Multivariate logistic regression procedures were used to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs).
Multivariate logistic regression procedures incorporated the complex survey sampling design.
These variables were then selected in a multivariate logistic regression via a stepwise procedure to build a multivariate model designed to identify factors independently and significantly associated with good neurological outcome.
For multivariate logistic analysis, a backward elimination procedure was performed to identify independent predictors for MGMT expression.
Multivariate logistic regression with forward selection procedure was used to identify those variables that independently contributed to the risk of short-term complications and long-term complications (statistically significant variables from univariate logistic regression).
Multivariate logistic analysis, using a backward procedure on the basis of likelihood ratios, was conducted to determine whether both tubular markers would be predictive factors for CKD stage 3 or greater.
All analyses were performed separately for women and men, as well as for the total group, and multivariate logistic regression analysis using a stepwise procedure, adjusting for significant interactions (considering a threshold level of 0.15), was used to assess the relationship between independent variables (mtDNA, APOE4 status, age and gender) and the outcome.
The multivariate logistic regression used a forward selection procedure and only those variables with significant p at 0.05 were included in the final model.
The odds of LVH in categories of impaired glucose regulation, compared with NGT, were analyzed by binary multivariate logistic regression analysis using a hierarchical procedure with priority enter of demographic variables, systolic BP, heart rate, and waist circumference (WC).
Model selection in the multivariate logistic regression analysis retained all 4 procedure variables (colonoscopy, sigmoidoscopy, polypectomy, DCBE in the past 4 years) as important predictors of indication, while it retained only 4 of 8 diagnostic variables from physician billing data.
