This finding led the
authors to build 2 separate regression models: one with
diagnosis and age and the other with diagnosis and L gap
for both SRS-22 and ODI as an effective solution for the
potential multicollinearity issue (owing to the existence of
correlated predictors in this model).