A Predictive Theory for the Calibration of Physical Functioning Patient Survey Items

Objectives: Theory development enables the modeling of relationships in physical functioning (PF), demonstrating the practical value of measurement theory relative to predictive control of the modeled parameters. Methods. Predictive theories provide highly practical foundations for efficient instrument calibration. PF theory hypothesizes the intentional use of the torso and extremities in tasks requiring varying degrees of coordination, balance, endurance, intensity, and sensitivity. In a pilot test of concept, a 14-item rating system based in this theory was used to score 37 PF activities listed as items in two patient-reported outcome measures, one of which was the SF-36’s PF-10. The instruments were calibrated simultaneously via Rasch analysis, using the theory-based ratings. The theoretical calibrations were then evaluated in a regression analysis for their capacity to predict the empirical calibrations, and this result was compared with a regression model employing nine of the 14 theory-based ratings. Results. Regression analysis shows the nine theory-based ratings to account for 94% of the variance in the empirical values, compared with 83% for the composite Rasch scale created from the 14-item rating system. Theoretical and empirical calibrations were equated to the same unit. Conclusions. Empirical data reproduce theoretical relationships between PF patient measures, response probabilities, and PF item calibrations. Previous research shows PF item calibrations to be invariant across samples, and that measures made from different instruments items are highly correlated. Theory-based reference standard metrics for PF measurement appear viable. Implications. An improved construct specification equation will combine the existing rating system with the log-frequency at which tasks appear in a typical day. Theory-based PF item calibration may soon scale relevant environments' individual challenges and efficiently allocate resources in targeted therapies