“…Comparisons between 3 groups were performed using a Kruskal Wallis test for quantitative variables, followed by a Dunn test when a significant difference was found; and Fisher's exact method for qualitative variables. Survival and the risk of a new institutionalization in the 3 groups of patients were estimated by a multivariate logistic regression analysis and for survival by the Kaplan-Meier method too (differences were assessed by the log-rank test) adjusted for the factors known in the published literature [6,[12][13][14][15]19,26,33,34] (semi-parsimonious method): age (< 80,81-85, 86-90, > 90), sex (female as reference), comorbidities (CIRS-G, ≤ or > 8) [29], number of drugs per day (< or ≥ 5), prefracture ADL6 (≤ 5 or = 6) [30], the ability to walk without assistance before surgery, time-tosurgery (≤ or > 48 hours) and the occurrence of delirium during hospitalization. We tested the association of survival / new institutionalization and cognitive status before surgery in UPOG by calculating the odds ratio (OR) or Hazard ratio (HR) and their 95 percent confidence interval (CI) with these prognostic variables.…”