Lemeshow S, Tserkovnyi A G, Tulloch J L, Dowd J E, Lwanga S K, Keja J
Int J Epidemiol. 1985 Sep;14(3):473-81. doi: 10.1093/ije/14.3.473.
A Monte Carlo simulation study was designed to evaluate the sample survey technique currently used by the Expanded Programme on Immunization (EPI) of the World Health Organization. Of particular interest was how the EPI strategy compared to a more traditional sampling strategy with respect to bias and variability of estimates. It was also of interest to investigate whether the estimates of population vaccination coverage were accurate to within 10 percentage points of the actual levels. It was found that within particular clusters, the EPI method was particularly sensitive to pocketing of vaccinated individuals, but the more traditional method gave more accurate and less variable results under a variety of conditions. However, the stated goal of the EPI, of being able to produce population estimates accurate to within 10 percentage points of the true levels in the population, was satisfied in the artificially created populations studied.
