Reliability of frailty assessment in the critically ill: a multicentre prospective observational study

Demand for critical care among older patients is increasing in many countries. Assessment of frailty may inform discussions and decision making, but acute illness and reliance on proxies for history-taking pose particular challenges in patients who are critically ill. Our aim was to investigate the inter-rater reliability of the Clinical Frailty Scale for assessing frailty in patients admitted to critical care. We conducted a prospective, multi-centre study comparing assessments of frailty by staff from medical, nursing and physiotherapy backgrounds. Each assessment was made independently by two assessors after review of clinical notes and interview with an individual who maintained close contact with the patient. Frailty was defined as a Clinical Frailty Scale rating > 4. We made 202 assessments in 101 patients (median (IQR [range]) age 69 (65-75 [60-80]) years, median (IQR [range]) Acute Physiology and Chronic Health Evaluation II score 19 (15-23 [7-33])). Fifty-two (51%) of the included patients were able to participate in the interview; 35 patients (35%) were considered frail. Linear weighted kappa was 0.74 (95%CI 0.67-0.80) indicating a good level of agreement between assessors. However, frailty rating differed by at least one category in 47 (47%) cases. Factors independently associated with higher frailty ratings were: female sex; higher Acute Physiology and Chronic Health Evaluation II score; higher category of pre-hospital dependence; and the assessor having a medical background. We identified a good level of agreement in frailty assessment using the Clinical Frailty Scale, supporting its use in clinical care, but identified factors independently associated with higher ratings which could indicate personal bias.