In order to determine the biological, sociocultural, and health-care characteristics associated with deaths from pneumonia among infants, this study used cases and controls to obtain quantitative information from a semistructured questionnaire given to those infants' mothers and to 28 doctors (15 who headed inpatient services and 13 who directed outpatient services). The cases were defined as children under 1 year of age who had died from pneumonia in 1995 in the Metropolitan Region of Chile (Greater Santiago). The controls were 118 children who survived after being hospitalized with pneumonia during the same period in the same area. The two groups were paired by age, month of hospitalization, and community of residence. A qualitative study was also done. It consisted of in-depth interviews with the mothers of 20 cases and of 5 controls and with 12 key informants from primary health care staffs. Of the 149 cases identified, 113 (75.8%) were studied, divided into two major groups: those who had died in the hospital (69%) and those who had died at home (31%). Of the deaths, 63.7% occurred during the first three months after birth. Both the cases and the controls had more biological and social risk than the Chilean population in general, although there were significant differences in risk between the cases and controls. The incidence of both low birthweight (< 2,500 g) and of very low birthweight (< 1,500 g) was higher in the two groups than in the general population and was significantly greater in the cases (40.7% and 13.3%, respectively) than in the controls (18.8% and 1.7%). Prematurity (gestational age < 37 weeks) was especially high among the children who had died in the hospital (40.3%), in contrast to 25.7% for those who had died at home and 17.1% among the controls. Those who had died in the hospital also had 10 times as many congenital disorders as did those who had died at home, and seven times as many as did the controls. According to their mothers, more than half (18/35) of the children who had died at home had not shown signs of pneumonia, which suggests sudden infant death syndrome or death from aspiration of food into the lungs. In terms of socioeconomics, the standard of living of the controls was lower than that of the cases. During 1995 or the second half of 1994, 60.3% of the fathers of the controls had been unemployed, as had 58.6% of the fathers of the cases. There were no differences between the cases and the controls with respect to the mothers' having noticed signs of illness or the mothers' knowledge of how to prevent pneumonia. The surveyed physicians generally believed that the primary cause of the deaths had been the high level of biological problems among those children. Since the group of children who died at home differed in both size and other characteristics from those who died in the hospital, there should be different strategies aimed at reducing deaths in the two groups. It is important to do autopsies in all deaths in order to determine if the cause of death is pneumonia, sudden infant death syndrome, or aspiration of food.