Introduction
Randomised controlled trials have been seen in recent decades as the ‘gold standard’ for clinical research.173 They are accepted as generally the best way to estimate the effectiveness and cost-effectiveness of interventions, as they make strong causal connections between interventions and their effects.174 However more than two thirds of published trials do not achieve their recruitment targets.175 McDonald et al. looked at 114 trials funded by the MRC or the NIHR HTA programme and found that 54% required an extension, while 41% experienced delays in starting recruitment.176 This has large cost implications for funders, resulting in trial extensions, and even complete failure. Why some trials recruit well and other suffer problems remains unclear.177
Several systematic reviews have sought to improve recruitment in trials. Prescott et al.177 attempted to identify factors that affected the effective running of trials, suggesting that recruitment problems may be reduced by piloting, using multiple recruitment strategies, making contingency plans for slow recruitment and using recruitment coordinators. However none of these approaches had been rigorously evaluated. Watson and Torgerson concluded that recruitment interventions were both sparse and often of poor methodological quality.178
Campbell et al. set out to identify factors associated with good and poor recruitment in multi-centred trials from a cohort of studies, a selection of case studies and a single in-depth case study.179 The main themes associated with success were flexibility, adaptability to unexpected issues and better training. The study suggested that the complex nature of multi-centred trials generated unexpected difficulties.
Recent methods to improve recruitment have achieved some success. One such method is the business model using marketing strategies; however these require further research to establish effectiveness and develop useable tools for medical research where these approaches, concepts and terminology are unfamiliar.180
Barnard et al. looked at the recruitment of participants into trials from a different perspective and aimed to identify different models that may be useful to RCTs when estimating the recruitment of participants.181 They noted that most trials use an unconditional model of recruitment and suggested that a new model was needed to predict recruitment to clinical trials which takes account of both centre and patient recruitment, recognising that one drives the other.
Health services research often recruits from primary care. However Bower et al.,175 exploring recruitment difficulties, responses to recruitment problems and the relationship between trial characteristics and recruitment, found that recruitment methods requiring GPs to consent patients into trials were particularly problematic. In a review of current literature, Bower also concludes that recruitment of patients into health research from primary care continues to be a major hurdle.182
Goodyear-Smith et al. sought to identify barriers to recruitment in primary care,183 including lack of time (exacerbated by the annual influenza vaccinations campaign in general practice), the need to identify staff responsible for decision making, the need to clarify the nature of the study, and the need to be flexible in accommodating practices. Strategies to improve recruitment included providing incentives to practices (both material and educational), using a personal approach, ensuring practices feel engaged, minimising disruption, streamlining processes, and using doctors to recruit doctors. They also suggested that smaller practices were easier to recruit than larger practices.
In the UK in recent years, national research networks have tried to improve some of the difficulties in recruiting to clinical trials and running them. These include specialised networks in key areas of NHS research such as cancer, mental health and stroke. Overarching networks also support research in areas like primary care and children’s research. The aim of these research networks is to improve the process of regulatory approvals, provide infrastructure and research support to trials and facilitate the recruitment of patients into trials.184 In Wales the NISCHR Clinical Research Centre (NISCHR CRC) was established in 2010 to provide research workforce to support and develop research activity within health and social care.185 However few reported trials have assessed the effectiveness of research networks. Those that have report that they have been successful in improving recruitment rates into trials.185–187
So we aimed to identify barriers to recruitment and factors that facilitated recruitment in the FolATED trial.
Recruitment methods
We recruited participants over three years in three centres across Wales. Recruitment methods included direct referrals from GPs, psychiatrists and other healthcare professionals, clinical database searches and self-referral. We reported the number of participants recruited by each method in each centre to identify differences.
Recruitment into the trial from primary care was facilitated by research staff employed directly by the trial or by NISCHR CRC. Their Clinical Studies Officers played a pivotal role in recruitment, including practice recruitment through visits and presentations, development and editing of regular newsletters to recruiting practices, and contributing to local and national meetings and conferences. NISCHR CRC staff also helped with computer searches in practices and invitation mailings from practices.
We collected data from each centre on recruitment, including whether from secondary care, general practices or community mental health services, and method of recruitment, for example whether participants were directly referred by healthcare professionals or as a result of invitations from computer searches in practices.
We sought qualitative feedback from centres through a reflective recruitment tool providing written accounts of recruitment strategies, barriers and facilitators in each site (). We analysed these data using a thematic approach. We also held monthly recruitment meetings and annual training events to identify recruitment difficulties and develop new strategies.
Recruitment strategies: questions for qualitative analysis
Recruitment performance
We used a wide variety of strategies to help with the recruitment and retention of both recruiters to the trial such as general practices and secondary care services, and trial participants (). We adopted a flexible approach to recruitment allowing the three centres to identify and deploy the strategies found to be successful in their area. However some strategies were universal such as the use of trial posters, payment of general practices, feedback to GPs of BDI-II scores, regular training events and research team meetings. Local variations included use of research networks such as NISCHR CRC, access to direct referrals to secondary care through flagging in referral notes, different methods of referral and participant reminders for appointments and venues.
Strategies for maintaining recruitment and retention
Methods used in recruitment
We recruited participants in three centres – North East Wales, North West Wales and Swansea (). Although several methods of recruitment were adopted by all centres, different methods were used in different centres. North West Wales, the most successful recruiting centre, acquired their referrals mainly from direct referrals to the psychiatric services, supplemented by computer searches at general practices and other referrals through secondary care. North East Wales acquired the majority of their referrals through computer searches at general practices and direct GP referrals supplemented by other mental health referrals. Swansea acquired the majority of their referrals through direct GP referrals, supplementing this with computer searches and other referrals from secondary care ().
Referral method by recruitment centre
Number and type of referrals.
Facilitators to recruitment
Although we adopted different methods of recruitment across sites, we identified several factors as aiding recruitment. A major facilitator was a good relationship with recruiting practices and services. This was seen as a key factor in North West Wales: they had a history of working with general practices and had established good working relationships with them. Elements important in building and maintaining a good relationship included regular contact with practices, fast responses to their queries, regular updates by newsletters and personal contact, and small tokens of appreciation like cards at Christmas.
A perception that participation conferred benefits on patients was also seen as a facilitator. Many GPs were pleased to offer something extra to long-term patients and saw quick access to a psychiatrist as beneficial, and time with the researcher as an alternative or supplement to counselling, which was often difficult to access. Some GPs felt this would be an alternative to patients visiting the surgery.
Relationships were a recurrent theme. A good patient–doctor relationship was viewed as a facilitator to direct referrals, as small surgeries had higher referral rates. Good relationships between participants and researchers were also seen as important in retaining participants. Follow-ups were face to face, helping relationships to develop. Continuity also contributed to the success of relationships with researchers encouraged to follow participants from beginning to end. Also important was personal contact through reminder letters and phone calls.
Awareness raising and networking with gatekeepers like practice managers were important in gaining access to recruiters. Initial contact with Local Health Boards was also seen as a facilitator by providing advice about key contacts, monthly meetings and identifying research-active practices. Furthermore aides-memoire facilitated recruitment by providing reminders to health professionals; these included pens, note paper and trolley tokens with the FolATED logo.
Barriers to recruitment
One common theme was the competing demands on general practice time and resources. Examples of commitments that preoccupied practices included preparing for QOF and the contemporaneous vaccination campaign. Also changes in staff and surgery relocation affected recruitment in some centres.
Barriers included perceptions that research might compromise practice, including concerns about increased visits to GPs, access to the practice database, confidentiality, disruption, expense, inappropriate referrals, consequences of participation for patients, in particular those found not suitable at screening, increased workload, availability of space in surgeries and even the validity of the study.
North East Wales reported that some practices were already taking part in other studies and did not feel they could recruit to another study. Also previous demands by researchers for information had made several practices reluctant to participate, fearing our study was connected in some way.
Staffing problems also reduced recruitment. The Swansea centre lost several research staff within a short time, thus halting participant recruitment during staff recruitment and training. North East Wales had only one researcher for several months before NISCHR CRC could help. North West Wales also had to restrict computer searches and mailings to focus on direct referrals.
Inclement weather also affected recruitment and retention during the winter of 2010. Participants and researchers were often unable to attend appointments, particularly in rural areas.
Discussion
This study has identified facilitators of, and barriers to, recruitment and retention in the FolATED trial. Facilitators included the importance of building good relationships with psychiatric services, practices and participants through the interpersonal skills of researchers and continuing feedback to recruiters. Also important was the potential benefit of the intervention, and participation in the trial, to patients, as perceived both by those recruiting and the participants themselves. Thus raising awareness and networking with gatekeepers like practice managers reportedly improved recruitment.
Barriers included the high demand on practitioners’ time, disruption to the surgery, consequences for ineligible patients, and worries about confidentiality. It is therefore vital that researchers design trials that minimise impact on surgeries and provide reassurance about effects on patients and their confidentiality. It is important that, when dealing with recruiters, researchers tread carefully. If future research is to be successful they need to nurture relationships and not make undue demands. As staffing levels often posed a threat to recruitment, researchers and funders need to be realistic when designing trials. Finally the FolATED trial experienced unforeseen disruption, notably from the volcano eruption in Iceland and the inclement weather of the winter of 2010.
This study also examined the strategies and methods used in the FolATED trial for the recruitment and retention of both those recruiting into the trial and participants in the trial. Recommendations include the need for a flexible approach to such recruitment and retention. An explicit recruitment strategy is essential in any trial; however multi-centred complex interventions present extra challenges which require a flexible and often creative approach. However we suffered from the slow regulatory systems and long waits to implement much needed changes. For example we waited several months to get approval for an improved poster and information leaflet for GP surgeries.
Monthly research team meetings provided the platform for sharing problems and exploring ideas with colleagues. This also encouraged the trial centres to work as a team. Annual training sessions also assisted in exploring and sharing ideas and experiences across centres.
The resources required to recruit and run a complex intervention should not be underestimated. There is a need to acknowledge that recruiting into trials and the day-to-day running of trials is laborous, and that the recruitment phase of a trial is particularly so. Alongside this, trials that employ labour-intensive methods like interviews must be costed accordingly. The competitive nature of bidding for trial funding often results in underfunded studies risking failure of trials owing to a lack of researchers on the ground to coordinate and perform the research. In the FolATED trial assistance from NHS staff and the emerging NISCHR CRC was crucial in centres with limited resources, where even the most successful recruiting centre could have handled more participants but for the limitation of research staff to provide follow-up interviews.
