The electronic frailty index shows a graded association with mortality, but not that identifying frailty changes what happens next

A systematic review and meta-analysis pooled studies that defined frailty with the electronic frailty index (eFI), which is derived from electronic health record data collected in the course of care, and followed older adults for mortality Meta-analysis (Du et al., PMID 42890831). Eight databases were searched from inception to 1 April 2025 and risk of bias was assessed with the Newcastle-Ottawa Scale. Twelve studies contributing 13 cohorts were included; 11 cohorts gave hazard-ratio estimates and two gave odds-ratio estimates, which were analysed separately.

Against older adults without frailty, the pooled hazard ratios for mortality were 1.81 for mild, 2.95 for moderate and 4.41 for severe frailty, a graded association. Heterogeneity was substantial in the main analyses, and the size of the association varied with follow-up duration, eFI cut-off definition, study setting and disease category. The authors state that the findings should not be interpreted as direct evidence of predictive performance. This account rests on the published abstract. The abstract does not say how many of the included studies were from the UK, and it reports no outcome other than mortality.

The benchmark is therefore narrow. The review supports using an electronic-records frailty index to stratify risk, which is a statement about who is more likely to die. It does not show that identifying frailty changes what happens to the person afterwards. That is the test set by the published pattern on structured frailty identification: a controlled evaluation showing that identification changes readmission, time at home or place of death. This review does not meet it, so the identification side of that pattern holds and its open condition stays open. It also confirms the reading in the synthesis for the week ending 2 October that scoring predicts death while nothing yet shows it changes the next step.

What follows identification in English general practice is the subject of a national audit published as a signal the same day.