Introduction Increased mortality has been demonstrated in older adults with coronavirus disease 2019 (COVID-19), but the effect of frailty has been unclear. Methods This multi-centre cohort study involved patients aged 18 years and older hospitalised with COVID-19, using routinely collected data. We used Cox regression analysis to assess the impact of age, frailty and delirium on the risk of inpatient mortality, adjusting for sex, illness severity, inflammation and co-morbidities. We used ordinal logistic regression analysis to assess the impact of age, Clinical Frailty Scale (CFS) and delirium on risk of increased care requirements on discharge, adjusting for the same variables. Results Data from 5,711 patients from 55 hospitals in 12 countries were included (median age 74, interquartile range [IQR] 54–83; 55.2% male). The risk of death increased independently with increasing age (>80 versus 18–49: hazard ratio [HR] 3.57, confidence interval [CI] 2.54–5.02), frailty (CFS 8 versus 1–3: HR 3.03, CI 2.29–4.00) inflammation, renal disease, cardiovascular disease and cancer, but not delirium. Age, frailty (CFS 7 versus 1–3: odds ratio 7.00, CI 5.27–9.32), delirium, dementia and mental health diagnoses were all associated with increased risk of higher care needs on discharge. The likelihood of adverse outcomes increased across all grades of CFS from 4 to 9. Conclusion Age and frailty are independently associated with adverse outcomes in COVID-19. Risk of increased care needs was also increased in survivors of COVID-19 with frailty or older age. [ABSTRACT FROM AUTHOR]