Community Versus Hospital Opioid-Related Overdose Deaths in Illinois

Abstract

Objective

To better understand approaches to reducing mortality from the opioid epidemic, we analyzed in-hospital versus community opioid-related overdose deaths in Illinois.

Methods

We used data from the Statewide Unintentional Drug Overdose Reporting System (July 2017 through December 2018) to identify deaths that occurred in hospitals and communities (ie, homes or public spaces). We used census tract–level data for 34 Illinois counties to create bivariate mapping by overdose death rates. We used logistic regression to analyze the association of demographic and overdose characteristics with the likelihood of death in a hospital versus a community.

Results

During the study period, 2833 opioid-related overdose deaths occurred in 24 Illinois counties, 655 (23.1%) of which occurred in the hospital; of 2178 community deaths, 1888 (86.7%) occurred in the same census tract as the decedent’s recorded residence and 1285 (59.0%) occurred in the decedent’s home. Non-Hispanic Black people were 1.63 (95% CI, 1.27-2.10) times more likely than non-Hispanic White people to die in a hospital. Decedents from suburban Cook County and other Chicago suburban counties were significantly more likely to die in the hospital than decedents from Chicago or other Illinois counties. Documentation of a previous overdose, history of opioid use, and having bystanders present were significantly associated with hospital deaths. Evidence of a rapid overdose, fentanyl present, or prescription opioids were significantly associated with deaths in a community.

Conclusions

The high number of opioid-related overdose deaths in the community illustrates the need to decriminalize illicit drug use and facilitate treatment seeking. Establishing supervised safe consumption sites may have the biggest effect in reducing the number of opioid-related overdose deaths.

Reflecting national trends, the number of opioid-related overdose deaths in Illinois declined slightly, from 2167 in 2018 to 2098 in 2019. ^1,2 The proportion of opioid-related overdose deaths that involves fentanyl or fentanyl analogs is large (70%) and growing. ³ Despite a leveling off in the number of overdose deaths in Illinois since 2017, emergency medical services (EMS) observed an increase in nonfatal overdoses (from 13 616 in 2018 to 14 592 in 2019) ¹ and an increase in multiple (ie, ≥2) administrations of naloxone (from about 20% of EMS encounters in 2013 to about 30% of EMS encounters in 2019), a sign of increased toxicity. ⁴ Illicit opioids, as opposed to prescription opioids, were responsible for 84% of overdose deaths in Illinois in 2019, an indication that limits to prescribing may have only a modest effect on the opioid-related death rate. ⁵ Most recently, a new spike in opioid-related overdose deaths appears to be related to the coronavirus disease 2019 (COVID-19) pandemic. ⁶

Opioid-related overdose deaths in Illinois are concentrated in the Chicago metropolitan area, with additional hotspots in smaller cities in Illinois. Death rates are highest in racial/ethnic minority neighborhoods with the greatest economic hardship. ⁷ In Illinois in 2018, non-Hispanic Black people had twice the overdose mortality rate per 100 000 population (32.8) of non-Hispanic White people (16.8) and 3 times the rate of Hispanic people (9.9). Although the number of deaths among non-Hispanic White people decreased 6.5% from 2018 to 2019, the number of deaths among non-Hispanic Black and Hispanic people increased by 9.1% and 4.1%, respectively. Meanwhile, the number of emergency department (ED) visits and hospitalizations for opioid-related overdose increased from 2013 to 2018. ⁴ Opioid addiction has also resulted in hundreds of thousands of nonfatal hospital visits and billions of dollars in hospital charges in Illinois during the past several years. ⁴

The objective of this study was to describe differences between opioid-related overdose deaths occurring in the community (ie, homes and public spaces) and opioid-related overdose deaths occurring in the hospital ED or inpatient setting in Illinois. To our knowledge, no study has investigated characteristics for opioid-related overdose deaths that occur in a community versus in a hospital, such as the characteristics of overdose victims who were transported and received hospital treatment compared with those who died on the scene or were dead on arrival. Our findings have important implications for estimating the demand for and effectiveness of EMS and for forecasting the additional hospital addiction treatment capacity and other initiatives that will be needed to reduce the number of opioid-related overdose deaths.

Methods

Data Source

The Statewide Unintentional Drug Overdose Reporting System (SUDORS) is a product of the Centers for Disease Control and Prevention’s (CDC’s) Enhanced State Opioid Overdose Surveillance (ESOOS) program, which provides funding to enhance drug overdose surveillance processes. ⁸ As part of an ESOOS grant, Illinois is required to abstract data from death certificates, coroner/medical examiner reports, toxicology reports, and autopsy reports on unintentional overdose deaths in the state. CDC defines an opioid-related overdose death as a drug poisoning death in which the death certificate or coroner/medical examiner report indicates that an opioid contributed to the drug overdose death. This analysis is based on data from 2833 opioid-related overdose deaths logged in SUDORS from July 2017 through December 2018, representing deaths occurring in 24 counties with >85% of all opioid-related overdose deaths in Illinois. Data were available for overdoses that occurred in 34 counties in Illinois (injury site); decedents’ residences included 44 of 102 counties in Illinois. The location of death was coded in SUDORS as hospital (ED or inpatient) or community (any other location). A total of 1131 (4.6%) decedents were non-Illinois residents and 184 (6.5%) had unknown addresses. The overall population of Illinois census tracts with at least 1 decedent was 6.0 million, approximately half the population of Illinois. Publicly available, de-identified SUDORS data on decedents are exempt from institutional review board review.

Decedent Characteristics

The SUDORS data set contains information on the decedent’s year and month of death, sex (male, female), age (5-24, 25-34, 35-44, 45-54, ≥55), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, other/unknown), educational attainment (≤high school diploma, some college, ≥college degree, unknown), homeless status (yes/no), and marital status (married, not married). Decedents’ residence was categorized as Chicago, suburban Cook County, suburban collar counties (Lake, McHenry, Kane, DuPage, and Will counties), or any other Illinois county, which we denoted as “downstate,” although this area includes northwestern, central, and southern Illinois. Downstate counties were not rural but generally included smaller cities and metropolitan areas. Each decedent’s Illinois census tract of residence was also matched to the 2018 American Community Survey (ACS) 5-year file to identify the number of people in the census tract living at or below the federal poverty level. ⁹ Census-tract poverty level was stratified as <5.0%, 5.0%-9.9%, 10.0%-19.9%, ≥20.0%, or non-Illinois resident; for bivariate poverty area results, we excluded 184 (6.5%) decedents with missing information on residence.

SUDORS data also contain numerous variables from coroner/medical examiner reports documenting other circumstances related to the incident. These yes/no variables included homeless status, the presence of bystanders at the overdose scene, whether the decedent had recently been released from prison or a behavioral health treatment facility, a history of pain treatment, history of opioid use, previous overdoses, evidence of a rapid overdose, evidence of prescription opioid at the overdose scene, EMS at the scene, whether naloxone had been administered, and whether fentanyl was involved in overdose. Data on location of death (eg, halfway house, jail, home of family member, abandoned building, motel, sidewalk, tent, airport, gas station) were not consistently available in structured format.

Bivariate Mapping

We computed death rates per 100 000 population for all decedent residence census tracts with a hospital or community death using 2018 ACS 5-year population estimates. ⁹ We stratified community death rates as tracts with 18-month death rates of <15, 15-24, 25-49, 50-99, and ≥100 deaths per 100 000 residents. We stratified hospital 18-month death rates as 1-19, 20-49, and ≥50 per 100 000 population. Census tract–level death rates for each location were mapped jointly using Esri ArcGIS PRO software (Esri).

Statistical Analysis

We determined the significance of the association of demographic and coroner/medical examiner variables with location of death (hospital vs community) by using the Pearson χ² test. We conducted logistic regression analysis using variables that were significant at P < .15 to determine adjusted odds ratios (aORs) for the likelihood of hospital versus community death. We included decedents who were missing data on residence information, but had all other data, in the reference category for the zip code tabulation area poverty level and Illinois region variables in multiple logistic regression analysis; aORs for these 2 variables were virtually identical after excluding the 184 missing observations. We performed all analyses using R (RStudio Global Corp) and SPSS version 26 (IBM Corp).

Results

A total of 2833 opioid-related overdose deaths occurred in the SUDORS counties during the 18-month study period, including 655 (23.1%) hospital deaths and 2178 (76.9%) community deaths (Table). Of the 2178 community deaths, 1285 (59.0%) occurred in the decedent’s home and 1888 (86.7%) occurred in the same census tract as the decedent’s residence. Hospital deaths were classified as being in the ED (65.8%) or inpatient setting (34.2%). The number of community deaths ranged from 129 (December 2018) to 167 (September 2017); the number of hospital deaths increased from 36 (July 2017) to 49 (October 2018) (Figure 1).

Table.

Characteristics of opioid-related overdose deaths in 24 Illinois counties, July 2017–December 2018 ^a

Characteristics	All deaths (N = 2833)	Community deaths (n = 2178)	Hospital deaths (n = 655)	P value b	Adjusted odds ratio (95% CI) c
Year				.07
2017	34.4	35.3	31.5		1.0 [Reference]
2018	65.6	64.7	68.5		1.18 (0.96-1.44)
Demographic characteristics
Sex				.07
Male	74.2	75.0	71.5		0.83 (0.67-1.02)
Female	25.8	25.0	28.5		1.0 [Reference]
Marital status				.43
Married	14.2	13.9	15.1		—
Not married	85.8	86.1	84.9		—
Age, y				.85
5-24	8.2	8.3	7.8		—
25-34	24.9	25.2	24.1		—
35-44	22.4	22.4	22.6		—
45-54	24.4	24.5	24.0		—
≥55	20.0	19.6	21.5		—
Race/ethnicity				.003
Non-Hispanic White	54.0	55.6	48.5		1.0 [Reference]
Non-Hispanic Black	26.5	25.0	31.6		1.63 (1.27-2.10)
Hispanic	10.1	9.8	11.0		1.36 (0.98-1.87)
Other/unknown	9.4	9.5	8.9		0.89 (0.59-1.32)
Education				.01
≤High school diploma	62.2	62.0	62.9		1.0 [Reference]
Some college	19.2	19.3	18.9		1.12 (0.88-1.43)
≥College degree	5.2	5.9	2.9		0.62 (0.37-1.05)
Unknown	13.3	12.8	15.3		1.49 (1.08-2.05)
Decedent residenced,e				<.001
Chicago	34.3	35.4	30.5		0.87 (0.66-1.17)
Suburban Cook County	17.6	16.5	21.2		1.65 (1.21-2.24)
Collar counties	20.8	19.7	24.3		1.76 (1.30-2.38)
Downstate Illinois	22.7	23.8	18.8		1.0 [Reference]
Non-Illinois resident	4.6	4.5	5.2		1.02 (0.59-1.77)
Homeless status
Yes	4.3	4.4	4.1	.75	—
No	95.7	95.6	95.9		—
Percentage of population in census tract of decedent’s residence living at or below the federal poverty level d				.04
<5.0	13.2	13.9	11.0		1.0 [Reference]
5.0-9.9	18.5	18.0	20.2		1.47 (1.04-2.06)
10.0-19.9	21.5	21.4	21.7		1.37 (0.97-1.93)
≥20.0	35.7	36.3	33.6		1.24 (0.87-1.78)
Non-Illinois resident	4.6	4.5	4.1		1.02 (0.59-1.77)
Overdose history and circumstances e
Decedent had previous overdoses	11.3	10.1	15.3	<.001	1.67 (1.27-2.21)
Decedent had history of opioid use	67.5	65.7	73.6	<.001	1.35 (1.09-1.67)
Decedent recently released from prison	3.5	3.7	2.9	.34	—
Decedent recently released from a behavioral health treatment facility	2.6	2.8	2.1	.38	—
No EMS on scene	3.5	3.0	5.2	.01	1.93 (1.23-3.04)
The presence of bystanders at the overdose scene	28.8	26.7	35.7	<.001	1.55 (1.27-1.89)
Evidence of a prescription opioid at the overdose scene	21.0	25.0	7.6	<.001	0.23 (0.17-0.31)
Evidence of rapid overdose	8.0	8.4	6.7	.15	0.73 (0.51-1.05)
Fentanyl was involved in overdose	71.8	72.5	69.5	.13	0.72 (0.58-0.88)
Decedent had a history of pain treatment	6.2	6.5	5.3	.29	—
Naloxone was administered f	29.5	17.0	71.1	<.001	— g

Abbreviation: EMS, emergency medical services.

^aAll values are percentage unless otherwise indicated. Not all percentages total to 100 because of rounding. Data source: Enhanced State Opioid Overdose Surveillance program. ⁸

^bVariables with bivariate P value ≥ .15 were not included in the model.

^cUsing the Pearson χ² test of significance, with P < .05 considered significant.

^dCollar counties include Lake, Kane, DuPage, McHenry, and Will. Downstate counties include all other Illinois counties except collar counties and Cook County.

^eExcludes 131 (4.6%) decedents with missing data on census tract. Those decedents were included in the reference category for regression analyses.

^fThe reference group for each variable is the negative of the statement indicated. For example, the reference group for “Decedent had previous overdoses” is “Decedent had no previous overdoses.”

^gNaloxone administration was not included in the model because it may have occurred after the decedent arrived at the hospital.

Figure 1.

Mean number of opioid-related overdose deaths occurring in a hospital and in the community, in 34 counties in Illinois, July 2017–December 2018. Data source: Statewide Unintentional Drug Overdose Reporting System. ⁸

A total of 1350 residence census tracts had at least 1 death: 91 (6.7%) census tracts had a community death rate of ≥100 deaths per 100 000 residents, and 67 (5.0%) census tracts had a hospital death rate of ≥50 deaths per 100 000 population. Only 15 census tracts with the highest hospital death rates were also among census tracts with the highest community death rates. Chicago and suburban Cook County accounted for 1471 (51.9%) opioid-related overdose deaths from the SUDORS data set (Figure 2), suburban collar counties accounted for 499 (17.6%) opioid-related overdose deaths (Figure 3), and downstate areas accounted for 643 (22.7%) opioid-related overdose deaths (Figure 4). The largest clusters of tracts with high hospital death rates in Cook County were on Chicago’s west side (along an expressway extending from Chicago into the near western suburbs) and on the South side, extending into the south suburbs. Several tracts with high community death rates (≥100 per 100 000 residents) were on the far south side and southwest suburban areas, farther from hospitals in Chicago. Collar-county maps show clusters of tracts with high death rates in the Aurora/Joliet area, Kankakee, northeast Lake County, and DuPage County. Clusters of tracts with high death rates downstate included areas around Peoria, Springfield, East St. Louis, and Rockford.

Figure 2.

Opioid-related overdose deaths in Chicago and suburban Cook County, Illinois, July 2017–December 2018. Data source: Statewide Unintentional Drug Overdose Reporting System. ⁸

Figure 3.

Opioid-related overdose deaths in suburban Chicago counties, July 2017–December 2018. Data source: Statewide Unintentional Drug Overdose Reporting System. ⁸

Figure 4.

Opioid-related overdose deaths in selected downstate Illinois counties, July 2017–December 2018. Data source: Statewide Unintentional Drug Overdose Reporting System. ⁸

Decedent Characteristics by Location of Death

Non-Hispanic Black people composed 26.5% and Hispanic people composed 10.1% of all SUDORS opioid-related overdose deaths during the study period, although these racial/ethnic groups had approximately equal SUDORS area population proportions (Table). We found significant differences in aORs between hospital and community deaths by race/ethnicity but not by sex or year of death. A larger proportion of non-Hispanic Black decedents died in the hospital than in the community (31.6% vs 25.0%), whereas a larger proportion of non-Hispanic White decedents died in the community than in the hospital (55.6% vs 48.5%). Non-Hispanic Black decedents were 1.63 times more likely than non-Hispanic White decedents to die in the hospital.

Decedents for whom education level was unknown were significantly more likely to die in the community than in the hospital (Table). Compared with residents of downstate Illinois, residents of suburban Cook County (aOR = 1.65; 95% CI, 1.21-2.24) and collar counties (aOR = 1.76; 95% CI, 1.30-2.38) were significantly more likely to die in the hospital. Residents in census tracts with lower income levels had greater odds of dying in the hospital than residents in the census tract with the highest income level (<5.0% of the population living at or below the federal poverty level), but the odds were significant only in the census tract in which 5.0%-9.9% of the population lived at or below the federal poverty level (aOR = 1.47; 95% CI, 1.04-2.06).

Of the overdose variables included in the regression model, documentation of a previous overdose, history of opioid use, and the presence of bystanders at the overdose scene were significantly associated with dying in the hospital versus the community. Although having no EMS on the scene composed only 3.5% of the sample, it was significantly associated with hospital death versus community death (aOR = 1.93; 95% CI, 1.23-3.04). Conversely, decedents who had evidence of a rapid overdose or that fentanyl was involved were significantly more likely to die in the community than in a hospital. Decedents with evidence of prescription opioids (21.0% of the sample) were 77% less likely to die in a hospital than in the community. People who died in the hospital were more than twice as likely as people who died in the community to have received naloxone.

Discussion

The SUDORS data showed that the number of opioid-related overdose deaths declined somewhat in Illinois from July to December 2017, which may have been related to the numerous medical and public health interventions implemented to reduce addiction-related morbidity and mortality. These efforts included more community-based behavioral health services for opioid use disorders ¹⁰ ; greater distribution of naloxone, including community and ED naloxone distribution programs; pharmacy dispensing via standing order ¹¹ ; increasing primary care physician buprenorphine training as part of medication for addiction treatment; and increased education and physician awareness about the risks associated with prescription opioids. ¹² However, despite these efforts, the numbers and rates of opioid-related overdose deaths remain high, and most deaths occur in the community. The number of opioid-related overdose deaths in Illinois began to rise in 2020 and may have reached a new peak since the onset of the COVID-19 pandemic in spring 2020, apparently related to more dangerous street drugs than were sourced before the epidemic, more frequent use of drugs alone, and lack of access to treatment for people recovering from drug addiction. ¹³

Our results are consistent with previous geospatial research findings of a bifurcated drug market in Cook County, with fentanyl-related overdose deaths concentrated in specific, racially/ethnically segregated Chicago neighborhoods, whereas non–fentanyl-related overdose deaths were concentrated in suburban areas. ¹⁴ This discrepancy in fentanyl-related overdose deaths may be one reason that suburban areas had a higher hospital death rate than community death rate, whereas Chicago had a higher community death rate than hospital death rate. The proportion of fentanyl-related overdose deaths was 77.5% in Chicago compared with 68.7% in suburban Cook County, 66.0% in collar counties, and 69.6% downstate. Similarly, only 5.9% of Chicago residents had overdosed previously, whereas 19.5% of collar-county residents and 15.7% of downstate residents had overdosed previously; evidence of prescription drug use was found among only 13.9% of decedents in Chicago but 29.4% of decedents in collar counties and 26.9% of decedents in downstate Illinois. These findings underscore differences in the opioid overdose crisis across Illinois regions.

It is unclear why non-Hispanic Black decedents were more likely than other racial/ethnic groups to be found alive and transported to the hospital. Post-hoc analyses revealed that a greater percentage of non-Hispanic Black decedents than decedents in other racial/ethnic groups had bystanders present at the overdose scene, although the difference was small (30.7% vs 28.0% for non-Hispanic White people). A significantly lower proportion of decedents with a documented history of overdoses (vs no documented history of overdoses) (–6.3%, P = .02) and opioid use (vs no opioid use) (–9.2%, P < .001) died in the hospital. Similarly, non-Hispanic Black decedents had a higher rate of fentanyl involvement in the overdose than residents in other racial/ethnic groups (9.8%, P < .001). However, evidence of prescription drug use, which was associated with lower odds (0.23) of in-hospital death compared with no evidence of prescription drug use, was lower among non-Hispanic Black decedents (10.8%) than among non-Hispanic White decedents (27.3%; P < .001).

The main finding of this study was that most decedents in the SUDORS data set died at the overdose scene and were pronounced dead when EMS arrived. These people died in their homes, the homes of friends and family, hotels, public spaces, or abandoned properties. A review of a sample of case narrative reports revealed few community deaths in which the decedent was transported to the hospital but was dead on arrival. Most (almost three-quarters) community decedents died alone, whereas approximately one-quarter had bystanders present but died before EMS arrived. Although many people experiencing an overdose who survive long enough to receive hospital treatment die, most are saved thanks to improved treatment protocols and earlier naloxone administration. ¹² However, people who die in the community before they can receive medical care, including many who had previous hospital encounters, have a need for a different prevention strategy.

Harm Reduction Strategies

The preponderance of community deaths speaks to the urgent need for a public health perspective focused on harm reduction, such as decriminalizing illicit drug use and establishing safe consumption sites. ¹³ Decriminalization could reduce the costs of policing, criminal prosecution, and incarceration; curb violence in the drug trade; and improve the safety of emergency responders. By reducing stigma, decriminalization would allow people who use drugs to more freely seek treatment. Such measures now include pretrial or postarrest diversion programs and assistance with housing, employment, and childcare services. Our findings indicate a need for additional approaches that focus on increasing the safety of drug users so that they can survive to participate in treatment and intervention services. Implementation of safe consumption sites is an especially applicable strategy for reducing fentanyl-associated mortality.

Safe consumption sites are also referred to as supervised consumption, overdose prevention, or safe injection sites. At these facilities, people can self-administer drugs that they obtain on their own. These facilities are often not run by medical professionals but can provide sterile use equipment. The goal of these facilities is harm reduction: reduce the risk of overdose death, reduce the transmission of blood-borne pathogens such as HIV and hepatitis C caused by sharing nonsterile needles and other equipment, and provide life-saving fentanyl test strips. ¹⁵ One study of an unsanctioned safe consumption site found that in 2 years, it prevented more than 2300 instances of injections that otherwise would have occurred in public spaces. ^16,17

The first pilot safe consumption site in Canada opened in Vancouver in 2003 with medical personnel. ¹⁸ The pilot found that people who used the facility used safer injection practices than street use with its attendant risk of infection and overdose, were referred for community resources and medical treatments, and had increased rates of starting treatment for substance use disorder. ¹⁸ From March 2004 through August 2005, the overdose rate was 1.3 per 1000 injections, of which heroin comprised 70% of overdoses, and naloxone was administered in 30% of overdoses, with no overdose-related deaths. ¹⁸ The opening of this site was associated with a reduction in overdose mortality of 35% during the next 2 years in the surrounding area. ¹⁹ Since then, multiple sites have opened in Canada, despite often restrictive legal requirements. ²⁰

A 2014 literature review demonstrated that across numerous studies, safe consumption sites were cost effective and had many benefits, including increased use of safer injection practices, increased referral to and initiation of substance use disorder treatment services, and reduced effects of public use, such as discarded syringes and resident complaints. ²¹ Safe consumption sites also showed large improvements in the rates of overdose death and overdose ambulance calls. ²¹ A 2017 systematic review showed similar results, including major reductions in the number of opioid-related overdose deaths in the areas around safe consumption sites and no recorded overdose deaths at those sites, despite safe consumption site users often being at high risk for overdose death. ²²

One legal safe consumption site was initially approved in 2019 in Philadelphia, but its fate is uncertain. ^23,24 Legislation legalizing safe consumption sites has been introduced in numerous states and has gained support from the American Medical Association. ²⁵ In Illinois, Governor J.B. Pritzker signed an executive order allocating funding to combat the opioid crisis, including potentially opening safe consumption sites, upon taking office in January 2020. ²⁶ That so many community deaths in Illinois did not have bystanders present, that so many died alone, is a powerful argument for the life-saving potential of safe consumption sites.

Limitations

This study had several limitations. First, the data were from decedents only and, therefore, offer no evidence on factors that may predict overdose survival. Second, although medical examiner reports contain a written narrative of the circumstances of a death, many characteristics of decedents are unknown and, in particular, SUDORS data lack important historical data on decedents’ drug use and medical treatment before their fatal overdose. Third, some variables such as homelessness and former incarceration are undercounted. Fourth, although medical examiner narratives include some examples of overdose victims who were found alive at the scene but died before arrival at the hospital, it was impossible to calculate the proportion of community deaths in which the decedent was found alive. Finally, SUDORS data do not include all opioid-related overdose deaths in Illinois and have limited data for many rural downstate counties.

Conclusion

Most opioid-related overdose deaths in the SUDORS data set occurred in the community. Data on the demographic and geographic variation in overdose death rates can be used to guide the location of safe consumption sites in areas with the highest rates of opioid-related overdose deaths. In Illinois, safe consumption sites should be considered in areas where the highest community overdose mortality rates occurred. To achieve the necessary legal and legislative reforms, epidemiologists and health care professionals will have to overcome substantial opposition from law enforcement and the public. Dissemination of harm reduction evidence will be crucial to the success of these efforts. Data from SUDORS can be used to monitor opioid-related overdose deaths and evaluate the effect of interventions such as safe consumption sites. Studies that use SUDORS data can contribute to coordinated harm reduction approaches to reducing opioid misuse and overdose.