PP 94 CONCURRENT AND PREDICTIVE VALIDITY FOR FRAILTY SYNDROMES IN THE ACUTE MEDICAL CARE SETTING

Abstract

INTRODUCTION

For some, population ageing is associated with increasing frailty. Existing frailty assessment scores exhibit poor predictive power for adverse events in the acute medical setting. We have published work validating a model based on frailty syndromes (cognitive impairment, falls, reduced mobility, pressure sores, functional dependence and anxiety/depression) on English administrative data(doi:10.1136/bmjopen-2015- 008457). We aim to explore concurrent (comparison with frailty index) and predictive validity (30-Day mortality, emergency readmission and institutionalization) for this model in the acute medical setting.

MATERIALS & METHODS

A prospective observational study in the Acute Assessment Unit of Chelsea and Westminster Hospital with convenience sampling from May - Dec 2013 of adult acute medical patient admissions. Data was abstracted by a researcher from patient records up to 36 hours from admission. Outcomes were retrieved at one month after index admission. Statistical analysis includes descriptive statistics, logistic regression and Area Under the Receiver Operator Characteristics Curves(AUC) for predictive power derived from predicted probabilities. Missing data analysis followed by multiple imputation (by regression of dataset) where appropriate. Frailty Syndromes models were adjusted for age, gender and number of readmissions in previous 6 months. A Frailty Index was created from 31 criteria from previously described methodology (DOI: 10.1186/1471-2318-8-24) RESULTS Frailty syndromes were prevalent in those >65 years (N=482; cognitive impairment-27.8%, falls- 42.1%, reduced mobility-5.2%, pressure sores-10.4% functional dependence-42.5% and anxiety/depression-31.3%). The frailty syndromes model had excellent concurrent validity with Frailty Index (AUC 0.83-0.85). The frailty syndromes model had moderate to good predictive power for adverse events at 30 days (inpatient mortality AUC 0.80-0.81, emergency readmission AUC 0.71-0.72, institutionalization AUC 0.63-0.65) in comparison to Frailty Index (inpatient mortality AUC 0.73-0.74, emergency readmission AUC 0.53-0.55, institutionalization AUC 0.52-0.58)

DISCUSSION

Frailty syndromes are a valid and useful for risk stratification in older persons requiring acute medical care.