Prediction of risk of in-hospital geriatric complications in older patients with hip fracture

Abstract

Hip fracture in older persons is a frequent reason for hospital admission and a substantial workload in orthopedic wards for geriatric liaison teams. However, robust patients who do not present in-hospital complications may not need geriatric liaison. For the sake of triage, we studied the ability of usual admission scores to identify patients who will not develop in-hospital complications, and who may therefore not be included in the overworked geriatric liaison teams. A retrospective cohort of consecutive community- living elderly patients (age ≥ 75 yrs), admitted for traumatic hip fracture in the orthopedic divisions of a teaching hospital over 18 months was examined. The predictive value of commonly used frailty scores (ISAR, VIP, KATZ) to rule out the incidence of three frequent and preventable in-hospital acute geriatric events (major behavioral problems, pressure sores, falls) was assessed by ROC curves and negative likelihood ratio (-LR). Of 145 older persons with hip fracture (median age 84 years; 76% women; 57% living alone, 44% with pre-existing geriatric syndromes), 81 (56%) presented some acute geriatric events (AGE), i.e. major behavioral problems (46%), pressure sores (19%) and/or falls (5%). The three frailty admission scores showed low power for AGE prediction (area under the ROC curve: 53- 58%) and identification of patients who will not present in-hospital AGE (-LR>0.5 at the most sensitive cut-off). None of the three scores helped in the triage of patients according to their risk of future in-hospital AGE. All older patients with hip fracture, irrespective of their admission frailty-robustness profile, should receive geriatric evaluation and intervention.