Community overdose surveillance: Fentanyl involvement in overdose deaths in rural Michigan

Highlights

• •Fentanyl was detected in approximately 70% of all deaths in 2020.
• •Overdose deaths increased by 72.4% between 2019 to 2020.
• •Stimulant use is rising and most stimulant-related deaths included fentanyl.

Abstract

Purpose

Examine fatal overdose toxicology trends to contribute toward understanding the outreach and treatment needs of people who use drugs in rural areas.

Methods

We describe toxicology results from overdose deaths that occurred between January 1, 2018, and December 31, 2020, in 11 rural counties in Michigan, a state with relatively high rates of overdose mortality. One-way ANOVA with Tukey's HSD posthoc tests were used to test statistically significant differences in the frequency of substances detected between years.

Findings

Decedents (N = 107) were male (72.9%), White (96.3%), non-military (96.3%), unemployed (71.0%), married (73.9%), and had a mean age of 47 years old. The number of observed overdose deaths increased considerably from 2019 to 2020, with an increase of 72.4%. Fentanyl was the most common substance detected and had a 94% increase during the three-year period to present in 70% of all the deaths in these counties in 2020. Among the deaths we examined where cocaine was detected, 69% also contained fentanyl, and in deaths where methamphetamine was detected, 77% also contained fentanyl.

Conclusion

Findings could inform rural health and outreach initiatives aimed at reducing overdose risks by providing education on the risks of stimulant and opioid couse but also the widespread saturation of illicit drugs that contain fentanyl. Lowthreshold harm reduction interventions are discussed amid limited prevention and treatment resources in rural communities.

1. Introduction

The overdose epidemic remains one of the most pressing public health issues. In 2020, the United Nations Office on Drugs and Crime (UNODC) World Drug Report estimated that approximately 585,000 people died of a drug overdose in 2017 Cross-cutting issues: Evolving trends and new challenges et al., 2020. In the United States (US), there have been approximately a half million deaths in the US over the past decade and over 107,622 deaths in 2021 alone (Mattson et al., 2021). Recent data suggest there are shifting trends relative to the type of opioids that were detected in rural overdose deaths. For instance, from 2015 to 2019 overdose deaths involving illicitly manufactured fentanyl and other opioids occurred in rural counties at an age-adjusted rate of 1.4 times fewer compared to urban counties (Hedegaard and Spencer, 2021). There is also evidence of rural-urban variation in the age-adjusted rates of overdose deaths where psychostimulants were detected. As of 2019, the rate of overdose deaths that involved methamphetamine in rural counties (6.7 per 100,000) was higher than the rate found in urban counties (4.8 per 100,000); whereas, age-adjusted rates of overdose deaths involving cocaine were lower in rural counties (2.4 per 100,000) compared to urban counties (5.3 per 100,000; (Hedegaard and Spencer, 2021).

Recent increases (i.e., 2019–2020) in the use of psychostimulants have been highlighted in Michigan (Magura et al., 2021) with research showing that these drugs are more frequently detected via urine testing among recently incarcerated detainees than opioids (Magura et al., 2021; Vohra et al., 2021). There is also evidence that overdose deaths involving psychostimulants have increased in Michigan from 1999 to 2019 (Hedegaard, 2022; Vohra et al., 2021); however, data used to examine polysubstance detection – such as a “goofball” the co-use of methamphetamine and heroin or fentanyl (Ivsins et al., 2022) – in overdose deaths often occurs in metropolitan jurisdictions (Nolan et al., 2019) and national-level data may have geographic biases due to suppressed cases (Tiwari et al., 2014; Volkow et al., 2022). Moreover, the extent to which polydrug co-use or contamination with fentanyl impact drug overdose mortality in rural counties is unclear (Park et al., 2021).

The lack of information, specifically toxicology data, on the specific substances causing overdoses in certain populations, such as rural communities or minority groups, contributes to a gap in research on the interaction between multiple substances. This may be due to limited resources for gathering and analyzing data (Quesinberry et al., 2019) Surveillance of rural overdose toxicology trends is important for several reasons (Tiwari et al., 2014; Volkow et al., 2022). It can help identify and track changes in overdose patterns over time, which can inform public health policy and resource allocation. Rural areas often have limited resources and providers offering substance use disorder treatment and overdose prevention (Lister et al., 2020; Lister and Lister, 2021; Parry et al., 2021) so more research on rural overdose toxicology trends – particularly trends in substances involved in polysubstance overdose fatalities – can help identify disparities in overdose rates between rural and urban areas, which can inform targeted efforts to address these disparities. There is evidence of fentanyl detection in postmortem toxicology across the US but rarely at granular county-level in rural areas.

Accordingly, in this study we examine toxicology records among overdose decedents during a 3-year period in 11 rural counties in Michigan. We use novel polydrug visualization efforts to display changes in overdose patterns involving multiple substances – including fentanyl, heroin, prescription opioids, cocaine, methamphetamine, and benzodiazepines. Understanding polysubstance use factors that are confined to rural areas may inform rural outreach initiatives aimed at preventing overdose fatalities.

2. Methods

2.1. Sample and procedures

Postmortem overdose toxicology records between January 1, 2018, and December 31, 2020, in 11 rural counties were obtained from the Mid-Michigan Medical Examiner Group. These data were obtained through a larger state funded grant program that concluded prior to the 2021 data collection period. These specific counties represent approximately one-fifth of all rural counties in Michigan and were selected based on their partnership with the Mid-Michigan Medical Examiner Group and their rural classifications, which were defined using the rural-urban continuum codes (RUCC) – a classification scheme that was developed by the U.S. Department of Agriculture (USDA).

The RUCC index assigns codes ranging from 1 to 9 to counties based on their degree of urbanization, proximity to urbanization, and population density, with code 1 representing the most urban areas and code 9 representing the most rural areas (U.S. Department of Agriculture Economic Research Service et al., 2020). Our sample was comprised of counties with RUCCs of 6 (n = 5, 45.45%), RUCC 7 (n = 4, 36.36%), and RUCC 9 (n = 2, 18.18%. Montcalm county (previously an RUCC 6) was included because of the similar population size to the other included counties despite its current RUCC of 2 following recent changes to its coding due to its proximity to the Grand Rapids-Kentwood metropolitan statistical area, of which has experienced rapid population growth over the past decade. An RUCC of 6 are counties that have a population of 2500–19,999 people and are adjacent to a metropolitan area, whereas an RUCC of 7 are counties that have a population of 2500–19,999 people but are note adjacent to a metropolitan area. Similarly, an RUCC of 8 has a population of less than 2500 people and is adjacent to a metropolitan area, while an RUCC of 9 is a county of less than 2500 people and is not adjacent to a metropolitan area (USDA Economic Research Service, 2020). For further information, a complete breakdown of RUCC categories is available online (USDA Economic Research Service, 2020). Each of the included county's RUCC are as follows: Alpena (RUCC=7); Clare (RUCC=6); Crawford (RUCC=7); Lake (RUCC=9); Mecosta (RUCC=6); Montcalm (RUCC=2); Montmorency (RUCC=9); Newaygo (RUCC=6); Oceana (RUCC=6); Ostego (RUCC=7); and Wexford (RUCC=7). See Fig. 1 for a map of the counties included in the current analysis.

Fig. 1.

Map of study counties.

In Michigan, the most rural counties are grouped into rural public health catchment areas as these counties do not have the resources to sustain their own county-level health department – unlike wealthier rural counties and more populated metro and urban areas. The counties included in this study are primarily situated in the northern lower-peninsula of the state and are representative of the larger rural health districts (Michigan Association for Local Public Health, n.d.). The Michigan Substance Use Vulnerability Index (MI-SUVI) is a standardized composite score used to measure each county's vulnerability to individual- and community-level adverse substance use outcomes (Michigan Department of Health and Human Services, 2020). This measure is comprised of substance use burden (i.e., overdose death rate; non-fatal overdose emergency department rate; opioid prescribing rate; and drug-related arrest rate), substance use resources (i.e., percentage of population within 30-minute drive of treatment center; percentage of population living within 15-minutes to a syringe service program; and buprenorphine prescribing rate), and social vulnerability (i.e., modified Centers for Disease Control and Prevention Vulnerability Index). As of 2020, most of the counties included in this analysis are the most vulnerable with MI-SUVI scores ranking within the 75th-100th percentile (i.e., Clare, Crawford, Lake, Montmorency, and Oceana) while several other counties are ranked within the 50th-75th percentile (i.e., Mecosta, Newaygo, and Ostego). Montcalm and Wexford counties were identified as being less vulnerable (i.e., 25th-50th percentile) and only Alpena county (0–25th percentile) was considered among the least vulnerable counties in the state. This study was determined to be non-human subjects research, and therefore exempt by university IRB.

Administrative data were utilized for this study. First, sociodemographic information on decedents’ age, sex, race, military status, employment status, and marital status was obtained using death certificate records. Other death certificate information that was included in our analysis was a decedent's place of death and injury (i.e., location of overdose event). The selection of sociodemographic variables was contingent on information that was captured in the death certificate records. Second, toxicology records provided data on the detection of substances present at the time of death. For the current paper, data collection of toxicology records was limited to substances that are commonly detected in overdose events.

All overdose deaths (N = 107) that occurred during the study period were included in the current analysis. Among the 107 deaths that occurred over the three-year study period, toxicology record data were provided specifically for 6 substances which were used in our analysis, including: fentanyl, heroin, prescription opioids, cocaine, methamphetamine, and benzodiazepines. During the three-year study period, all toxicology testing was completed by NMS Labs in two separate locations in Pennsylvania – greater than 90% of all cases were sent to the same lab. Regardless of the lab, all results adhered to the same detection threshold set by NMS Labs and each toxicology screen was individually interpreted by a licensed medical examiner. Each respective toxicology screen identifies all substances, including metabolites and analogues, present in the decedent at their time of death with detection thresholds determined by the ordered drug panels. All respective metabolites and analogues of primary drugs were considered positive for the primary drug. Fentanyl was detected with the fentanyl metabolite norfentanyl as well as the following analogs: 1) Carfentanil; 2) Acetyl fentanyl; 3) Acryl fentanyl; 4) Para fluorobutyryl fentanyl; 5) Methoxy acetyl fentanyl; 6) Furanyl fentanyl; 7) Butyry fentanyl; 8) Valeryl fentanyl; and 9) Fluoro fentanyl. In addition, the fentanyl precursor chemical 4-ANPP was detected along with a synthetic opioid known as U-47,700. Positive heroin cases were counted when 6-MAM (the primary metabolite for heroin) was detected. In the absence of 6-MAM, the detection of codeine and morphine without any other opioids were counted as positive heroin cases (Harruff et al., 2015; Ray et al., 2017). Prescription opioid types were categorized by the detection of codeine, morphine, oxycodone, oxymorphone, hydrocodone, hydromorphone, dihydrocodeine, and tramadol. There were no distinctions made between pharmaceutical and illicit fentanyl.

2.2. Analytic plan

Frequency and descriptive statistics of all fatal overdose cases were conducted to determine sociodemographic characteristics, the place of death, and the number of substances detected in each death. One-way ANOVA with Tukey's HSD post-hoc tests were used to test statistically significant differences in the frequency of substances detected between years. Following an examination of the substances detected over time, a post-hoc Pearson's chi-square test was conducted to determine a correlation between a syringe being found at the scene and a toxicology detection of fentanyl.

To analyze polydrug combinations, we used network analysis techniques previously described in Ray et al., (2020) and conducted the analysis in R by visualizing the frequency of co-occurrences among fentanyl, heroin, prescription opioids, cocaine, methamphetamine, and benzodiazepines. In the resulting visualization, the weight (thickness) of each edge (line) indicates the frequency (count) of cooccurrences between the connecting nodes each year. The node size indicates the frequency of each specific substance thus bigger nodes indicate more frequent polydrug occurrences while edge weights can be compared across years to examine changing frequencies of unique polydrug combinations.

3. Results

A full description of the study sample can be found in Table 1. The number of observed overdose deaths was relatively consistent from 2018 (n = 28) to 2019 (n = 29). However, there was a considerable increase in observed overdose deaths from 2019 to 2020, with a percentage increase of 72.4% (see Table 1). Decedents were male (72.9%), White (96.3%), non-military (96.3%), unemployed (71.0%), married (73.9%), and had a mean age of 47-years-old. The place of injury most frequently occurred at the decedent's own residence (71.0%), followed by other residence (12.1%), other location (9.3%), and then unknown location (7.5%). The place of death most often reported was the decedent's residence (69.2%), follow by ER/Outpatient (12.1%), other residence (11.2%), and Other (7.5%). Approximately 38.3% of deaths were reported to have a syringe(s) present and nearly 64.0% had other paraphernalia present.

Table 1.

Study sample characteristics.

	Total		2018		2019		2020
	N = 107		N = 28		N = 29		N = 50
Age
18–25	6	5.6%	1	3.6%	1	3.4%	4	8.0%
26–35	22	20.6%	3	10.7%	8	27.6%	11	22.0%
36–45	23	21.5%	6	21.4%	7	24.1%	10	20.0%
46–55	29	27.1%	11	39.3%	5	17.2%	13	26.0%
56–65	24	22.4%	6	21.4%	7	24.1%	11	22.0%
Over 65	3	2.8%	1	3.6%	1	3.4%	1	2.0%
Sex
Female	29	27.1%	9	32.1%	6	20.7%	14	28.0%
Male	78	72.9%	19	67.9%	23	79.3%	36	72.0%
Race/Ethnicity
White	103	96.3%	28	100.0%	28	96.6%	47	94.0%
Black	2	1.9%	0	0.0%	1	3.4%	1	2.0%
Hispanic	2	1.9%	0	0.0%	0	0.0%	2	4.0%
Education
< High School	11	10.3%	1	3.6%	1	3.4%	9	18.0%
High School/GED	11	10.3%	2	7.1%	2	6.9%	7	14.0%
> High School	11	10.3%	3	10.7%	3	10.3%	4	8.0%
Unknown	74	69.2%	22	78.6%	23	79.3%	30	60.0%
Employment
Employed	31	29.0%	6	21.4%	5	17.2%	20	40.0%
Unemployed	22	20.6%	5	17.9%	6	20.7%	11	22.0%
Disabled	1	0.9%	0	0.0%	0	0.0%	1	2.0%
Retired	1	0.9%	1	3.6%	0	0.0%	0	0.0%
Unknown	52	48.6%	16	57.1%	18	62.1%	18	36.0%
Marital Status
Never Married	29	27.1%	5	17.9%	8	27.6%	16	32.0%
Married	17	15.9%	7	25.0%	3	10.3%	7	14.0%
Married but separated	5	4.7%	0	0.0%	2	6.9%	3	6.0%
Widowed	1	0.9%	1	3.6%	0	0.0%	0	0.0%
Divorced	23	21.5%	5	17.9%	6	20.7%	12	24.0%
Unknown	32	29.9%	10	35.7%	10	34.5%	12	24.0%
Military History (Yes)	2	1.9%	0	0.0%	2	6.9%	0	0.0%
Place of Injury
Residence	76	71.0%	20	71.4%	21	72.4%	35	70.0%
Other Residence	13	12.1%	3	10.7%	3	10.3%	7	14.0%
Unknown	8	7.5%	3	10.7%	3	10.3%	2	4.0%
Other	10	9.3%	2	7.1%	2	6.9%	6	12.0%
Place of Death
Residence	74	69.2%	20	71.4%	21	72.4%	33	66.0%
Other Residence	12	11.2%	2	7.1%	3	10.3%	7	14.0%
ER/Outpatient	13	12.1%	5	17.9%	3	10.3%	5	10.0%
Other	8	7.5%	1	3.6%	2	6.9%	5	10.0%

As illustrated in Fig. 2, fentanyl was the most common substance detected with a 94% increase during the observed three-year period (F (2104) = 4.959, p = 0.009), Tukey's HSD post-hoc test (p = 0.008, 95% C.I. = −0.61, −0.08), including a 70% increase of fatal overdoses in 2020 alone. Cocaine presence had a statistically significant decrease from 28% in 2018 to only 8% in 2020 (F (2104) = 3.097, p = 0.049), Tukey's HSD post-hoc test (p = 0.039, 95% C.I. = 0.01, 0.40) while methamphetamine detection increased from 18% in 2018 to about 30% in 2019 and 2020. Benzodiazepines decreased from 36% in 2018 to 20% in 2020.

Fig. 2.

Trends in rural overdose toxicology

Note: p-values represent statistically significant changes in detected drugs

*p<0.05, **p<0.01.

Among the deaths examined where cocaine was detected, 69% also contained fentanyl, and in deaths where methamphetamine was detected, 77% also contained fentanyl. Fig. 3 illustrates the polydrug detections among these substances across each year that data were analyzed. As illustrated here, fentanyl has been the dominant substance detected in overdose deaths since 2019 in rural Michigan – initially in combination with methamphetamine and more recently alone, as fentanyl was detected without any of these substances in 28% of the 2020 cases.

Fig. 3.

Polydrug Trends in 11 Rural Michigan Counties by Year, 2018 through 2020 (N = 107).

4. Discussion

In this paper we describe toxicology trends of overdose deaths that occurred between January 1, 2018, and December 31, 2020, in 11 rural counties. We found that fentanyl was the dominant substance detected in overdose deaths during each year of the study period. The current findings suggest that fentanyl and methamphetamine co-use is an emergent trend although the degree to which methamphetamine is contaminated with fentanyl in the rural US remains unclear (Baker et al., 2021; Dai et al., 2021; Daniulaityte et al., 2022; Lockwood et al., 2021; Papamihali et al., 2021; Park et al., 2021). These findings are not intended to generalize across all rural contexts of the United States. However, the counties in the current study share federal distinction as rural counties with those across the country (i.e., have similar population sparsity and distance from large urban areas). Drug patterns, however, may vary by region of the country, and thus some of our information might generalize best with rural areas of states on the eastern half of the US where access to more potent opioids occurred earlier than in the western US. Our study findings may provide a few directions for future research. These data underscore the importance of rural regional research, where trends may diverge or align with general patterns. As a result, region-specific investigations (Lister et al., 2019) have the potential to provide important public health information, in line with Domain 1 of the CDC's Four Domains of Chronic Disease Prevention (Bauer et al., 2014) regarding epidemiological and surveillance data for overdose-related mortality.

A recent (i.e., January 2018 to March 2020) cross-sectional study of survey data of people who use drugs in rural areas across 10 states (i.e., Illinois, Kentucky, New Hampshire, Massachusetts, North Carolina, Ohio, Oregon, Vermont, West Virginia, and Wisconsin) found that the prevalence of non-fatal overdose was highest among those who self-reported using both methamphetamine compared to using either of the drugs independently (Korthuis et al., 2022). Our study extends the findings by Korthius et al. (2022) by illustrating that a combination of methamphetamine and fentanyl is increasing among fatal overdoses in rural areas. Interestingly, our findings suggest that benzodiazepine detections declined over the study period. This is counter to recent research that has identified increases in the detection of benzodiazepines in urban settings across North America (Pardo, 2022). Moreover, a recent study out of Toronto, Canada that utilized data from a similar timeframe as our study (i.e., 2019–2020) and found that estimates of co-occurring detection of fentanyl and methamphetamine was approximately 15.8% (Scarfone et al., 2022), which is considerably lower than what was found in the current analysis (i.e., 77%).

Based on prior research, our findings suggest two potential patterns of use – one pattern in which individuals may be co-using methamphetamines and fentanyl, and another patter where individuals may be fentanyl naïve, meaning that they were exposed to fentanyl via methamphetamine that was contaminated with fentanyl (Baker et al., 2021; Bolinski et al., 2022; Corser et al., 2022; Dai et al., 2021). Despite the potential for polydrug combinations that involve an opioid and methamphetamine, addressing the opioid aspect should be a priority for many organizations and first responders involved in efforts to curb overdose deaths. Moreover, responding to methamphetamine overdose can be challenging, as the drug can cause a wide range of physical and psychological symptoms and there are few response options (e.g., there is not currently a naloxone equivalent for methamphetamine overdose).

Naloxone distribution and other harm reduction supports (e.g., overdose prevention sites) would likely be effective interventions for those who co-use methamphetamine and those who may be fentanyl naive given that fentanyl was detected in most deaths. For those who may be fentanyl naïve, tools such as fentanyl test strips (FTS) may be helpful in detecting fentanyl and regulating use and there is evidence that rural residents are willing utilize FTS (Peiper et al., 2018). It should be noted that an awareness of the presence of fentanyl is a limited but potentially useful overdose prevention tool, as this technology can assists people in applying appropriate harm reduction strategies (e.g., titrating dosage) to reduce the risk of an overdose amid an increasingly volatile drug market supply (Jacka et al., 2020; Peiper et al., 2018; Zibbell et al., 2021). However, the awareness of fentanyl may not deter drug use as some individuals purposefully use fentanyl for a number of reasons (e.g., preference for fentanyl high or increased opioid tolerance; Foglia et al., 2021; McKnight and Des Jarlais, 2018).

In Michigan, FTS are legally accessible via syringe service programs (SSP) and the state's health department. At the time of writing, only three of the studied rural counties have implemented low-threshold (i.e., free of charge and publicly available) naloxone vending machines. Slightly less than half (i.e., Alpena, Clare, Newaygo, Lake, & Wexford) of the counties included in our analysis have SSPs and three counties have recently (i.e., since 2020) implemented naloxone vending machines (i.e., Crawford, Lake, Wexford. Our findings suggest that there may be a need to expand these services in each of the studied counties and there is a need for more research on the implementation and sustainability of harm reduction services in rural areas. Further, low-cost outreach services (e.g., naloxone distribution) are promising given the concerns about outpatient treatment and confidentiality among people who use drugs in rural areas (Hammarlund et al., 2018; Wieczorek, 2017)

Qualitative research has illustrated the critical importance of harm reduction services for individuals living in rural areas – insofar that access to naloxone and FTS were reflected upon as being sources of hope and community resilience (Bolinski et al., 2022). However, rural residents have stressed that access to these services can improve as structural and attitudinal barriers to harm reduction persist among providers in the rural US (Parry et al., 2021; Victor et al., 2022). Overdose prevention sites (OPS) are another evidence-based intervention that have not been formally implemented in the rural US. To date, there are only two recognized OPSs pilot programs in the US (Harocopos et al., 2022), whereas Canada opened their first OPS in 2003 (Kerr et al., 2017), and as of 2019 (Government of Canada, 2020), 39 are operating across the country while hundreds more are in operation globally (Levengood et al., 2021).

Outreach in rural Michigan counties may also consider services that are tailored to stimulant drugs, such as symptoms consistent with “overamping”, as these drugs are increasingly detected in postmortem toxicology records (Harding et al., 2022; Mansoor et al., 2022). Other promising options for improving access to MOUD include expanding telehealth services at rural spoke clinics, subsidizing transportation costs for rural patients facing burdensome travel to opioid treatment programs, and policy changes that enact COVID-19 flexibilities (e.g., extended take-home doses) long-term or create opportunities for methadone dispensing at community pharmacies or primary care providers (Lister and Lister, 2021).

4.1. Limitations

Our results do not represent all rural counties in Michigan as this study only looked at fatal overdoses that occurred within areas that were associated with the Mid-Michigan Medical Examiners group. Although administrative data may be subject to internal measurement error although they do offer a critical access point for monitoring drug use behavior and drugs that may be attributed to overdose deaths. The data presented in this paper are the most recent data available in the communities studied and it may be impactful in formulating public health response in rural areas, the timeliness of “real time” data about fatal and non-fatal drug overdose is a persistent challenge. That is, the drugs detected in toxicology records that are reported here may not accurately characterize current overdose deaths given the rapid evolution of the overdose crisis. Another limitation is that we could not account for every fentanyl analog which restrains the generalizability of the current findings with extent research from neighboring states (e.g., Ohio). This limitation is particularly crucial because in the neighboring states of Ohio and Indiana, there has been a wide range of fentanyl analogs identified over time, and some of these analogs have had a significant impact on overdose mortality rates (Daniulaityte et al., 2020 ;Daniulaityte et al., 2019; Huynh et al., 2020;Rosenblum et al., 2020). Future research should consider recruiting primary data among individuals in current use who live in rural communities to understand behavioral and systemic patterns which may increase the risk of fatal and non-fatal overdose exposure.

5. Conclusion

This paper presents the evolving public health threats to accidental overdose deaths in rural Michigan communities. We describe an ongoing overdose crisis at the county-level in which synthetic opioids contribute to the greatest level of overdose mortality, but we also show the emergence of methamphetamine and cocaine in rural toxicology records. This calls for specific responses, such as naloxone provision and MOUD to attend to the primary driver of overdose deaths – synthetic opioids. The use of other drugs that contribute to comparatively less mortality, such as methamphetamine, may require other customized interventions and the continued pursuit of an effective replacement medication for individuals who use methamphetamine is warranted (Trivedi et al., 2021). Moreover, outreach efforts and interventions should also consider the increased risk of fentanyl-related mortality among people who inject drugs. Drug checking technologies are one low-barrier tool that could aid in detecting the presence of fentanyl and rural policymakers should consider expanding these services via peer-driven outreach efforts.

Declaration of Competing Interest

None.