Abstract
INTRODUCTION: Frailty has emerged as an important construct to support clinical decision-making during the COVID-19 pandemic. However, doubts remain related to methodological limitations of published studies. METHODS: Retrospective cohort study of all people aged 75 + admitted to hospital in England between 1 March 2020 and 31 July 2021. COVID-19 and frailty risk were captured using International Classification of Disease-10 (ICD-10) diagnostic codes. We used the generalised gamma model to estimate accelerated failure time, reporting unadjusted and adjusted results. RESULTS: The cohort comprised 103,561 individuals, mean age 84.1, around half female, 82% were White British with a median of two comorbidities. Frailty risk was distributed approximately 20% low risk and 40% each at intermediate or high risk. In the unadjusted survival plots, 28-day mortality was almost 50% for those with an ICD-10 code of U071 (COVID-19 virus identified), and 25-35% for those with U072 (COVID-19 virus not identified). In the adjusted analysis, the accelerated failure time estimates for those with intermediate and high frailty risk were 0.63 (95% CI 0.58-0.68) and 0.67 (95% CI 0.62-0.72) fewer days alive respectively compared to those with low frailty risk with an ICD-10 diagnosis of U072 (reference category). CONCLUSION: In older people with confirmed COVID-19, both intermediate and high frailty risk were associated with reduced survival compared to those with low frailty risk.