Implementation of Overdose Prevention in Maryland: Implications for Resource Allocation, Program Scale-Up, and Evaluation.

Overdose mortality in the United States continues to climb, with Maryland being one of the hardest hit states. We summarized implementation of overdose prevention and response programs in Maryland and identified associations between opioid overdose deaths by jurisdiction in 2019 and implementation of overdose programs by 2021. Data on program implementation are from Maryland's Opioid Operational Command Center (OOCC) Program Inventory. OOCC coordinates the state's response to overdose, and their Program Inventory tracks implementation of 145 programs across 12 domains (e.g., public health, education, and judiciary), including 10 programs designed to broaden naloxone access. The level of program implementation was dichotomized as substantial implementation versus other levels (i.e., partial, planned, and none). We estimated associations between per capita opioid overdose deaths and substantial implementation of: all 145 programs in the Inventory, programs within each of 12 domains, and 10 naloxone programs. Data on program implementation and overdose mortality are summarized at the jurisdiction level. Across jurisdictions, the median proportion of programs with substantial implementation was 51% across all programs and 70% among naloxone programs. Overdose mortality was associated with subsequent substantial implementation of programs within the public health domain (p = .04), but not in the other 11 domains. We did not find evidence that per capita overdose deaths in 2019 spurred overdose program implementation by 2021, with the exception of public health programs. The OOCC Program Inventory is a novel way to track implementation across jurisdictions. Findings can inform the implementation and evaluation of overdose programs in other jurisdictions across the United States.

An Exploration of Emergency Department Visits for Home Unintentional Injuries Among Children With Autism Spectrum Disorder for Evidence to Modify Injury Prevention Guidelines.

OBJECTIVES: Current childhood injury prevention guidance is anchored by a child's age. For example, children are considered at high risk for falls at ages 4 years and less, and guidance for prevention focuses on these ages. However, these guidelines may not be adequate for children with autism spectrum disorders (ASD). METHODS: This retrospective chart review examined injury characteristics for children with ASD receiving treatment in a pediatric emergency department between 2014 and 2016. Bivariate statistics determined injury demographic correlates. Chart narratives were also coded using traditional content analysis to determine the mechanism that caused the home injury. RESULTS: The sample (27 cases) was mostly male (89%), Black (48%), with a mean age of 7.8 (SD, 4.9) years. The most common mechanism was a fall (44%), followed by self-injurious behavior (33%), and then burns (22%). All cases identified at least 1 mechanism contributing to the injury, in 2 cases, 2 items were mentioned. Of the 29 items identified, most involved a house feature that was not stairs (24%); some included stairs, furniture, or a combination of foreign object and grill (21%); and few identified food/beverage/liquid (11%). CONCLUSIONS: Analyses suggest that unintentional home injury prevention for children with ASD may require prevention guidance extended through older ages.

Impact of unsolicited reporting notifications on providers' prescribing behavior: An experimental study on Maryland PDMP data.

BACKGROUND: Unsolicited Reporting Notifications(URNs) have been a component of Maryland's Prescription Drug Monitoring Program (PDMP) since 2016. We evaluated the effect of URNs on providers' prescription behaviors. METHODS: This is a quasi-experimental study of providers who were issued at least one URN from January 2018 to April 2021. Providers for whom URNs were not successfully delivered were designated as a comparison group. The outcome variables were average daily opioid and benzodiazepine prescriptions, average morphine milligram equivalents per patient, and proportion of overlapping opioid and benzodiazepine, either with or without muscle relaxant prescriptions. Changes were compared before versus after the issuance of a URN among the intervention and comparison groups using "Generalized Estimation Equation" and "Generalized Linear" Models. We also conducted stratified analyses by types of URN, including notifications for multiple provider episodes (MPE), overdose fatality (ODF), and dangerous drug combinations (DDC). RESULTS: The average daily number of opioids prescriptions (3.3% decrease in the intervention group vs 22.7% increase in the comparison group, P<0.001), co-prescription of opioids and benzodiazepines either with muscle relaxants (68.0% decrease vs. 36.1% decrease, P<0.001), or without muscle relaxants (6.0% decrease vs. 16.3% increase, P<0.001), significantly reduced after the first URN regardless of the type of URN. Stratified analysis by types of URNs showed that ODF and DDC URNs had a significant effect on most of the outcomes of interest. CONCLUSION: The findings suggest that unsolicited reporting, especially particular types of URNs including ODF and DDC, is associated with subsequent changes in unsafe prescribing behaviors.

Disruptions to naloxone training among lay and occupational responders in Maryland during the emergence of COVID-19: early impacts, recovery, and lessons learned.

Background: : Opioid overdose death rates increased during the COVID-19 pandemic. Disruptions in community-based naloxone trainings could have reduced the likelihood of overdose reversal and increased the chances of a fatal overdose. We investigated changes in the number of people trained in naloxone administration and distribution in Maryland before, during, and after COVID-related stay-at-home orders. Methods: : Data on naloxone training are from the Maryland Department of Health. We used interrupted time series models to estimate changes in average monthly number of people trained: [1] pre-interruption (4/2019-3/2020), [2] 1-month post-interruption (4/2020-5/2020), and [3] 12-months post-interruption (4/2020-3/2021). Trainees were classified as lay (e.g., people who use drugs) or occupational (e.g., law enforcement officers and harm reduction workers) responders. Results: : There were 101,332 trainees; 54.1% lay, 21.5% occupational, and 23.4% unknown responder status. We observed a decrease in the average monthly number of trainees in the pre-interruption period (-235, p<0.001), a larger decrease of 93.2% during the 1-month post-interruption (-846, p=0.013), and an increase 12-months post-interruption (+217, p<0.001). There was a significant decrease among occupational responders 1-month post-interruption, and a significant increase among lay responders in the 12-month post-interruption period. Conclusions: : Findings suggest a marked decrease in naloxone trainees immediately after stay-at-home order, followed by a moderate rebound in the 12-months after stay-at-home order. The decrease in occupational responders trained may have limited access to naloxone, but would likely have been offset by increases in number of lay responders trained. Strengthening lay and occupational responder connections could maintain naloxone distribution during public health crises.

Unsolicited reporting notifications (URNs) through Maryland's prescription drug monitoring program (PDMP): Characteristics of providers.

Objectives: Unsolicited reporting is the activity of analyzing Prescription Drug Monitoring Program (PDMP) data and then sending unsolicited reporting notifications (URNs) to prescribers to notify them of their outlier prescribing behavior. We aimed to describe information about prescribers who were issued URNs. Methods: A retrospective study of Maryland's PDMP data from Jan.2018-Apr.2021. All providers who were issued ≥ one URN were included in analyses. We summarized data on types of URNs issued by provider type and years in practice using basic descriptive measures. We also performed logistic regression analysis to provide odds ratio and estimated marginal probability of issuing ≥ one URN to providers in the Maryland health care workforce in comparison with physicians as reference group. Results: A total of 4,446 URNs were issued to 2,750 unique providers. Odds ratio (OR) and the population estimated probability of issuing URNs were higher among nurse practitioners [OR: 1.42, 95% Confidence Interval (CI): 1.26-1.59] followed by physician assistants [OR: 1.87, 95% CI: 1.69-2.08], compared to physicians. Physicians and dentists with >10 years in practice comprised the majority of providers who were issued URNs (65.1% and 62.6%, respectively), while majority of nurse practitioners had been in practice for <10 years (75.8%). Conclusion: Findings indicate a higher probability of issuing URN for Maryland's physician assistants and nurse practitioners, compared to physicians, and an overrepresentation of physicians and dentists with longer and nurse practitioners with shorter practice experience. The study suggests education programs on safer prescribing practices and management of opioids should target certain types of providers.

Past 15-year trends in adolescent marijuana use: Differences by race/ethnicity and sex.

BACKGROUND: The potential for increases in adolescent marijuana use is an important concern given recent changes in marijuana policy. The purpose of this study was to estimate trends in marijuana use from 1999 to 2013 among a national sample of US high school students. We examine changes over time by race/ethnicity and sex. METHODS: Data are from the National Youth Risk Behavior Survey (YRBS), which involves biennial, school-based surveys that generate nationally representative data about 9th-12th grade students in the United States. Students self-reported sex, race/ethnicity, and marijuana use (i.e., lifetime use, past 30-day use, any use before age 13). We generated national estimates of the prevalence of marijuana use for the time period, and also tested for linear and quadratic trends (n=115,379). RESULTS: The prevalence of lifetime marijuana use decreased modestly from 1999 to 2009 (44% to 37%), and has increased slightly since 2009 (41%). Other marijuana use variables (e.g., past 30-day use) followed a similar pattern over time. The prevalence of past 30-day use from 1999 to 2013 for all groups and both sexes was 22.5%, and it was lowest among Asians and highest among American Indian/Alaska Natives. Although boys have historically had a higher prevalence of marijuana use, results indicate that male-female differences in marijuana use decreased over time. CONCLUSION: Despite considerable changes in state marijuana policies over the past 15 years, marijuana use among high school students has largely declined. Continued surveillance is needed to assess the impact of policy changes on adolescent marijuana use.