Efficiency of EPI cluster sampling for assessing diarrhoea and dysentery prevalence

Abstract

This study examines the efficiency of EPI cluster sampling in assessing the prevalence of diarrhoea and dysentery. A computer was used to simulate fieldwork carried out by a survey taker. The bias and variance of prevalence estimates obtained using EPI cluster sampling were compared with those obtained using simple random sampling and cluster (stratified random) sampling. Efficiency ratios, calculated as the mean square error divided by total distance travelled, were used to compare EPI cluster sampling to simple random sampling and standard cluster sampling. EPI cluster sampling may be an appropriate low-cost tool for monitoring trends in the prevalence of diarrhoea and dysentery over time. However, it should be used with caution when estimating the prevalence of diarrhoea at a single point in time because of the bias associated with this cluster sampling method.

PIP: Diarrhea is one of the major causes of morbidity and mortality among children in developing countries, and more than 12,000 people are estimated to die per day from diarrheal disease in such countries. Diarrhea has also been shown to affect child growth and is an important cause of malnutrition which can adversely affect child mortality. The infectious nature of diarrhea suggests that fecal contamination of water and food contributes to its transmission. That is, a contaminated source of water, if serving several families, can cause a cluster of illness. Simple random sampling and cluster sampling are discussed as examples of strategies which can be used to measure the prevalence of diarrhea and dysentery in a community. The authors investigated the cost of conducting World Health Organization (WHO) Expanded Program on Immunization (EPI) cluster sample surveys, expressed as the distance travelled to visit the sample subjects, and the bias and variance of the estimates derived with that design. The efficiency of EPI cluster sampling was specifically evaluated with regard to the assessment of the prevalence of diarrhea and dysentery. A computer was used to simulate fieldwork conducted by a survey taker. The bias and variance of prevalence estimates obtained using EPI cluster sampling were compared with those obtained using simple random sampling and cluster sampling. EPI cluster sampling was determined to be an appropriate low-cost tool with which to monitor trends in the prevalence of diarrhea and dysentery over time. It should, however, be used with caution when estimating the prevalence of diarrhea at a single point in time because of bias.