Abstract
BACKGROUND: Estimates of the substantial clinical benefit (SCB) for knee arthroplasty have been endorsed by the Centers for Medicare and Medicaid Services but have not been validated. We examined validity by comparing SCB estimates to an externally validated gold standard of meaningful improvement. METHODS: There were two datasets, the Knee Arthroplasty pain coping Skills Training (KASTPain) no-effect randomized clinical trial and the Osteoarthritis Initiative (OAI) data on patients who had a knee arthroplasty. Both studies captured preoperative and multiple postoperative Western Ontario and McMaster Universities Osteoarthritis Index Physical Function scores. A total of 396 participants were included from the OAI, and 384 participants were from KASTPain. RESULTS: The validity of Latent Class Analysis estimates was supported. Compared to the gold standard, KASTPain participants meeting/exceeding the SCB generally ruled in a good outcome, while change scores less than the SCB did not rule in a poor outcome. The diagnostic validity of SCB estimates in OAI was weak. CONCLUSIONS: We found evidence of weak validity for SCB estimates, particularly for the subset of participants not meeting the SCB. Future work should explore the validity of baseline quartile-adjusted SCB estimates obtained at multiple sites. LEVEL OF EVIDENCE: Level 1 Diagnostic test study.