Comparison of regression models between quantitative and categorical variables in the studies on the quality of life of the elderly

The aim this study was to evaluate the effect of socioeconomic and health aspects variables and WHOQOL-BREF and WHOQOL-OLD domains on the overall quality of life of the elderly by two statistical models: multiple linear regression quantitative and multiple categorical Poisson regression. The regression analyzes performed in this study were based on data from a cross-sectional study with the population of elderly users from four Living Center for Senior Citizens in the Municipality of Cuiaba-MT, considering a random sample of 317 elderly. Generally, the quantitative linear regression model is applied in quality of life studies, however, the scores of this variable do not often follow a linear behavior, even when performing a transformation of the data of the response variable. In this situation and also considering that this study is transversal, an alternative was the application of the multiple Poisson regression model with the categorized dependent variable scores. In this study, only the multiple Poisson model was adequate to the data and it was found that the variables that together presented the greatest risks in quality of life were: perception of ill health , no walking and scores of the physical and psychological domains; all below than the median.