External validation of four models predicting treatment success after interdisciplinary multimodal pain treatment

Read full text Summarise with VIRAT
Authors: Michel GCAM Mertens, Sander M. J. van Kuijk, Erik van der Graaff, Fredrick Zmudzki, Rob Smeets
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Externally validate, recalibrate, and update four prediction models for interdisciplinary multimodal pain treatment (IMPT) success in chronic musculoskeletal pain (CMP). Routine care data from individuals with CMP undergoing a 10-week IMPT was analyzed. Success was assessed using four outcomes: patients’ recovery perspective in disability, physical and mental quality of life, and disability. We evaluated 63 demographic and candidate predictors, primarily patient reported outcome measures. Data of 2204 individuals were analyzed, achieving success rates of 38%, 28%, 29%, and 52% for recovery perspective, physical and mental quality of life, and disability, respectively. After updating, calibration was similar to the original models, but discrimination was slightly reduced (−0.06 to −0.02). Furthermore, the four models included 19 predictors (one consistent across all) and demonstrated strong calibration and mostly acceptable discrimination (AUC 0.66–0.74). Decision curve analysis indicated greater net benefit for the updated models than treat-all or treat-none across clinically relevant thresholds. Standardized patient-reported outcome measures can predict IMPT success effectively. “Treatment control” (i.e., expected treatment effect) emerged as the most consistent predictor. Predictor relevance varied by outcome, underscoring the importance of careful outcome selection. These results support patient-centered care, and tailoring interventions to individual factors for optimal treatment success. Routinely assess beliefs about treatment control and pain self-efficacy at intake, as these were the most informative predictors of treatment success across models and can be directly targeted in rehabilitationAgree with each patient on the outcome domain that will define success (i.e., recovery perspective, physical or mental health-related quality of life, or disability) because predictors and benefits differ across these endpointsUse the prediction models in consultations to guide shared decision-making and selectively initiate interdisciplinary multimodal pain treatment at thresholds where benefits outweigh burden and costsIntegrate standardized patient-reported outcome measures into routine workflows to generate treatment success predictions and align goals with the outcome domain most important to each patient Routinely assess beliefs about treatment control and pain self-efficacy at intake, as these were the most informative predictors of treatment success across models and can be directly targeted in rehabilitation Agree with each patient on the outcome domain that will define success (i.e., recovery perspective, physical or mental health-related quality of life, or disability) because predictors and benefits differ across these endpoints Use the prediction models in consultations to guide shared decision-making and selectively initiate interdisciplinary multimodal pain treatment at thresholds where benefits outweigh burden and costs Integrate standardized patient-reported outcome measures into routine workflows to generate treatment success predictions and align goals with the outcome domain most important to each patient

Fields of Study & Subjects

Musculoskeletal pain and rehabilitation Pain Management and Treatment Medical Practices and Rehabilitation

Publication details

Journal / source
Figshare
Publication date
22 March 2026
Language
EN
Document type
Dataset