The Clinical Frailty Scale can be used retrospectively to assess the frailty of patients with hip fracture: a validation study

PURPOSE: Frailty is a common clinical syndrome affecting hip fracture patients. Recognising and accurately assessing frailty status is important in clinical and research settings. The Rockwood Clinical Frailty Scale (CFS) is a commonly used instrument and demonstrates a strong correlation with mortality and length of hospital admission following hip fracture. What is not understood, however, is the validity of retrospectively assigned CFS scores in hip fracture patients. The aim of this study was to assess the validity of retrospective non-orthogeriatrician assigned CFS scores in hip fracture patients. METHODS: Hip fracture patients from a single major trauma centre were assessed and CFS scores were assigned prospectively by non-orthogeriatric clinicians (n = 57). A subset of these patients were also assigned a prospective CFS score by a specialist orthogeriatrician (n = 27). Two separate blinded observers (non-orthogeriatric clinicians) assigned CFS scores retrospectively using electronic patient records alone. Agreement and precision was examined using the Bland–Altman plot, accuracy was assessed using R(2) statistic and inter-rater reliability was assessed using quadratic weighted Cohen’s kappa. RESULTS: Seventy percent of the cohort were female with an average age of 83. Agreement was high between prospective non-orthogeriatrician assigned CFS scores and retrospective non-orthogeriatrician assigned CFS scores, with a low bias (0.046) and good accuracy (R(2) = 73%). Good agreement was also seen in comparisons between prospective orthogeriatrician assigned CFS scores versus retrospective non-orthogeriatrician assigned scores, with a low bias (0.23) and good accuracy (R(2) = 78%). Good inter-rater reliability was seen between blinded observers with a quadratic weighted Cohen’s kappa of 0.76. CONCLUSIONS: Retrospective CFS scores assigned by non-orthogeriatricians are a valid means of assessing frailty status in hip fracture patients. However, our results suggest a tendency for non-orthogeriatricians to marginally overestimate frailty status when assigning CFS scores retrospectively. LEVEL OF EVIDENCE: 3.