Introduction
The Birmingham Repeat Dispensing Pilot was one of the Pharmacy Wider Role Projects funded by the Department of Health. It involved 350 patients from two medical practices and the 7 pharmacies most closely associated with them. One practice was based in an inner city area of low socio-economic status and one in a residential outer city suburb. The repeat dispensing phase ran for six months. This paper focuses upon the views of patients and GPs who were involved in the project obtained by interview after the completion of the project.

Method
After completion of the intervention, structured interviews were conducted with 206 patients and 16 carers. The patient interviews covered experience and attitudes of the intervention. Semi-structured interviews with the lead GPs from each of the practices covered their experiences of the project. They were tape recorded and transcribed for analysis.

Results
The majority of patients who completed the six months the liked the repeat dispensing system (86%); the main reason given was convenience. A small sub-group were dissatisfied. Overall 15% of patients were lost during the six-month study with a noticeably greater loss from the inner city practice. Patients withdrawing included both those on stable long-term repeats and those where medication changes were taking place. There was some evidence that communications in the pharmacy improved - 96% of patients felt that service was more friendly and personal. However, 93% still did not know the name of their pharmacist and 96% felt that the chemist had made no change to their medication. In the interviews with GPs clear differences emerged between the two practices which help to explain the outcome. Attitudes to working through the project process were different. At the onset both practices had stated that they ran a formal repeat prescribing system but experience revealed that the inner city practice had a very informal system and had been unable to deal with changes to prescriptions or with the protocols required by the project. The inner city GP said that patients withdrew from the project, it seemed to be the easiest solution to solve their process problems. Nobody on the practice side had ownership of the project. By comparison in the suburban practice the repeat prescribing system was highly organised and all partners conformed to its use: few patients withdrew from the project. GPs from both practices noted that there had been no improvement in communications between themselves and the pharmacists. In the inner city practice, contact prior to the project was minimal and the GP was sceptical of commercial competitiveness between pharmacists. In the suburban practice there were good links prior to the project and the GP was genuinely disappointed by the failure to develop these.

Discussion
The repeat dispensing process was generally liked by patients but is not suitable for all repeat patients. In spite of a project framework developed to encourage communication, the overall level of interaction between the pharmacists and patients and pharmacist and GPs was disappointing. Based upon our findings, we suggest that the following are essential requirements for optimal operation of a repeat dispensing system: careful selection of the patient group, in the GP practice a well developed repeat prescribing systems and a commitment to repeat dispensing and in the pharmacy a good relationship between the pharmacy and the practice prior to the project and a commitment from the pharmacist to develop communications with both patients and prescribers.