Comparing Approaches to Link SF-36 PF-10 Scores to PROMIS Physical Function: A Validation Study in Three Clinical...

Area of Science:

• Health Outcomes Research
• Psychometrics
• Clinical Measurement

Background:

• Physical function (PF) is a key patient-reported outcome (PRO) across various conditions.
• Disparate PRO measures (PROMs) hinder score comparability and interpretability.
• The Patient-Reported Outcomes Measurement Information System (PROMIS®) developed a standardized T-score metric using item response theory (IRT) to address this.
• Linking algorithms enable conversion of scores from different PROMs to the PROMIS-PF metric for harmonization.

Purpose of the Study:

• To validate and compare two established linking methods for translating SF-36 PF-10 scores to the PROMIS-PF metric.
• To assess the reliability of these methods in diverse clinical populations.

Main Methods:

• Two linking approaches were evaluated: item-level linking and cross-walk tables.
• PROMIS-PF T-scores derived from SF-36 PF-10 scores were compared against directly observed PROMIS-PF20a T-scores.
• The study included patients from cardiology, rheumatology, and psychosomatic medicine.

Main Results:

• All linking methods showed high correlation with observed PROMIS-PF20a T-scores (Pearson correlation ≥ 0.84).
• Negligible practical differences were observed at the group level (standardized mean difference < 0.2).

Conclusions:

• Established linking methods reliably translate SF-36 PF-10 scores to PROMIS-PF T-scores across clinical samples.
• This eliminates the need for re-estimating models for new datasets.
• Cross-walk tables are recommended as a practical method for score conversion, avoiding complex statistical modeling.