Surveys to measure programme coverage and impact: a review of the methodology used by the expanded programme on immunization.

To improve the health status of their populations most countries are developing their ability to provide primary health care. This ultimately depends on information for planning supervision and monitoring of health activities. Data are needed to define the need for health services the efficiency of existing services as well as their impact on morbidity and mortality. Such imformation can best be obtained from surveys. The WHO and other international agencies have been active in promoting the use of surveys and as an example through its Expanded Program on Immunization (EPI) WHO aims to ensure the availability of immunization for all children in the world by the year 1990. This is a vital effort towards the goal of health for all by the year 2000. EPI developed an appropriate system for data collection which could be implemented in a relatively standardized manner from 1 country to another. The primary purpose of the methodology was to assess the level of immunization coverage but becasue of its success it was also adopted for other purposes and suggests limitations modifications and alternatives to meet the needs of different health programs. The sampling strategy utilized was the probability proportionate to size (PPS) cluster sampling. The EPI survey as it is currently carried out for determining immunization coverage involves the detailed review of immunization status of about 210 children by trained reviewers. Field methodology involve identifying precisely what the population is and which age groups within the population are of particular interest and determining which individuals within the cluster to study. Random selection methods are explained. A distinct problem with the EPI methodology is the risk that surveys of adjacent households could either over or underestimate the true population coverage depending on where the starting households happen to be. Leaving selection of successive households to the interviewer creates another opportunity for bias. With some modifications the EPI methodology has been applied to studies of the incidence of poliomyelitis neonatal tetanus diarrhea and studies of morality due to measles. Careful consideration of sample size is necessary. An evaluation of the EPI sampling strategy via computer simulation is presented. An alternative method is the Lot Quality Assurance Sampling (LQAS) technique.