To perform this multivariate logistic regression, a multiple variable logistic model was used.
However, gender was reinserted in the multiple variable logistic model for its importance as a socio-demographic variable, in addition to it maintaining a statistical p-value of less than 0.2 [ 22].
The presence of an interaction between vasopressor dose variables and baseline SOFA was evaluated using an interaction term in three-variable logistic models (vasopressor variable, SOFA variable and vasopressor*SOFA variable).
The explanatory variables with a cut-off of p-value < 0.1 from univariate analyses were used and included into the multi-variable logistic models.
Single-variable logistic models were used to select the 2- and 3-mile density measures for mammography facilities and public transportation respectively.
In total, 22 variables were assessed in the single variable logistic regression models.
Potential covariates were initially examined in single variable logistic regression models.
For the primary outcome variable, logistic regression models will be created, using ITT analysis.
With questionnaire response as the dependent variable, logistic regression models were performed in order to determine predictors for questionnaire response.
Multiple variable logistic regression models showed shorter height, lower BMI, higher systolic BP and lower IOP to be independent predictors of a small eye (table 3).
To adjust for possible confounders and to detect potential interactions with the "treatment group" variable, logistic regression models will be constructed, adjusted for the baseline level.
