Sampling design for a national health care information system

Abstract

The principles and planning of a sample of consultations undertaken in the public health services belonging to or financed by INAMPS (Brazilian federal social welfare service), are gave. It was describe the epidemiological profile of the population assisted over a twelve-month period. The sampling technique of clustering in two stages, with probabilities proportional to size (PPS), was used. The health services are the Primary Sampling Unit (PSU) and the estimated number of consultations for each one, over the period, is the measure of its size (M.). The Brazilian states were taken as strata named domains. The sample size of consultations of Health Services was defined by means of simulations studies. The overall sampling fraction was defined by f= n/N = a.Mj/;å;Mj . b/Mj; where b = 250. The calculations, which include intervals for the systematic sampling of consultations (Ic) in the second stage of selection, defined by Ic = Mj/(;å;Mj) : n/N; are also given.