Abstract
Background
Frailty and delirium commonly coexist in acutely ill older adults, but they are variably measured. The Clinical Frailty Scale (CFS) and the 4AT scores are advocated as standardised tools to assess these conditions. We have developed risk categories based on these scales to predict mortality.
Methods
Two-graph ROC curve analysis derived thresholds at 5.5 for CFS and 1 for 4AT for predicting mortality, from which three composite Risk-Categories were created: “Low-Risk” represents low scores for both CFS (1–6) and 4AT (0); Intermediate-Risk” represents either high CFS (7–9) or high 4AT (1–12) scores; and High-Risk” represents both high CFS and 4AT scales. These Risk-Categories were used to predict in-hospital or 30-day mortality using logistic regression, and up to 27 months since admission using Cox regression; adjusted for age, sex, Charlson comorbidity index, anticholinergic burden and polypharmacy.
Results
There were 1192 patients (57.1 % women) of mean age 86.1 yr (SD=7.1) consecutively admitted to a hospital. Compared to those in the “Low-Risk” category (reference), in-hospital mortality odds-ratios (ORs; 95 %CI) were greater for those in the “Intermediate-Risk” category: OR=1.74 (1.11–2.72), and “High-Risk” category: OR=2.72 (1.47–5.02). Corresponding values for within 30-day mortality were: OR=1.75 (1.18–2.60) and OR=3.03 (1.76–5.21). Risk of death within 27 months of admission was increased in the “High-Risk” category: hazard ratio=1.46 (1.14–1.87). The association of CFS and mortality was partially mediated by 4AT.
Conclusion
Mortality in hospital was approximately doubled in “Intermediate-Risk” and tripled in “High-Risk” patients, which persisted after discharge. These risk categories are a useful tool for identifying high-risk patients.
Frailty and delirium commonly coexist in acutely ill older adults, but they are variably measured. The Clinical Frailty Scale (CFS) and the 4AT scores are advocated as standardised tools to assess these conditions. We have developed risk categories based on these scales to predict mortality.
Methods
Two-graph ROC curve analysis derived thresholds at 5.5 for CFS and 1 for 4AT for predicting mortality, from which three composite Risk-Categories were created: “Low-Risk” represents low scores for both CFS (1–6) and 4AT (0); Intermediate-Risk” represents either high CFS (7–9) or high 4AT (1–12) scores; and High-Risk” represents both high CFS and 4AT scales. These Risk-Categories were used to predict in-hospital or 30-day mortality using logistic regression, and up to 27 months since admission using Cox regression; adjusted for age, sex, Charlson comorbidity index, anticholinergic burden and polypharmacy.
Results
There were 1192 patients (57.1 % women) of mean age 86.1 yr (SD=7.1) consecutively admitted to a hospital. Compared to those in the “Low-Risk” category (reference), in-hospital mortality odds-ratios (ORs; 95 %CI) were greater for those in the “Intermediate-Risk” category: OR=1.74 (1.11–2.72), and “High-Risk” category: OR=2.72 (1.47–5.02). Corresponding values for within 30-day mortality were: OR=1.75 (1.18–2.60) and OR=3.03 (1.76–5.21). Risk of death within 27 months of admission was increased in the “High-Risk” category: hazard ratio=1.46 (1.14–1.87). The association of CFS and mortality was partially mediated by 4AT.
Conclusion
Mortality in hospital was approximately doubled in “Intermediate-Risk” and tripled in “High-Risk” patients, which persisted after discharge. These risk categories are a useful tool for identifying high-risk patients.
| Original language | English |
|---|---|
| Article number | 100180 |
| Number of pages | 9 |
| Journal | Archives of Gerontology and Geriatrics Plus |
| Volume | 2 |
| Issue number | 3 |
| Early online date | 15 Jun 2025 |
| DOIs | |
| Publication status | Published - 1 Sept 2025 |