Geriatric Pharmacoepidemiology

6A-02 - Adverse Drug Events of Opioid Treatment Strategies for Pain After Motor Vehicle Crash

Tuesday, August 26, 2025

1:45 PM - 2:00 PM ET

Co-Author(s)

AZ

Andrew R. Zullo

Brown University School of Public Health, United States

Presenting Author(s)

SB

Steven Balog

Brown University School of Public Health, United States

Co-Author(s)

MK

Marzan A. Khan

Brown University School of Public Health, United States
MP

Melissa Pfeiffer

Center for Injury Research and Prevention, Children's Hospital of Philadelphia, United States
MR

Melissa Reister

Brown University School of Public Health, United States
AC

Allison E. Curry

Center for Injury Research and Prevention, Children's Hospital of Philadelphia, United States
FB

Francesca Beaudoin

Brown University School of Public Health, United States
AO

Arman Oganisian

Brown University School of Public Health, United States
BM

Brandon D.L. Marshall, PhD

Professor
Department of Epidemiology, Brown University School of Public Health, United States
PD

Patience Dow

Brown University School of Public Health, United States
BD

Beth Dana

Brown University School of Public Health, United States
NJ

Nina Joyce, PhD

Assistant Professor
Department of Epidemiology, Brown University School of Public Health, United States

Background: Many motor vehicle crash (MVC)-related injuries occur among older adults every year in the United States, leading to acute or chronic pain. Although opioids are a mainstay of pain management after MVC in older adults, they have significant risks of adverse drug events (ADEs) and the optimal post-MVC opioid treatment strategies are unknown.

Objectives: To evaluate the effects of three pairs of key opioid treatment strategies on opioid-related ADEs following MVC among older adults.

Methods: We linked data on New Jersey drivers’ licensing and police-reported crashes with Medicare claims to emulate three target trials consisting of crash-involved drivers aged >67 years from May 1, 2007 to November 16, 2017 who initiated an opioid within 10 days after MVC. The unit of analysis was a person-crash, where persons are eligible for each emulation multiple times and can contribute multiple crashes. We evaluated the comparative risks of ADEs associated with three sets of opioid prescribing strategies: 1) greater versus lesser days’ supply (≥6 days versus ≤5 days), 2) higher versus lower doses (≥30 morphine milligram equivalents (MME) versus < 30 MME) and 3) tramadol versus other opioids over 45 days of follow-up. The outcome was opioid-related ADEs, including dependence, respiratory insufficiency, and accidental overdose. We estimated inverse probability of treatment weights to correct for potential confounding and selection bias. Covariates included in the weight estimation encompassed demographics, indications for opioids, comorbidities, frailty, prior medications, calendar time, and prescriber specialty. Using the predicted values of marginal structural models, we estimated weighted intention-to-treat risks and risk ratios. We constructed 95% confidence limits (CLs) using the non-parametric percentile bootstrap method.

Results: Our study included 510 person-crashes [mean (standard deviation) age: 76.12 (6.60) years, 59.8% female, 90.4% White race]. The weighted risks (95% CLs) of opioid-related ADEs were: days’ supply [≤5 days: 5.25 (2.93, 8.00), ≥6 days: 3.20 (1.21, 6.70)]; dosing [ < 30 MME: 4.28 (2.38, 7.36), ≥30 MME: 5.13 (2.80, 8.02)]; tramadol [non-tramadol: 5.59 (3.42, 8.06), tramadol: 10.37 (1.35, 24.27)]. Risk ratios (95% confidence limits) were 1.64 (0.64, 4.57), 0.83 (0.45, 1.76), and 0.54 (0.21, 4.01) for days’ supply, dose, and tramadol treatment strategies, respectively.

Conclusions: Different opioid treatment strategies did not produce significant differences in the risk of ADEs after MVCs, but there was considerable uncertainty around the effect measures. Prescribers should continue to judiciously minimize dose and duration when prescribing opioids for MVC-related pain in older adults.