As a result of 12 years of civil war in Lebanon, it has been impossible to collect regular morbidity information at the primary level. This report is based on a case-control analysis of various health problems as identified from a population based health survey in Beirut in 1983-1984. Cases of headache, backpain and peptic ulcer, as identified from this survey of 2752 households, were matched for age, sex, and neighbourhood with controls from the same sample. Cases and controls were compared for the presence of various characteristics as collected in the household interview. Headaches were more prevalent in females and in the higher educational categories, and the odds ratio was 1.3 (95 per cent confidence interval 1.01-1.68) for the married compared to the non-married. In comparisons of backpain, the odds ratio for alcohol consumption was 2.40 (1.14-5.08), and for belonging to skilled and unskilled labour categories of occupation it was 2.33 (1.05-5.15) when the analysis was limited to the employed group only. Although the peptic ulcer cases were of lower educational background compared to their controls, no other findings were identified in this third case-control comparison. The methodological shortcomings of such studies and the various interpretations of the findings are presented in the discussion.