The 130th Annual Meeting of APHA

Abstract #41655

David J. Fitch, PhD, Departamento de Matematica, Universidad del Valle de Guatemala, Apartado Postal No. 82, 01901, Guatemala, Guatemala, 502 364 0336, dfitch@uvg.edu.gt

EPI is a method for selecting village houses and the children who live therein for the purpose of estimating vaccination rates. It was developed during the successful efforts of the WHO to eliminate smallpox. The interviewer goes to a village, selected with probability proportional to available size figures, throws a pencil into the air, walks off in the direction to which it points counting houses, and then using a random number procedure selects one as the starting house. The second is the one closest to the first, the third is closest to the second, etc. Selection continues until data have been collected on 7 children. It is not a probability sampling procedure. If, e.g., there are 40 houses to the north but only 20 to the south, those houses to the north have only half the probability of being selected. The method can be expected to yield biased estimates in two situations where correction is possible. Underestimation will occur in villages where average number of children per mother, and vaccination rates are negatively correlated. The consequences are that we can expect EPI to overestimate vaccination rates. This could be corrected using modern methods of estimation, as could bias where village size figures are incorrect. My hope has been that the WHO will let us work with them to introduce these corrections, and that we can then go on to develop methods for probability selection. Two possible procedures will be presented.

Learning Objectives:

• At the conclusion of this session, the participant in this session will know that the EPI method, promoted by the WHO for undertaking health surveys is not statistically sound, biased and inefficient. She will also know something about better alternatives. With this information she can undertake more effective health surveys which will help reduce the number of children needlessly dying in poor countries.

Keywords: Child Health, Statistics

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

Handout (.doc format, 45.0 kb)