Can Patient Frailty Be Estimated from Inpatient Records? A Prospective Cohort Study

Background: Frailty is a common condition in older people affecting around 1.8 million people in the UK and is independently associated with adverse outcomes. Frailty is commonly measured with the Clinical Frailty Scale (CFS) which involves a face to face assessment. There is evidence the CFS can be assigned retrospectively from information in patient records, but no research has assessed whether scores can be assigned using routine inpatient records. We aimed to assess whether a CFS could be accurately assigned using hospital inpatient records in a UK setting. Methods: Forty newly admitted patients aged 65 and over were prospectively recruited at Bradford Royal Infirmary. A CFS score was assigned to each patient using a face to face assessment. A CFS score was independently assigned to each patient solely using information in the inpatient records. A quadratically weighted Cohen’s Kappa was used to estimate inter-rater reliability (IRR) between the assessments. Results: A Kappa of 0.84 was estimated for the IRR between the scores derived from a face to face assessment and those based solely on inpatient records, indicating good agreement. The Kappa score fell to 0.7 when a 4 category collapsed version of the CFS was used. The IRR appeared unaffected by the number of times a patient had been admitted recently. Conclusion: The CSF may accurately be determined from data routinely recorded in inpatient records. Retrospectively derived scores can, therefore, be used in analysis of patient outcomes of older people in research and service improvement.