Prediction of Vitamin D Deficiency Among Tabriz Elderly and Nursing Home Residents Using Stereotype Regression Model

Objectives: Vitamin D deficiency is one of the most important health problems of any society. It is more common in elderly even in those dwelling in rest homes. By now, several studies have been conducted on vitamin D deficiency using current statistical models. In this study, corresponding proportional odds and stereotype regression methods were used to identify threatening factors related to vitamin D deficiency in elderly living in rest homes and comparing them with those who live out of the mentioned places. Methods & Materials: In this case-control study, there were 140 older persons living in rest homes and 140 ones not dwelling in these centers. In the present study, 25(OH)D serum level variable and age, sex, body mass index, duration of exposure to sunlight variables were regarded as response and predictive variables to vitamin D deficiency, respectively. The analyses were carried out using corresponding proportional odds and stereotype regression methods and estimating parameters of these two models. Deviation statistics (AIC) was used to evaluate and compare the mentioned methods. Stata.9.1 software was elected to conduct the analyses. Results: Average serum level of 25(OH)D was 16.10±16.65 ng/ml and 39.62±24.78 ng/ml in individuals living in rest homes and those not living there, respectively (P=0.001). Prevalence of vitamin D deficiency (less than 20 ng/ml) was observed in 75% of members of the group consisting of those living in rest homes and 23.78% of members of another group. Using corresponding proportional odds and stereotype regression methods, age, sex, body mass index, duration of exposure to sunlight variables and whether they are member of rest home were fitted. In both models, variables of group and duration of exposure to sunlight were regarded as meaningful (P