Importance Opioid misuse remains a critical public health concern, associated with increased risk of overdose, psychiatric comorbidity, and societal costs. While machine learning (ML) analyses have been used to identify features associated with opioid misuse across various clinical settings, few studies have examined both modifiable and non-modifiable features. Understanding modifiable features may inform targeted prevention and intervention strategies to reduce opioid misuse.

Objective To identify key features associated with opioid misuse severity in patients with chronic pain taking long-term prescription opioids.

Design, Setting, and Participants This analysis used data from clinical trials investigating a skills-based pain management intervention. Participants were 314 community-dwelling adults with chronic pain who were taking daily opioid medications (≥ 10 morphine-equivalent daily dose) for at least 3 months. Data was extracted from baseline assessments.

Outcomes and Measures Opioid misuse severity was assessed using the Current Opioid Misuse Measure (COMM). Thirty-six demographic and clinical features were evaluated, including PROMIS symptom domains (e.g., pain rating, pain interference, physical function, fatigue, sleep disturbance, depression, anxiety, and anger) as well as emotional ambivalence, pain catastrophizing, trauma exposure, and substance use.

Results Among the seven ML algorithms (Random Forest, XGBoost, Support Vector Regression, LASSO regression, Ridge Regression, Elastic Net, and Multilayer Perceptron), the Elastic Net model demonstrated the strongest performance, yielding the highest correlation with COMM scores (mean r = 0.61, 95% CI [0.40, 0.73]) and the lowest root mean square error (mean RMSE = 4.16, 95% CI [3.41, 4.80]). Feature ablation analysis identified anger (Δr = 0.053), emotional ambivalence (Δr = 0.022), pain catastrophizing (Δr = 0.008), and fatigue (Δr = 0.002) as the most influential features associated with the COMM scores. Shapley Additive exPlanations (SHAP) analysis confirmed that higher levels of these key features were associated with higher COMM scores.

Conclusions and Relevance Emotional factors, particularly anger, emerged as key features associated with the severity of opioid misuse. These findings suggest that interventions targeting emotion regulation, especially anger management, may reduce opioid misuse among individuals receiving long-term opioid therapy for chronic pain.

The authors have declared no competing interest.

This study was funded by NIH NIDA K24DA053564 (BDD) NIH NIDA T32DA035165 (YW, SK, SCM)

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Stanford Institutional Review Board approved this study and obtained informed consent from all participants before any study procedure.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data produced in the present study are available upon reasonable request to the authors.