Association between frailty and 30-day outcomes after discharge from hospital

Background:Readmissions after hospital discharge are common and costly, but prediction models are poor at identifying patients at high risk of readmission. We evaluated the impact of frailty on readmission or death within 30 days after discharge from general internal medicine wards.Methods:We prospectively enrolled patients discharged from 7 medical wards at 2 teaching hospitals in Edmonton. Frailty was defined by means of the previously validated Clinical Frailty Scale. The primary outcome was the composite of readmission or death within 30 days after discharge.Results:Of the 495 patients included in the study, 162 (33%) met the definition of frailty: 91 (18%) had mild, 60 (12%) had moderate, and 11 (2%) had severe frailty. Frail patients were older, had more comorbidities, lower quality of life, and higher LACE scores at discharge than those who were not frail. The composite of 30-day readmission or death was higher among frail than among nonfrail patients (39 [24.1%] v. 46 [13.8%]). Although frailty added additional prognostic information to predictive models that included age, sex and LACE score, only moderate to severe frailty (31.0% event rate) was an independent risk factor for readmission or death (adjusted odds ratio 2.19, 95% confidence interval 1.12–4.24).Interpretation:Frailty was common and associated with a substantially increased risk of early readmission or death after discharge from medical wards. The Clinical Frailty Scale could be useful in identifying high-risk patients being discharged from general internal medicine wards.Readmissions within 30 days after hospital discharge are common and costly occurrences. Although many studies have attempted to identify patients at highest risk of readmission, neither experienced clinicians nor experienced researchers using rigorously developed administrative data-rich algorithms can accurately predict which patients will not successfully transition back into the community.^1–6 This suggests that currently unrecognized factors likely play a major role in readmission risk. Identification of these factors would be important for future initiatives to reduce readmission rates by targeting resources to those at highest risk.Frailty is a frequently underdiagnosed condition, with prevalence estimates ranging from 27% to 80% among inpatients^7–9 and from 4% to 59% among older adults living in the community,¹⁰ depending on the frailty measure used and the population evaluated. Frailty is a multidimensional syndrome of decreased reserve and resistance to stressors leading to increased vulnerability to adverse outcomes.^11–14 The 2 models of frailty most commonly used in the literature are the phenotype model (e.g., the approach proposed by Fried and colleagues,¹⁵ which is based on 5 objective variables assessed at one point in time that do not include psychosocial and cognitive variables) and the cumulative deficit model (e.g., the Clinical Frailty Index, which is based on a mix of more than 30 variables capturing function in many domains over time).^16–18Although the gold standard for frailty assessment is a comprehensive geriatric assessment by a multidisciplinary team, both the phenotype and cumulative deficit models appear reasonably accurate for identifying frailty. However, both are somewhat cumbersome for routine use at the bedside.¹² For these reasons, the Clinical Frailty Scale was developed and relies on clinical judgment based on history taking and clinical examination. The Clinical Frailty Scale is easy to administer at the bedside; has been used by physicians, allied health professionals and research assistants; does not require any special equipment; is highly correlated with the Fried frailty index (r = 0.8);¹⁷ and appears to be valid, reliable and reproducible.¹⁹ Some risk-prediction models, such as the LACE Index, have tried to incorporate frailty, but they did not find it to be a significant independent variable, possibly owing to the frailty measure used. A systematic review of 30 risk-prediction models for hospital readmission found that only 2 included functional status.⁴We conducted a study to evaluate whether frailty identified using the Clinical Frailty Scale is an independent predictor of death or readmission within 30 days after discharge from hospital.