Mediating effect of depressive mood and walking speed on the association between injurious falls and changes in cognitive performance in the subsequent time intervals. Notes: Numbers are β coefficients (p values). The linear mixed-effects regression model used to test the association between injurious falls (at least 1 vs none) considered as a time-varying variable, and changes in MADRS or walking speed is adjusted for age, sex, educational level, living arrangement, history of falls, pain, number of chronic diseases, number of fall-risk-increasing drugs, and number of unplanned hospitalizations in the previous 3-year period (as time-varying variable). The linear mixed-effects regression model used to test the association between injurious falls (at least 1 vs none) and cognitive performance in the following time interval also includes the mediators (MADRS and walking speed, as continuous variables).