

Telephone interviews were undertaken with four groups of stakeholders (funders, principal investigators, trial managers and ethics committee chairs) to explore their views on the practicality and acceptability of undertaking nested trials of recruitment methods. These semi-structured interviews were transcribed and analysed thematically.

Twenty people were interviewed. Respondents were familiar with recruitment difficulties in primary care and recognised the case for 'nested' studies to build an evidence base on effective recruitment strategies. However, enthusiasm for this global aim was tempered by the challenges of implementation. Challenges for host studies included increasing complexity and management burden; compatibility between the host and nested study; and the impact of the nested study on trial design and relationships with collaborators. For nested recruitment studies, there were concerns that host study investigators might have strong preferences, limiting the nested study investigators' control over their research, and also concerns about sample size which might limit statistical power. Nested studies needed to be compatible with the main trial and should be planned from the outset. Good communication and adequate resources were seen as important.

Although research on recruitment was welcomed in principle, the issue of which study had control of key decisions emerged as critical. To address this concern, it appeared important to align the interests of both host and nested studies and to reduce the burden of hosting a recruitment trial. These findings should prove useful in devising a programme of research involving nested studies of recruitment interventions.