Deficits comprising multi-dimensional frailty indices based on routine data: sub-analysis of a scoping review

Poster ID
2319
Authors' names
S Dlima1; A Hall1; A Aminu1; C Todd1; E Vardy12
Author's provenances
1. School of Health Sciences, University of Manchester; 2. Oldham Care Organisation
Conditions

Abstract

Introduction

The frailty index (FI) is a frailty assessment tool calculated as the proportion of the number of deficits, or “things that individuals have wrong with them”, to the total number of variables in the index. Routine health and administrative databases are valuable sources of deficits to automatically calculate FIs. There is large heterogeneity in the deficits used in FIs. This sub-analysis of a scoping review on routine data-based FIs aimed to describe and map the deficits used in multi-dimensional FIs.

 

Methods

Seven databases were searched to find literature published between 2013 and 2023. The main inclusion criterion was multi-dimensional FIs constructed from routinely collected data. Multi-dimensional FIs should have deficits in at least two of the following categories: “symptoms/signs”, “laboratory values”, “diseases”, “disabilities”, and “others”.

 

Results

Of the 7,526 publications screened, 61 distinct FIs were identified from 60 included studies. Most FIs were developed in hospital settings (n=19). The most dominant data source of deficits to calculate the FIs was hospital records (n=23). The median number of deficits used in the FIs was 36 (range = 5–72). We identified 611 unique deficits that comprised the FIs. Most deficits were either “diseases” (34.4%; n=205) or “symptoms/signs” (32.1%; n=196), followed by “disabilities” (17.0%; n=101), “others” (10.1%; n=60), and “laboratory values” (8.3%; n=49). Forty-seven deficits were present in ≥20% of the FIs (≥12 FIs). The most common “disease” was diabetes, “symptom/sign” was depression, “disability” was hearing loss, and “laboratory value” was anaemia & haematinic deficiency.

 

Conclusion

These findings highlight the reactive approach to frailty assessment, as most of these FIs were calculated from hospital data and used symptoms/signs and diseases as deficits. Given the heterogenous manifestations and long-term impacts of frailty, using a more proactive approach that leverages non-clinical routine data is warranted to prevent frailty development and progression.

 

Presentation