Second, as a more general principle, the multivariable model is developed after a univariate screening for statistically significant predictors. We are concerned as to why the authors did not use univariate analysis for selecting the significant predictors for their multivariable model. On the other hand, how they did their multivariable analysis is unclear. The stepwise method is a commonly used approach for model building in biomedical research. It may be argued that the value of regression coefficients is highly reliant on univariate prescreening of variables and type of multivariable analysis. 5,6