Patterns of intra-familiar distribution of undernutrition: methods and applications for developing societies.

OBJECTIVE: To propose a method to assess patterns of intra-familiar distribution of undernutrition and apply it to different socioeconomic strata of the Brazilian population. DESIGN: A large nationally representative cross-sectional anthropometric survey undertaken in 1989 is the primary source of information. Undernutrition was defined as body mass index (adults) and weight-for-age (children) below the 5th percentile of a healthy and non-malnourished reference population. Log-linear models were used to assess patterns of intra-familiar distribution of nutritional status in four income strata. SUBJECTS: Two thousand, one hundred and seventy-four families composed by at least one child 6-36 month-old and his/her father and mother. SETTING: All regions in Brazil. RESULTS: Undernutrition was significantly associated among household members only for the 25% poorest families (P < 0.0001). In this group, the presence of undernutrition in the mother or the father increased 1.6-1.9 times the risk of undernutrition in the child and the presence of undernutrition in the father made it 2.7 times more frequent in the mother. The relatively small prevalence ratios suggest that even in extremely poor families only a small proportion of undernutrition could be attributed to common household determinants. CONCLUSION: Our results are consistent with the hypothesis that in transitional societies undernutrition would appear as a global family problem only for those at the earlier stages of the nutrition transition. Policies and strategies to overcome undernutrition should take this fact into account.

[Sampling design for the implementation of a national information system on the demand for ambulatory services]. / Delineamento amostral para a implantação de um sistema nacional de informações de demanda ambulatorial.

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 (Mj). 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/epsilon 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/(epsilon Mj) : n/N; are also given.