Search

The wide spread distribution of statistical software recommended for multivariate analysis as well as the ease in handling it can lead the users into adopting wrong measures if they do not pay attention to the theoretical principles behind those methods. With a view to bringing out some of these principles some steps for the data analysis of a case-control study undertaken in the city of S. Paulo-Brazil from March, 1993 to February, 1994 in order to test the association between diabetes mellitus and ischaemic heart disease after adjusting for potential confounders and/or modifiers of effect are presented. Methodologic issues are emphasized in the development of four steps: a) the data bank structure: b) the calculation of statistical power; c) the definition of variables strata and codification and d) the choice of the logistic regression method.

The reliability of information about mothers' and fathers' education, weight and height at birth, history of diarrhoea, duration of exclusive breast feeding and age of introduction of cows' milk products, selected from a structured questionnaire used in home interviewers was examined in a sample of 38 cases and 38 controls from a study related to the risk factors of insulin-dependent diabetes mellitus. The repetition of the questions was done by telephone. The agreement between the answers of both interviewers was verified using the kappa statistic (categorical variables) and the intra-class correlation coefficient (quantitative variables). The results enable one to conclude that the information is reproducible.

Many studies in this field justify the hypothesis that early weaning is one of the causes of anemia in the first year of life. This study seeks to discover the prevalence of iron deficiency anemia and its relation to the duration of breast-feeding among infants aged 0-12 months.A sample of 317 infants of four school health centers in the city of S. Paulo (Brazil) was studied. The information about diet was obtained by interviewing the mothers. The presence of anemia was verified by hemoglobin concentration, using the cianometahemoglobin method, and the criteria recommended by the World Health Organization (WHO) were utilized for the diagnosis. The duration of exclusive breast feeding was calculated by using the life table technique for censored data.The prevalence of anemia was found to be 14.5% for the whole population and 22.6% for infants aged more than 180 days. No association was found between anemia and time of breast-feeding; the median time were the same for both groups, anemic and non-anemic.The results observed are more probably related to the demographic characteristics of the population studied and by no means lessen the importance of exclusive breast feeding in the prevention of anemia.

Several works have been reported diabetes as an independent risk factor for ischaemic heart disease. The use of different methodologies have been an obstacle to a comparison of these studies. The study realized has sought to test the association between diabetes and ischaemic heart disease, after adjusting for known confounders and/or modifiers of effect. The study was designed as a case control study and the period of data collection was one year (March/93 to February/94). The cases were compared with three kinds of controls in two base-populations: primary and secondary. The sample was composed of 833 individuals of both genders of 30-69 years of age living in the city of S. Paulo, Brazil. Logistic regression was the statistical method used for the analysis of the data. The results showed that diabetes was not an independent risk factor for IHD. Hypertension, hypercholesterolemia, smoking and family antecedents of cardiovascular diseases were considered major risk factors for ischaemic heart disease. The interaction between diabetes and exposure levels of the others variables did not present statistical significance. Some methodological issues are presented to explain different magnitudes of effect according to the different kinds of controls. It is concluded that the presence of major risk factors in the models contributed to the disappearance of the association between diabetes and ischaemic heart disease.

Data of a case-control study of esophageal cancer were used as an example of the use of multivariate analysis with stratification and logistic regression. Eighty-five cases and 292 controls were classified according to sex, age and smoking and drinking habits. The point estimates of the odds ratios were similar, and the techniques were considered complementary.

The association between esophageal cancer and smoking and drinking habits, instruction and nutritional factors was examined by means of the utilization of data of a case-control study which was undertaken in the city of S. Paulo (Brazil). Eighty five cases were compared with two hundred and ninety two hospital controls, with different diseases, including other kinds of cancer. The crude estimations of the odds ratios were calculated for all the variables. Logistic regression was used in the next steps of the analysis. The drinking habit [odds ratio = 3.68, 95% confidence interval (1.74 - 7.78)], the smoking habit [odds ratio 4.86; 95% confidence interval (1.95 - 12.13)] and the frequent eating of hot pepper [odds ratio = 2.48; 95% confidence interval (1.46 - 4.23)] are important risk factors for the disease. The estimate of odds ratio for smoking ordinary cigarette was 3.43 (1.31-8.97) and for smoking of corn straw hand-rolled cigarrette was of 4.18 (1.38-12.66).

This study relates to a case-control analysis for the purpose of verifying the association between oral clefts and possible risk factors. The analysed variables were: place of mother's residence (urban/rural), pollution, parental diseases, mother's diseases during the first four months of pregnancy, intake of drugs related to this period, heredity, smoking habits, alcohol consumption and X-ray examinations during pregnancy or X-ray examinations prior to pregnancy. There were 450 cases of clefts of whom 354 had a cleft lip with or without cleft palate and 96 had a cleft palate. The relative risk (RR) for each variable by was estimated points and at a 95% of confidence interval and multivariate analysis was applied. As regards cleft lip with or without cleft palate, the risk factors are heredity (RR = 4.96), epilepsy in the mother (RR = 2.39) and the intake of drugs such as anti-inflammatory substance in the first four months of pregnancy (RR = 2.59). Related to cleft palate, the risk factors are heredity (RR = 2.82) and pollution (RR = 2.58).

The results of a cross-sectional (prevalence) study of high blood pressure, carried out in 1987 in Araraquara County in the north of S. Paulo State, Brazil are presented. The group studies numbered 1,199 people, including 533 men and 666 women, of from 15 to 74 years of age. The results show a high prevalence of the disease (32.0% for men and 25.3% for women), with a tendency to increase with age, up to the 40's (men) and the 50's (women). There was also a greater percentage of hypertensives in the mulatto and black groups, in the obese, and among those of lower family income, with a lower level of schooling and with less well remunerated jobs. However, the logistic regression showed very small odds ratios associated with these risk factors, which could be attributed to the way these factors were considered.

A cross-sectional study for prevalence of arterial hypertension in the population aged 15-74 years of age of the urban area of Araraquara County, 250 km from the city of S. Paulo, S. Paulo, State, Brazil, in 1987, was performed. The questionnaires presented to 1,199 people (533 men and 666 women) at the interview consisted of regarding sociodemographic variables, as well as the use of tobacco (smoking), the ways in which tobacco was used and the habit of inhaling the smoke. The sample was taken by the procedure of clustering, carried out in three stages. The sample was equiprobabilistic. The prevalence of tobacco smoking was very high being of 45.2% for men and 22.8% for women. The ex-smokers accounted for 15.9% of men and 8.0% of women. The men smoked much more than the women. The poorer smoked more than the richer, in both sexes. Among men, the prevalence of smoking was inversely proportional to duration of schooling, but that difference was not noted in women. As there have been other studies regarding the high prevalence of obesity and high blood pressure, it was concluded therefore that the population of Araraquara, an average town of the affluent urban interior of S. Paulo State, has a high frequency of risk factors for chronic non-transmissible diseases.