This 2001 structured literature review synthesizes measurement property data from 43 studies on the WOMAC. It asks: across which patient groups and interventions does the WOMAC reliably and validly measure outcomes? The paper covers test-retest reliability, internal consistency, construct validity, known group validity, and responsiveness by intervention type.
When designing a hip or knee outcomes study, subscale selection matters as much as instrument selection.
Use WOMAC pain and physical function as primary endpoints — both have demonstrated reliability and responsiveness across arthroplasty and drug trial contexts. Treat stiffness as secondary at best: it has good internal consistency but consistently fails test-retest reliability, likely because just 2 items cannot generate stable scores.
When calculating sample size, always use effect size data from your specific intervention and patient group. An arthroplasty effect size of 2.4 SRM will dramatically underspower a drug trial where effect sizes run 0.17-0.94. This is a common and costly design error.
The omission of stiffness subscale data from some studies creates a downstream problem: without stiffness, you cannot compute the WOMAC global score. Decide whether stiffness is a clinically meaningful outcome for your study before excluding it.
This 2001 structured literature review synthesizes measurement property data from 43 studies on the WOMAC. It asks: across which patient groups and interventions does the WOMAC reliably and validly measure outcomes? The paper covers test-retest reliability, internal consistency, construct validity, known group validity, and responsiveness by intervention type.
When designing a hip or knee outcomes study, subscale selection matters as much as instrument selection.
Use WOMAC pain and physical function as primary endpoints — both have demonstrated reliability and responsiveness across arthroplasty and drug trial contexts. Treat stiffness as secondary at best: it has good internal consistency but consistently fails test-retest reliability, likely because just 2 items cannot generate stable scores.
When calculating sample size, always use effect size data from your specific intervention and patient group. An arthroplasty effect size of 2.4 SRM will dramatically underspower a drug trial where effect sizes run 0.17-0.94. This is a common and costly design error.
The omission of stiffness subscale data from some studies creates a downstream problem: without stiffness, you cannot compute the WOMAC global score. Decide whether stiffness is a clinically meaningful outcome for your study before excluding it.