Objectives-To measure the processes of care, health benefits and costs of outreach clinics held by hospital specialists in primary care settings. Design-The study was designed as a case-referent (comparative) study in which the features of 19 outreach clinics (cases) were compared with matched outpatient clinics (controls). The measuring instruments were self administered questionnaires. Patients were followed up at six months to reassess health status. The specialties included in the study were cardiology, ENT, general medicine, general surgery, gynaecology and rheumatology. Setting-Specialist outreach clinics in general practice in England, with matched outpatient clinic controls. Subjects-Consecutive patient attenders in the outreach and outpatient clinics, their specialists, the outreach patients' general practitioners, practice managers and trust accountants. Patients' response rate at baseline: 78% (1420). Main outcome measures-Patient satisfaction, doctors' attitudes, processes and health outcomes, costs. Results-Outreach patients were more satisfied with the processes of their care than outpatients, their access to specialist care was better than that for outpatients and they were more likely to be discharged. Doctors reported that the main advantages of the outreach clinic were improved patient access to specialists and convenience for patients, in comparison with outpatients, and most GPs and specialists felt the outreach clinic was "worthwhile". At six month follow up, the health status of the outreach sample had significantly improved more than that of the outpatients on all eight sub-scales of the HSQ-12, but this was probably because of their better starting point at baseline. The impact of outreach on health outcomes was small. The NHS costs of outreach were significantly higher than outpatients. An increase in outreach clinic size would reduce cost per patient, but would lead to the loss of most of the clinics' benefits. Conclusions-While the process of care was of higher quality in outreach than in outpatients, and the eYciency of care was also greater in the latter, the eVect on patients' health outcomes was small. Responsiveness to patients' views and preferences is an essential component of good quality service provision. However, the greater cost of outreach raises the issue of whether improvements in the quality and eYciency of health care, without a substantial impact on health outcomes, is money well spent in a publicly funded health service. On the other hand, the real costs of outreach in comparison with outpatients clinics can probably only be truly estimated in a longitudinal study with a resource based costing model derived from documented patient attendances and treatment costs over time in relation to longer term outcome (for example, at a two year end point).