The International Clinical Epidemiology Network (INCLEN) was established in 1982 to strengthen the research capacity of medical schools in the developing world through the development of Clinical Epidemiology Units (CEUs). The role of these units is to promote a rational approach to clinical and health care decision making, drawing on the methods of clinical epidemiology, biostatistics, health economics and health social science. This paper summarizes the evolution of the INCLEN model and the experience to date. Progress with Phase 1, the designation of sites for CEU development and the provision of advanced research training by developed country training centres has been substantial. The network now consists of 27 units: 26 in developing country medical schools in Asia, Latin America, India and Africa and 1 in France. More than 60% of the target of 270 fellows have completed training and returned to take up faculty positions in their unit. The remainder will be trained and on site by 1995. The non-return rate of fellows (2%) is very low. Research productivity is significant given only 60 fellows have been working in their CEUs for more than 3 years following the completion of training. An appropriate balance between hospital and community-based research is evident and changes in clinical and health care policy have been made based on the research conducted. The educational responsibilities of all units include courses and workshops in critical appraisal and clinical epidemiology for medical trainees and colleagues. Graduate training programs have emerged in 3 units so far. Major challenges lie ahead as we move into Phase 2 of the project--self sustainability and the transfer of training responsibility to the CEUs. The problems encountered during Phase 1 will need to be addressed. These include time protection for research, the limited availability of research funds, the low priority given to research careers and the poor linkage between health researchers and government policy makers. Our experience echos the recommendations of the recent report of the Commission on Health Research for Development, namely that donors and national governments should give increased priority to the role of health research in less developed countries. We conclude that with continuing support and special attention to the problems encountered, the INCLEN approach can contribute to ensuring that the medical establishment is part of the solution rather than the problem faced by health systems in less developed countries.