Modeling opioid overdose events recurrence with a covariate-adjusted triggering point process

Area of Science:

• Public Health
• Epidemiology
• Biostatistics

Background:

• Substance use disorder, especially opioid-related, presents a significant public health challenge in the U.S.
• Accurate prediction of opioid overdose events is vital for timely healthcare interventions.
• Existing research has not quantitatively modeled the impact of prior overdose events on future occurrences.

Purpose of the Study:

• To develop and evaluate a novel statistical model for predicting opioid overdose events.
• To simultaneously assess various risk factors and the triggering effect of past events on future opioid overdoses.
• To improve risk stratification for patients with opioid use disorder.

Main Methods:

• Proposed a covariate-adjusted triggering point process model.
• Utilized U.S. state-wise Medicaid reimbursement claims data for model assessment.
• Compared the proposed model's performance against established prediction models.

Main Results:

• The proposed model demonstrated superior prediction accuracy, achieving the lowest Mean Absolute Errors and Mean Absolute Percentage Errors for 30- to 180-day-ahead predictions.
• The study confirmed the statistical significance of incorporating a triggering mechanism for predicting recurrent opioid overdose events.
• The triggering mechanism accounted for approximately 47% of event recurrence on average.

Conclusions:

• The covariate-adjusted triggering point process model offers enhanced accuracy for predicting opioid overdose events.
• Accounting for the temporal clustering (triggering effect) of past events is critical for improving predictive models.
• This approach can aid healthcare providers in better identifying and managing patients at high risk of recurrent opioid overdose.