Surveys conducted in the context of the Demographic and Health Surveys (DHS) programme are an important source of data on health of families in developing countries. Both at the national and international level, DHS surveys provide much-needed data on fertility and family planning, on mortality and nutrition, and on health services utilization. The use of uniform survey instruments allows detailed international and subnational comparisons of health status and health care. Limitations of the DHS surveys are also discussed.
PIP: Demographic and Health Surveys (DHS), funded by the USAID, were first initiated in 1984 as an expansion of World Fertility Survey (WFS) type data. DHS collected qualitative information on health and nutrition, as well as WFS data on fertility, child mortality, and family planning. DHS surveys used nationally representative samples of women of childbearing age and, most recently, samples of males. National government organizations usually administered the survey and analyzed results. A core questionnaire was adapted to each country's special needs. A list of the 44 countries conducting a DHS was provided with additional information on the most recent year/s of the survey and number of respondents. The core questionnaire included questions on fertility and mortality, anthropometry, family planning, maternity care, child feeding, vaccination, child morbidity, and AIDS. The surveys were useful in providing a wide variety of health indicators and health services indicators. Data quality checks were conducted continuously in order to improve instruments, to assure trained field personnel, to use concurrent data entry and editing, and to provide feedback to interviewers during field administration of the instrument. Results were published very quickly, and tabulations were available within 2-3 months after field work is completed. The limitations included reporting and recall bias, particularly for age or other retrospective data relying on memory of a past event. Omissions were not considered a serious problem. Individual level data required more careful interpretation. District level analysis was preferred because it corresponded with major health program levels. Samples of 1000-1500 women were required for valid estimation of fertility and child mortality. Expansion was considered unlikely because of the current length of the questionnaire. Consideration was given to supplementary modules or to inclusion of saliva or blood testing for AIDS. Countries adopting the health goals of the World Summit for Children could use DHS for base line information.