Background: Over the past two decades, England has experienced a notable rise in prescription opioid use, leading to some of Europe’s highest opioid-related hospital admissions [1], which nearly doubled between 2008 and 2018 [2]. Patients with Rheumatic and Musculoskeletal Diseases (RMDs) are particularly vulnerable due to multimorbidity and long-term opioid use risk [3]. Machine learning-based clinical prediction models (ML-CPMs) could improve current patient risk assessment by addressing nonlinear relationships.

Objectives: This study aimed to develop and evaluate CPMs based on regression and ML, using nationally representative UK data, to estimate the risk of opioid-related hospitalisations in RMD patients with non-cancer pain.

Methods: A retrospective cohort study was conducted using the Clinical Research Practice Datalink, a large-scale UK primary care electronic health record database, from 2006-2021. Patients were included if they were ≥18 years old, new opioid users, and diagnosed with one or more RMDs, including osteoarthritis, fibromyalgia, rheumatoid arthritis, axial spondyloarthritis, systemic lupus erythematosus, or psoriatic arthritis. Patients with a history of cancer were excluded. On drug opioid-related hospital admissions were identified using Hospital Episode Statistics Admitted Patient Care and Accident and Emergency Attendances data and included harms such as opioid poisoning, fractures, falls, delirium, sleep disorders, and constipation. Candidate predictors were selected based on clinical relevance from prior literature. Patients were considered ‘at risk’ from the time they received their first new opioid prescription and were censored at death or loss of follow-up. Predictive models using Cox proportional hazards (CPH) and Random Survival Forest (RSF) were developed using time-to-event patient data. Predictive performance of the models was evaluated, including calculating the area under the receiver characteristic operator curve (AUC) using 5-fold cross validation.

Results: Over the 15-year study period, 1,127,357 unique patients (60.7% female, median age [IQR]: 62 [25]) who were newly initiated on opioids were analysed. Opioid-related hospitalisations occurred in 39,541 patients (3.5%). Key risk factors identified in the CPH model included very high initial morphine milligram equivalents (≥200 MME/day; hazard ratio [HR]: 5.26, 95% CI: 5.03–5.49), high initial MME/day (120–199 MME/day; HR: 4.54, 95% CI: 4.23–4.88), high Charlson Comorbidity Index scores (HR: 2.31, 95% CI: 2.21–2.42), a history of suicide attempts or self-harm (HR: 2.07, 95% CI: 1.95–2.19), and alcohol dependence (HR: 1.77, 95% CI: 1.67–1.87). The CPH and RSF models demonstrated comparable predictive performance, with mean AUCs of 0.81 and 0.82, respectively.

Conclusion: High MME/day at initiation, multimorbidity, and histories of self-harm and alcohol dependence were found to be top factors associated with opioid-related hospitalisation. Greater vigilance where these factors exist could guide targeted interventions to reduce opioid-related hospitalisations. In this study both machine-learning and CPH models demonstrated good predictive performance.

REFERENCES: [1] Pierce, M., van Amsterdam, J., Kalkman, G. A., Schellekens, A. & van den Brink, W. Is Europe facing an opioid crisis like the United States? An analysis of opioid use and related adverse effects in 19 European countries between 2010 and 2018. Eur Psychiatry 64 , e47 (2021). https://doi.org:10.1192/j.eurpsy.2021.2219 .

[2] Friebel, R. & Maynou, L. Trends and characteristics of hospitalisations from the harmful use of opioids in England between 2008 and 2018: Population-based retrospective cohort study. Journal of the Royal Society of Medicine 115 , 173-185 (2022). https://doi.org:10.1177/01410768221077360 .

[3] Huang, Y.-T., Jenkins, D. A., Peek, N., Dixon, W. G. & Jani, M. High frequency of long-term opioid use among patients with rheumatic and musculoskeletal diseases initiating opioids for the first time. Annals of the Rheumatic Diseases , ard-2023-224118 (2023). https://doi.org:10.1136/ard-2023-224118 .

Acknowledgements: Funded by a FOREUM– Foundation for Research in Rheumatology Career Research Grant and the National Institute for Health and Care Research (NIHR).

Disclosure of Interests: None declared.

© The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ). Neither EULAR nor the publisher make any representation as to the accuracy of the content. The authors are solely responsible for the content in their abstract including accuracy of the facts, statements, results, conclusion, citing resources etc.