Using an electronic frailty index to predict adverse outcomes in geriatric COVID-19 patients: data from the Stockholm GeroCovid study

Background

The Clinical Frailty Scale (CFS) is a strong predictor for worse outcomes in geriatric COVID-19 patients, but it is less clear whether an electronic frailty index (eFI) constructed from routinely collected electronic health records (EHRs) provides similar predictive value. This study aimed to investigate the predictive ability of an eFI in comparison to other frailty and comorbidity measures, using mortality, readmission, and the length of stay as outcomes in geriatric COVID-19 patients.

Methods

We conducted a retrospective cohort study using EHRs from nine geriatric clinics in Stockholm, Sweden, comprising 3,405 COVID-19 patients (mean age 81.9 years) between 1/3/2020 and 31/10/2021. Frailty was assessed using a 48-item eFI developed for Swedish geriatric patients, the CFS, and Hospital Frailty Risk Score (HFRS). Comorbidity was measured using the Charlson Comorbidity Index (CCI). We analyzed in-hospital mortality and 30-day readmission using logistic regression and area under receiver operating characteristic curve (AUC). 30-day and 6-month mortality were modelled by Cox regression, and the length of stay by linear regression.

Results

Controlling for age and sex, a 10% increase in the eFI was associated with higher risks of in-hospital mortality (odds ratio [OR]=2.84; 95% confidence interval=2.31-3.51), 30-day mortality (hazard ratio [HR]=2.30; 1.99-2.65), 6-month mortality (HR=2.33; 2.07-2.62), 30-day readmission (OR=1.34; 1.06-1.68), and longer length of stay (β=2.28; 1.90-2.66).

The CFS, HFRS and CCI similarly predicted these outcomes, but the eFI had the best predictive accuracy for in-hospital mortality (AUC=0.775).

Conclusions

An eFI based on routinely collected EHRs can be applied in identifying high-risk geriatric COVID-19 patients.