Developing and validating a frailty score based on patient-reported outcome 3 months after stroke: A Riksstroke-based...

Area of Science:

• Gerontology
• Neurology
• Health Services Research

Background:

• Frailty is prevalent post-stroke, significantly impacting patient outcomes.
• Existing frailty measures are often clinician-rated and time-intensive.
• Patient-reported outcome measures (PROMs) offer a scalable alternative.

Purpose of the Study:

• To develop and validate a patient-reported frailty score using the Swedish Stroke Register (Riksstroke).
• To assess the score's psychometric properties and prognostic validity.

Main Methods:

• Analysis of 19,470 stroke survivors' responses to nine patient-reported items.
• Mokken Scale Analysis and exploratory factor analysis for dimensionality.
• Item response theory (IRT) models, including bifactor analysis, for score computation.
• Differential item functioning (DIF) analysis for measurement fairness.
• Prognostic validity assessed via survival analysis (Kaplan-Meier, Cox regression).

Main Results:

• All items met scalability criteria; factor analysis identified Physical Functioning and Well-being/Mental Health facets.
• A bifactor IRT model demonstrated the best fit, capturing a general frailty dimension.
• Minimal DIF observed for sex and education; modest age-related effects.
• Higher frailty scores significantly correlated with increased all-cause mortality.

Conclusions:

• A robust, fair, and prognostically valid frailty score can be derived from Riksstroke PROMs.
• This approach enables efficient, cost-effective frailty monitoring in stroke survivors.
• The findings support leveraging routine PROMs for quality improvement in stroke care.