A unique case study approach to training medical students in community diagnosis techniques was initiated at the Medical College of Ohio at Toledo. This paper describes the five elements of this teaching method: preliminary specification of target community and data base; group problem-solving requirement; specification of desired output; defined performance objectives; and regularly scheduled time for analysis. Experience with the case study method over two years was evaluated to identify specific strengths and weaknesses. The identified strengths include use of limited educational time to introduce community health problems, development of experience in a collegial team work setting, and specific awareness of the types of data useful to the analysis of community health service problems. Negative evaluations suggested that the method was not conducive to the development of skills in three areas: ability to establish the relative importance of health problems in communities; ability to identify an appropriate health system response to a community health problem from feasible alternatives; and ability to anticipate the community impact of health program modifications or improvements. Potential explanations for these deficiencies include: need for increased didactic support in the classroom for particular skill areas; need to establish a direct field experience in community diagnosis; inappropriateness of the data base used for evaluation of particular skills; and the probability that quantitative analysis, as used in this evaluation, may not be sufficient in and of itself to measure the outcome of a community diagnosis experience.