Predicting prolonged work absence due to musculoskeletal disorders: development, validation, and clinical usefulness of prognostic prediction models

• DOI: 10.1007/s00420-025-02129-8
• Publication Date: 2025-04-08T19:32:50Z
• Authors: Tarjei Rysstad; Margreth Grotle; Adrian C. Traeger; Lene Aasdahl; Ørjan Nesse Vigdal; Fiona Aanesen; Britt Elin Øiestad; Are Hugo Pripp; Gwenllian Wynne-J...; Kate M. Dunn; Egil A. Fors; Steven J. Linton; Anne Therese Tveter
• Keywords: Statistic; Predictive modelling; Disability pension; Bootstrapping (finance)

Abstract Purpose Given the lack of robust prognostic models for early identification individuals at risk work disability, this study aimed to develop and externally validate three prolonged absence among on sick leave due musculoskeletal disorders. Methods We developed multivariable logistic regression using data from 934 4–12 weeks disorders, recruited through Norwegian Labour Welfare Administration. The predicted outcomes: (1) > 90 consecutive days, (2) 180 (3) any new or increased assessment allowance disability pension within 12 months. Each model was validated in a separate cohort participants (8–12 leave) different geographical region Norway. evaluated performance discrimination ( c -statistic), calibration, assessed clinical usefulness decision curve analysis (net benefit). Bootstrapping used adjust overoptimism. Results All showed good predictive external validation sample, with -statistics exceeding 0.76. predicting days performed best, demonstrating calibration -statistic 0.79 (95% CI 0.73–0.85), providing net benefit across range thresholds 0.10 0.80. Conclusions These models, particularly one may facilitate secondary prevention strategies guide future trials. Further refinement are necessary optimise test their larger samples.