In 1980, diarrhoea was the leading cause of child mortality, accounting for 4.6 million deaths annually. Efforts to control diarrhoea over the past decade have been based on multiple, potentially powerful interventions implemented more or less simultaneously. Oral rehydration therapy (ORT) was introduced in 1979 and rapidly became the cornerstone of programmes for the control of diarrhoeal diseases. We report on the strategy for controlling diarrhoea through case management, with special reference to ORT, and on the relationship between its implementation and reduced mortality. Population-based data on the coverage and quality of facility-based use of ORT are scarce, despite its potential importance in reducing mortality, especially for severe cases. ORT use rates during the 1980s are available for only a few countries. An improvement in the availability of data occurred in the mid-1990s. The study of time trends is hampered by the use of several different definitions of ORT. Nevertheless, the data show positive trends in diarrhoea management in most parts of the world. ORT is now given to the majority of children with diarrhoea. The annual number of deaths attributable to diarrhoea among children aged under 5 years fell from the estimated 4.6 million in 1980 to about 1.5 million today. Case studies in Brazil, Egypt, Mexico, and the Philippines confirm increases in the use of ORT which are concomitant with marked falls in mortality. In some countries, possible alternative explanations for the observed decline in mortality have been fairly confidently ruled out. Experience with ORT can provide useful guidance for child survival programmes. With adequate political will and financial support, cost-effective interventions other than that of immunization can be successfully delivered by national programmes. Furthermore, there are important lessons for evaluators. The population-based data needed to establish trends in health service delivery, outcomes and impact are not available in respect of diarrhoea, as is true for malaria, pneumonia and other major childhood conditions. Standard indicators and measurement methods should be established. Efforts to change existing global indicators should be firmly resisted. Support should be given for the continuing evaluation and documentation activities needed to guide future public health policies and programmes.