Community-based comprehensive primary healthcare programmes are a widely-promoted strategy for improving child survival in less-developed countries, but limited documentation exists concerning their effectiveness in actually reducing child mortality. This study examined the impact of a community-based comprehensive primary healthcare programme on child survival in Bolivia. Mortality rates from two intervention areas where Andean Rural Health Care (ARHC) had been conducting child-survival activities for 5-9 years were compared with those from two geographically-adjacent comparison areas that lacked such activities and that were virtually identical to the intervention areas in socioeconomic characteristics. Vital events were registered at the time of regular visit to all homes. In the comparison areas, limited services were available which reached only a small percentage of the population, while in the intervention areas, prenatal care, immunizations, growth monitoring, nutrition rehabilitation, and acute curative services were readily available to the entire population. In 1992-1993, the annual rates of mortality of children, aged less than five years, were 205.5 per 1,000 and 98.5 per 1,000 in the comparison and intervention areas respectively. The absolute difference in mortality of 107.0 deaths per 1,000 (95% confidence interval [CI], 72.7-141.3 per 1,000) represented 52.1% (95% CI, 35.2-68.8%) lower mortality of children aged less than five years in the intervention areas compared to the control communities. These results suggest that the provision of community-based, integrated health services can significantly improve child survival in poor countries. Better-designed and larger field trials of community-based comprehensive primary healthcare programmes in multiple regions of the world are needed to provide a stronger scientific basis for developing this approach further in developing countries.