The latent difference score model is a viable alternative to arithmetic difference score-based anchor-driven minimal important change calculation

Background Traditional MIC estimates contain measurement error from PROMs. The LDSM offers a viable alternative to estimate MIC controlling for measurement error. We also propose a conceptual framework for considering the multiple steps necessary to calculate anchor-based MICs. Methodology Data from KASTPain, a no-effect multi-center randomized clinical trial of 364 participants with knee osteoarthritis and who underwent knee arthroplasty, were used to illustrate the LDSM to estimate MIC. Based on a structural equation modeling framework, we used LDSMs to estimate MICs of three commonly used patient-reported outcomes of pain and functional status. Results We reported nine MIC estimates, three PROMs crossed with three post-intervention measurement occasions, from the LDSM. Results from the receiver operating curve analysis indicated that the area under curve was high (median: 0.88 range: 0.77–0.90) for all analyses. Conclusions Latent variable modeling of PROM change excludes PROM measurement error and may offer advantages over more traditional forms of MIC estimation.