UC Irvine

The opioid crisis remains one of the most serious public health challenges in the United States. Over the past two decades, opioid-related harms have evolved from a crisis driven largely by prescription pain relievers to one increasingly shaped by illicitly manufactured opioids. This evolving epidemic has prompted growing interest in harm-reduction approaches aimed at reducing morbidity and mortality. This dissertation examines whether harm reduction can reduce opioid-related harms at multiple intervention points in the United States. Guided by harm reduction theory and the risk environment framework, I evaluate three distinct strategies: provider education, naloxone co-prescribing, and fentanyl test strip (FTS) policy reform. Together, these interventions target different stages of the overdose pathway, from reducing unnecessary opioid exposure in clinical care to expanding overdose response and drug-checking capacity in community settings. The first chapter evaluates whether psychiatric continuing medical education for primary care providers reduces opioid prescribing among their patients. Using Inland Empire Health Plan data on 11,975 patients linked to 180 providers in Southern California, I compare patients treated by providers who completed the Train New Trainers Primary Care Psychiatry Fellowship with patients treated by non-participating providers. After adjustment for provider characteristics, seasonality, and time trends, patients of TNT-trained providers received 0.226 fewer opioid prescriptions per quarter after training, equivalent to an estimated reduction of approximately 2,706 prescriptions per quarter and more than 10,800 annually across the study population. These findings suggest that provider-focused psychiatric and substance use training may function as an upstream harm-reduction strategy by improving clinical decision-making and reducing downstream reliance on opioids. The second chapter evaluates California’s AB 2760 naloxone co-prescribing mandate using interrupted time-series methods applied to quarterly emergency department visits and mortality data from the California Overdose Surveillance Dashboard. I also applied interrupted time-series and synthetic control methods to monthly mortality data from the National Vital Statistics System. After policy implementation, opioid-related emergency department visits exceeded expected levels in multiple quarters of 2019. Mortality models indicated no declines in prescription-opioid, synthetic-opioid, and any-opioid overdose mortality. These findings suggest that naloxone co-prescribing alone did not offset worsening overdose risk in California during the fentanyl era. The third analytic chapter examines whether state FTS policies reduce overdose mortality in the United States using 2016–2023 NVSS mortality data, ARIMA pre-whitening, and Callaway-Sant’Anna difference-in-differences models under staggered adoption. Overall, I find that FTS policy domains showed different relationships with overdose mortality. In the pre-COVID period, the signal emerged for any opioid and any drug overdose mortality: state laws that authorized FTS sale or free distribution aligned with statistically detectable reductions, whereas possession decriminalization showed no comparable association. In the post-COVID period, estimates across all three domains shifted toward the null with wide uncertainty, consistent with a setting in which pandemic-era disruptions and rapidly changing drug markets may have overwhelmed or obscured incremental benefits of legal reforms. Heterogeneity analyses further suggest that associations were stronger in large-population states and in states with lower baseline opioid mortality. Taken together, my findings indicate that harm reduction policies and interventions to curb the opioid epidemic have shown mixed and often limited population-level effects. Future avenues of research could include more precise measurement of policy implementation, access, and uptake in practice; closer attention to heterogeneity across populations and places; and stronger evaluation of how these interventions operate under changing drug-market conditions.