AANA/ASES JOINT WEBINAR: Shoulder instability: When to operate?-Mather-Taylor Article

Mather-Taylor Article

Pdf Summary

This study presents the development and validation of a general-purpose, publicly available predictive model for assessing shoulder instability following a first-time anterior shoulder dislocation (FTASD). Conducted by Mather et al. at Duke University Medical Center, the model employs a Markov decision process combined with Monte Carlo simulation to estimate long-term outcomes over a 10-year horizon. It considers key patient variables such as age and gender, which are demonstrated as primary predictors of recurrent instability.<br /><br />The model simulates health states including initial dislocation, stable shoulder, recurrent instability, and revision surgical stabilization, allowing transitions based on probabilities sourced from high-level evidence and expert opinion when necessary. Utility measures used the Western Ontario Shoulder Instability index (WOSI), a disease-specific quality of life metric, while accounting for 'disutility' periods representing surgical recovery impacts.<br /><br />Validation against two external cohorts— a Swedish prospective cohort (Hovelius et al.) and a US military cohort (Bottoni et al.)— showed close alignment. For example, the model projected 49% stable shoulders at 10 years versus 52% observed in Hovelius et al., and a 76% predicted recurrence rate closely matched the military cohort after adjusting risk parameters.<br /><br />Base case simulations for 50,000 patients indicated a 63% recurrence over 10 years, with 12.6% undergoing primary stabilization surgery. The model’s flexibility allows scenario customization, demonstrated by example predictions for an 18-year-old male (77% recurrence risk year one, 32% stable at 10 years), a 30-year-old female painter, and the US military population.<br /><br />Strengths include incorporation of published data, adaptability to different populations, and the potential to support shared clinical decision-making by providing individualized prognostic estimates. Limitations are related to input data quality, unmodeled variables (e.g., activity level beyond age/gender, certain pathoanatomic factors), and assumptions regarding transition probabilities such as lack of spontaneous stability regain.<br /><br />Overall, this model offers a valuable tool for clinicians, policymakers, and researchers to inform treatment decisions for FTASD by integrating multifactorial data and patient preferences. It encourages ongoing refinement as new evidence emerges and supports improving patient-centered care through enhanced prognostic communication.

Keywords

shoulder instability

first-time anterior shoulder dislocation

predictive model

Markov decision process

Monte Carlo simulation

Western Ontario Shoulder Instability index

recurrence risk

revision surgical stabilization

patient-specific prognosis

clinical decision-making