Establishing thresholds for meaningful within-individual change using longitudinal item response theory

Area of Science:

• Psychometrics
• Health Outcomes Research
• Statistical Modeling

Background:

• Meaningful within-individual change (MWIC) is crucial for interpreting patient-reported outcome measures (PROMs).
• Transition ratings (TR) are often used as anchors to define MWIC.
• Traditional methods for MWIC estimation have limitations, including ignoring floor/ceiling effects and measurement error.

Purpose of the Study:

• To introduce and evaluate a novel approach for MWIC estimation using longitudinal item response theory (LIRT).
• To compare the performance of LIRT with traditional statistical methods for MWIC estimation.

Main Methods:

• A Graded Response LIRT model was developed for baseline and follow-up PROM data, incorporating a TR item for latent change.
• The LIRT threshold parameter for the TR was used to establish the MWIC threshold on the latent metric.
• Observed PROM score MWIC thresholds were estimated, and LIRT was compared to traditional methods using example data.

Main Results:

• The LIRT model demonstrated good fit to the data.
• LIRT-estimated MWIC thresholds ranged from 3 to 4 points of score improvement.
• Traditional methods yielded a wider range of MWIC estimates (2-10 points), influenced by the proportion of self-rated improvement.

Conclusions:

• Traditional anchor-based MWIC analyses are susceptible to study-specific conditions.
• LIRT provides a more robust and promising analytical framework for establishing MWIC thresholds in PROMs.