An Study On Self Reported Recurrent Falls

922 WordsMay 1, 20174 Pages
Outcome Variable The primary outcome for this study was self-reported recurrent falls in the previous twelve months which is categorized base on the participants’ response to the above two questions into: no fall, single fall, and recurrent falls. Independent Variables The following variables based on the literature, were selected for this study: 1. Demographic variables: Age group (65-74; 75-84; 85 and older), Sex (Male; female), BMI classification (Underweight; Normal weight; Overweight/Obese), Marital status (Married/Common law; Widowed; Divorced/Separated; Single), Education (Less than secondary school; Secondary school; Post-secondary). 2. General health & functional condition: Able to walk without help (Yes; No), Perceived health…show more content…
Variables showing statistically significant associations with a p-value < 0.25 from the univariate analysis were first examined for collinearity and then were entered into the multivariate analysis. Using “Proc Surveylogistic”, multinomial logistic regression models for the prediction of the single fall and recurrent falls were conducted, considering non-fallers as the reference group. At every step of the modeling process, a variable selection decision was made using alpha value of 0.05 as the significance level [28]. When a variable was removed, testing for confounding by comparing the values of the estimated regression coefficients between the two models at each stage were carried out. Any change in the regression coefficient values of 20% or more was considered an indication of confounding and the variable was retained in the model. Based on the literature, all major interactions among the variables in the model were tested, using 0.05 as the level of significance. Finally, the predictive power of the model was evaluated using a Cox & Snell pseudo R-square [29]. Statistical tests were based on a two-tailed level of significance, with an alpha value of 0.05. RESULTS In tabulating, analyzing and reporting the results we adhered to the guidelines by Statistics Canada. Therefore, all total
Open Document