Introduction
In this multi-centre study of the medicines management roles of informal carers, local people were recruited and trained as interviewers in each of four study areas. The aim of this paper is to characterise the teams of interviewers who participated in this study, to describe the processes of their recruitment, training, support and debriefing and to identify how these structures and processes affected research outcomes.

Method
This study was conducted in four randomly selected English health authority areas: North Oxfordshire, Coventry, North Sefton and Southall. These were selected following stratification by the proportion of elderly people, proportion of people from ethnic minority groups and socio-economic status to ensure inclusion of diverse population groups.

Potential interviewers were recruited through local media. Applications were reviewed and suitable candidates shortlisted for interview. Successful applicants attended a training day and were allocated to randomly selected pharmacies. The interviewers had two roles: identification and recruitment of the carer sample in the pharmacy, using a screening questionnaire, and to arrange and conduct semi-structured interviews with the carers and associated care-recipients in their own homes. Training was conducted in each of the areas and ongoing support provided throughout the period of data collection. The interviewers were employed on a temporary basis, invoicing the University for hours worked. They were paid a standard hourly rate for recruitment in the pharmacy and a set fee for each interview undertaken. A debriefing day was held at the end of each period to take feedback about improvements that could be made within the study.

Results
Advertisements in local newspapers were the primary method of recruitment; however, in Southall this was supplemented with local radio advertising, which was found to be more effective. The number of applications ranged from 8 in North Oxfordshire to 33 in North Sefton enabling teams of 5-7 interviewers to be appointed in each area. Each team varied by age, sex and background (eg. personal experience of being a carer, languages spoken); one team was generally younger, with a high proportion of tertiary qualifications, another was generally older with less formal qualifications. On-going support for interviewers was deemed essential to the success of the project. Throughout the project many issues emerged including ethical dilemmas, interviewer safety, and terms and conditions of employment.

Interviewers were requested to recruit and interview 10 carers. This target was achieved by 7 of the 25 interviewers across the four areas; however, 4 of these 7 conducted 15 or more. Interviewers were also asked to recruit and interview the associated care-recipients where possible. Marked variation in these success rates was also evident. During the debriefing sessions benefits were identified by interviewers in terms of personal development and community involvement, and by the project team in that processes were refined as the study progressed and study objectives were successfully achieved.

Conclusion
Local people contribute many strengths and challenges to multi-centre research. Careful planning and flexible management can ensure successful outcomes and mutual benefits to the researchers and local communities.

We would like to acknowledge the contribution of Jenny Newbould to the management of the project in North Oxfordshire.