Reliability and Construct Validity of Three Movement-Evoked Pain Methods in Individuals With Knee Osteoarthritis

This study aimed to compare 3 movement-evoked pain (MEP) calculation methods following commonly used performance-based tests in knee osteoarthritis and to investigate their construct validity, test-retest reliability, Standard Error of Measurement (SEM), and smallest detectable change (SDC).

Test-retest reliability study as secondary analysis from an observational longitudinal study.

Department of Physiotherapy at the Federal University of São Carlos (Brazil).

A total of 111 participants.

Not applicable.

Three performance-based tests were performed on 2 visits. The Western Ontario and McMaster Universities Osteoarthritis Index was applied. Pain was measured using the Numeric Pain Rating Scale at rest and immediately after each task. Three MEP calculation methods were evaluated: MEP-index (postmovement minus rest), MEP-max (maximum pain across tasks), and MEP-avg (average postmovement pain). Differences among methods were analyzed using the Friedman test. Construct validity was assessed using Spearman correlation coefficient based on predefined hypotheses and reliability was assessed with intraclass correlation coefficients (ICCs). SEM and SDC were also calculated.

Significant differences were found among all MEP methods (P<.001). Construct validity was confirmed only for MEP-avg (77.7%). Reliability was excellent for MEP-max (ICC=0.70, 95% CI 0.59-0.78) and MEP-avg (ICC=0.84, 95% CI 0.77-0.89), but poor for MEP-index (ICC=0.38, 95% CI 0.11-0.57). SEM and SDC were 1.43 and 3.9 for MEP-max, 1.33 and 3.7 for MEP-avg, and 1.53 and 4.2 for MEP-index.

MEP-avg showed validity, higher reliability, and lower SDC, making it the preferred method for clinical research and practice.