In each case, after bivariate associations were examined, variables with P < .10 or strong biologic plausibility were included in the model selection process; a P value of <.10 was chosen as a threshold to include variables in the model to ensure that marginally significant variables with strong predictive effects were not inappropriately excluded.
← all excerpts
Risk Factors for Development of Carbapenem Resistance Among Gram-Negative Rods.
1
—
—