Developing and validating a risk prediction model for acute care based on frailty syndromes

Area of Science:

• Gerontology
• Health Services Research
• Clinical Epidemiology

Background:

• Population ageing is associated with increased comorbidity, functional dependence, and reduced quality of life.
• Frailty pathophysiology is not fully understood, leading to a lack of consensus on its definition and operationalization.
• Existing frailty scores in acute care settings demonstrate limited predictive power for clinically significant outcomes.

Purpose of the Study:

• To evaluate the utility of frailty syndromes, as defined by national guidelines, as a risk prediction model for elderly individuals in acute care settings.
• To assess the predictive performance of frailty syndromes for key clinical outcomes including inpatient mortality, 30-day readmission, and institutionalization.

Main Methods:

• Analysis of English National Health Service (NHS) secondary care emergency admissions for patients aged over 65 years (N=2,099,252) from 2012.
• Development of multivariable logistic regression models using International Classification of Diseases, 10th revision (ICD-10) codes for frailty syndromes, demographics, and prior hospital utilization.
• Model performance was evaluated using Receiver Operating Characteristic (ROC) curves and risk decile analysis on a split dataset (60% train, 40% test).

Main Results:

• Frailty syndrome models achieved ROC scores ranging from 0.624-0.659 for inpatient mortality.
• Predictive performance for institutionalization was between 0.63-0.654, and for 30-day emergency readmission, it ranged from 0.57-0.63.
• These results indicate a moderate predictive capability for the frailty syndrome models.

Conclusions:

• Frailty syndromes represent a valid predictor of outcomes pertinent to acute care for the elderly population.
• The predictive power of these models aligns with other established scores in the literature.
• While simple and clinically relevant, the current predictive power of frailty syndromes is insufficient for direct clinical application in acute care decision-making.