OBJECTIVE--To determine the mortality and health effects from the current civil war in Kabul, Afghanistan. SUBJECTS--One resident population and one displaced population. DESIGN--Between November 22 and December 16, 1993, we conducted a retrospective, population-based, household survey, interviewing 312 displaced families and 300 resident families. RESULTS--During the 285 days before the survey, the highest average daily crude mortality rate and the mortality rate for those younger than 5 years (0.9 and 2.6 per 10,000 population per day, respectively) were among residents who had lived at their current location for 10 months or less. The average daily crude mortality rate and the mortality rate for those younger than 5 years were lower among displaced persons (0.6 and 1.9 per 10,000 per day) and lowest among residents who had lived at their current location for more than 10 months (0.5 and 0.6 per 10,000 per day). Overall, the most common cause of death for both groups was gunshot or other war trauma; for children younger than 5 years, deaths resulting from measles, diarrhea, and acute respiratory tract infection predominated. CONCLUSIONS--While provision of basic public health measures would likely decrease mortality among both displaced and resident populations, the most urgent health need is for a cessation of hostilities against the civilian population. During humanitarian relief operations, organizations should not focus exclusively on persons identified as displaced.