Of all populations affected by cholera, refugees are at particular risk of infection due to overcrowding and poor sanitation. Between 15 March and 17 May 1988, 951 cases of cholera were registered at the cholera treatment centre in a Mozambican refugee camp in Malawi. The epidemic duration was 65 days. Vibrio cholerae biotype E1 Tor serotype Inaba was isolated. To identify high-risk groups and potential risk of acquiring the disease, an epidemiologic investigation was conducted. The attack rate of recorded cases was 2.6% with a range from 0.9 to 5.1% for different sections of the camp. The case fatality rate was 3.3% and decreased from week 1 to week 6. The epidemic started in the section near the market place and radiated out. A matched-pair case-control study of food and water consumption was performed early in the outbreak. It showed that cases were more likely to use shallow wells (surface wells) instead of boreholes compared to controls (OR = 4.5, CI = 1.0-20.8, P = 0.04) and that cases were more likely to have had contact with the market than controls (OR = 3.5, CI = 0.7-16.8, P = 0.09). None of the food items available at the market was more likely to be preferred by cases than controls. Recommendations included early case finding and treatment, temporary closure of the market, tetracycline prophylaxis of contacts, and water chlorination.