# Bridging patient-reported outcomes and performance assessments in older adults: linking the Short Physical Performance Battery to the standardised PROMIS Physical Function scale

## TL;DR

This study connects two ways of measuring physical function in older adults, allowing for better comparison and use in research and clinical settings.

## Contribution

The study links the SPPB performance tool to the PROMIS Physical Function scale for the first time.

## Key findings

- SPPB and PROMIS-PF20a scores were highly correlated (latent correlation = 0.89).
- Item response theory assumptions were satisfied, supporting the linking model.
- Linked T-scores showed stable agreement across different subsamples.

## Abstract

Assessment of physical function, a key outcome in geriatric research, relies on either patient-reported or performance-based assessments. While several patient-reported instruments have been successfully linked to the standardised Patient-Reported Outcomes Measurement Information System (PROMIS) Physical Function scale, commonly used performance-based tools, such as the Short Physical Performance Battery (SPPB), remain uncalibrated to this scale. This lack of standardisation limits interpretability, comparability, and integration of physical function data across instruments, studies, and clinical settings.

To link SPPB scores to the PROMIS Physical Function T-score metric in older adults.

This analysis is part of the Standardizing-PF project, a prospectively designed cross-sectional study examining the possibility of mapping patient-reported and performance-based assessments onto a common scale. In the present study, 556 older adults (mean age 74 years) from different clinical and community-based settings subsequently completed a generic 20-item PROMIS Physical Function short form (PROMIS-PF20a) and the SPPB. Assumptions of item response theory modelling were investigated. We estimated a unidimensional item response theory-based linking model and derived cross-walks to convert SPPB scores into standardised PROMIS PF T-scores.

SPPB and PROMIS-PF20a were highly correlated (latent correlation = 0.89); assumptions of item response theory modelling were fulfilled. After linking, agreement between observed and linked T-scores was stable across several subsamples.

The SPPB can be meaningfully linked to the PROMIS PF T-score metric, enabling standardised interpretation, comparison, and aggregation of performance-based and self-reported physical function in older adults. We provide a user-friendly score cross-walk table to facilitate application in clinical practise and standardisation in geriatric research.