20 Comparison of Six Frailty Screening Tools in Patients Aged 65+ with An Arm Fragility Fracture

Frailty is associated with an increased risk of falling and fracture, but not routinely assessed in fracture clinic. Early identification and management of frailty among older people with arm fragility fracture could help avoid further falls and fractures, especially of the hip. We evaluated the feasibility of assessing frailty in a busy fracture clinic.

People aged 65+ years with an arm fracture in one acute trust were recruited. Frailty was assessed in fracture clinics using six tools: Fried Frailty Phenotype (FFP), FRAIL scale, PRISMA-7, electronic Frailty Index (e-FI), Clinical Frailty Score (CFS), and Study of Osteoporotic Fracture (SOF). The sensitivity and specificity of each tool was compared against FFP as a reference. Participants identified as frail by 2+ tools were referred for Comprehensive Geriatric Assessment (CGA).

100 patients (mean age 75 years±7.2; 20 men) were recruited. Frailty prevalence was 9% (FRAIL scale), 13% (SOF), 14% (CFS > 6), 15% (FFP; e-FI > 0.25), and 25% (PRISMA-7). Men were more likely to be frail than women. Data were complete for all assessments and completion time ranged from one minute (PRISMA-7; CFS) to six minutes for the FFP which required most equipment. Comparing with FFP, the most accurate instrument for stratifying frail from non-frail was the PRISMA-7 (sensitivity = 93%, specificity = 87%) while the remaining tools had good specificity (range 93%–100%) but average sensitivity (range 40%–60%). Twenty patients were eligible for CGA. Five had recently had CGA and 11/15 referred were assessed. CGA led to 3–6 interventions per participant including medication changes, life-style advice, investigations, and onward referrals.

It was feasible to assess frailty in fracture clinic and to identify patients who benefitted from CGA. Frailty prevalence was 9%—25% depending on the tool used and was higher among men. PRISMA-7 could be a practical tool for routine use in fracture clinics.