Evaluating demographic variation in no-cost naloxone distribution in North Carolina: a cross-sectional survey study

Abstract

Background

The objective of this study was to identify naloxone distribution resources, as well as to evaluate potential differences in naloxone distribution by demographic population utilizing a naloxone distribution score.

Methods

We utilized a cross-sectional online survey of potential no-cost naloxone distributors in North Carolina (NC) to identify their program type, where programs distribute no-cost naloxone, and to whom no-cost naloxone distribution was distributed in 2023. This information was utilized to create a naloxone distribution measure at the ZIP code level.

Results

A total of 241 survey responses for programs were identified as distributors of no-cost naloxone to community members in NC. At least one no-cost naloxone distributor served all demographic populations in over 92% of NC ZIP codes. The Non-Hispanic (NH) White group had the highest mean naloxone distribution score of 30.2 (SD = 8.9) across all ZIP codes, suggesting greater distribution to this demographic. After adjusting for the size of the race/ethnicity population in GEE models, all minority race/ethnicity groups had significantly lower naloxone distribution than the NH White group.

Conclusions

The naloxone distribution score calculated suggests NH White populations may be served by no-cost naloxone distributors more frequently when compared to other populations.

Introduction

Despite a recent downward trend in the overall rate of drug overdose deaths in the United States (US), differences in drug overdose death rates have widened across specific demographic groups, particularly among Non-Hispanic (NH) Black and Hispanic populations compared to NH White populations [1]. American Indian and Alaska Native (AI/AN) populations also continue to experience the highest rate of drug overdose deaths (31% higher than White NH populations), while Asian populations have the lowest rate [1]. In North Carolina, differences in drug overdose death rates were exacerbated during the COVID-19 pandemic [2], with persistently high death rates among AI/AN and Hispanic populations despite overall decreases in drug overdose deaths in the state [3].

Prior to 2017, NH White men had the highest opioid overdose death rate, as a result, treatment and opioid overdose prevention efforts were targeted toward White people who use drugs (PWUD) [4]. During this same time period, Black and Hispanic people were disproportionately criminalized for non-prescribed drug use and continue to be incarcerated for drug-related charges at higher rates than their White counterparts [5, 6]. For example, data from the National Survey on Drug Use and Health found prevalence of prior-year drug use was significantly higher among NH White populations than among Black populations; however, Black populations had nearly double the rate of drug-related arrests [7, 8]. This disproportionate response to drug use may have contributed to studies indicating a reluctance among PWUD to call emergency medical services, particularly in rural and underserved areas [9–11]. Given the sociocultural underpinnings of and fear associated with seeking emergency services, a community-driven approach may be necessary to reduce overdose deaths.

In response to elevated opioid overdose death rates, major efforts have been enacted to mitigate opioid overdose fatalities, including the allocation of public funds for the distribution of naloxone at no cost [12]. Naloxone is an opioid antagonist that reverses an opioid overdose and restores breathing; it can be safely administered by community members [13, 14]. Community-based naloxone distribution is an effective tool that reduces opioid overdose deaths by enabling bystanders to intervene in the event of an overdose without waiting for emergency services to arrive [15, 16]. Community-based naloxone distribution to PWUD and their social networks reduces opioid overdose fatalities and is extraordinarily cost-effective, with each dollar spent resulting in $2,742.00 in cost savings through fatality reduction [17, 18]. Syringe service programs (SSPs), health departments, and emergency medical services are among a few of the community-based programs that distribute no-cost naloxone at the community-level [19, 20].

The distribution of no-cost naloxone varies across areas and populations. An analysis of Rhode Island and Massachusetts community-based naloxone distribution found that urban municipalities and those with a higher percentage of Black residents, had higher naloxone distribution rates [212]. In multivariate analyses, however, naloxone distribution rates for racial/ethnic minorities were lower than the rate for NH White residents in Massachusetts [21]. An interrupted time-series evaluation also found that with the onset of the COVID-19 pandemic, there was a surge in opioid overdose deaths among Black residents, with no compensatory increase in naloxone distribution to this population [22]. Survey data of a northern Midwest Tribe in the US found that among indigenous participants that actively use drugs, only 40% had recently received naloxone [23]. Understanding patterns of no-cost naloxone distribution methods to different demographic groups can help contextualize the differences in opioid overdose and identify opportunities to improve distribution.

Efforts to measure naloxone distribution have proven challenging due to the varied nature of distribution channels, which make it difficult to track distribution to the end user. The objective of this study was to identify no-cost naloxone distribution sources and evaluate differences in naloxone distribution across population groups. North Carolina is an ideal state for this evaluation given its racial/ethnic diversity, including a large AI/AN population [24], and its substantial opioid overdose burden [25].

Methods

Study design

A cross-sectional survey was used to assess no-cost naloxone distribution in North Carolina (NC). The study protocol was reviewed by the Institutional Review Board at the University of North Carolina-Chapel Hill and determined to be non-human subjects research (IRB: 22–2664).

Study sample

As there is no comprehensive list of programs that distribute naloxone in NC, a community advisory board was convened to identify all potential no-cost naloxone programs in the state [26]. Our community advisory board consisted of twelve individuals who were purposively selected to support this project with geographic, racial/ethnic, and programmatic/role diversity throughout NC. This advisory board, in collaboration with the study team, identified lists of programs and contact information, which is described elsewhere [26]. In total, 548 contacts from 415 identified potential no-cost naloxone-programs were identified by community experts on the study team and advisory board.

Recruitment

Survey distribution occurred from June 2023 to January 2024. Identified contacts were emailed a link to an online Qualtrics survey. Each email was personalized to include the contact’s name and program with which the contact person was affiliated. If individuals were the contact person for more than one program, they were instructed to complete separate surveys for each program. Up to four follow-up emails were sent to non-responders. If there was no response to emails, non-responders were then called up to three times using a standardized script to request survey completion. Those who did not respond to phone contact were then mailed a flyer with a QR code linked to the Qualtrics survey. Survey respondents with valid contact information were provided with a $20 Amazon gift card.

Survey response reconciliation

Reconciliation of survey responses is described in detail elsewhere [26]. In some cases, multiple contacts were identified for the same potential naloxone-distributing program; all identified contacts were surveyed to maximize the program-level response rate. For programs with multiple survey responses, we retained the most complete response using a standardized process. Some respondents filled out the survey on behalf of multiple naloxone distribution entities (i.e. a health department and a post-overdose response team), which was determined by their survey response and through triangulation with community experts on the study team and advisory board [26].

Measures

The survey consisted of 12 questions, including program type (via a select all that apply question; i.e., treatment program, syringe service program (SSP), health department, Emergency Medical Services (EMS)); the physical address where the program was located; if the program was fixed site/mobile/both; the years in which the program distributed no-cost naloxone to community members (2017–2023); and all ZIP codes in which the program distributed naloxone (or towns if ZIP codes were unknown). Respondents also reported the racial and ethnic groups to whom they distributed no-cost naloxone in the past 6 months, including NH White, NH Black, NH American Indian/Alaska Native (AI/AN), NH Asian, NH Native Hawaiian/Pacific Islander, and Hispanic populations. With input from the community advisory board, these racial and ethnic groups were selected to match categories used by the U.S. Census Bureau. For each racial and ethnic group, respondents indicated the frequency with which they dispensed no-cost naloxone in the past six months: “never” (1), “rarely” (2), “sometimes” (3), or “often” (4).

Estimation of racial and ethnic populations at the ZIP code level

Our analysis required having racial and ethnic population breakdowns in the same geographic units as the survey responses (ZIP code). Because ZIP code level data are not readily available, we used a technique called population-weighted areal interpolation [27, 28] which uses higher resolution data to estimate values between non-nesting geographic units. For data inputs, we gathered race and ethnicity data (using the same groups as the survey) from the American Community Survey’s (ACS) 5-year estimates (2017–2021) at the block group level and census block-level population data from the 2020 U.S. Census. First, the 2020 block centroids (with population counts) were overlain with the ZIP code polygon layer. The resulting output was aggregated by U.S. Census block group and ZIP code tuple pairs in which each tuple contained the corresponding 2020 population. Each tuple population was divided by the total population of the corresponding block group, resulting in a proportion of each block group’s population in each tuple (e.g., if 400 residents of block group A were located in ZIP code X and 600 residents of block group A were located in ZIP code Y, then the result would be block group A—ZIP code X = 0.4 and block group A—ZIP code Y = 0.6). The block group population count data from the ACS (for each racial and ethnic breakdown) were then multiplied by the weights, and then the counts were aggregated by ZIP code to calculate the ZIP code-level count. Race and ethnic group-specific population proportion values were also calculated for each ZIP code by dividing the group-specific population count of the ZIP code by the total population count of the ZIP code.

Data analysis

All data were analyzed using SAS version 9.4. First, data at the entity-level were analyzed descriptively. Means and standard deviations of the frequency of distributing no-cost naloxone to each race/ethnicity group were calculated in total, and by program type. Because of the correlated structure of our data (i.e. individual contacts provided responses on behalf of multiple programs), linear mixed models with random effects for respondent were used to test whether there were significant differences in (1) distribution frequency ratings by race/ethnicity group within each program type, and (2) distribution frequency ratings for each race/ethnicity group by program type.

Second, the data were analyzed at the ZIP code level. The naloxone distribution (ND) score at the ZIP code level was calculated in two steps. First a total distribution measure was calculated for each race/ethnicity group. Second, the measure was converted to proportion to construct the final distribution measure.

Equation 1: First Step of Calculating the Naloxone Distribution score: Calculating the Naloxone total distribution measure.

where NT_r,z is the total distribution measure for the r-th race/ethnicity group in ZIP code z. E is total number of entities reporting distribution of Naloxone to ZIP code z. NF_e,r is the value (1 = never, 2 = rarely, 3 = sometimes, or 4 = often) provided by the entity e in relation to race/ethnicity r to the question “In the past 6 months, please estimate how often your organization distributed naloxone to the following populations?”.

Equation 2: Calculating the Naloxone Distribution Score

where ND_r,z is Naloxone distribution score for the r-the race/ethnicity group in ZIP code z. G is the number of race/ethnicity groups in ZIP code z. NT_g,z is the total distribution measure for race/ethnicity r in ZIP code z. r ∈ {NH White, NH Black, NH American Indian/Alaska Native, NH Asian, NH Native Hawaiian/Pacific Islander, Hispanic}. z ∈ {27,006, 27,007, …., 28,909}.

For ease of interpretation, the ND_r,z was further converted to percentages by multiplying it by 100. The naloxone distribution score is a relative measure, with a higher number indicating a greater frequency of no-cost naloxone distribution to the race/ethnicity of interest when compared to distribution to all other races/ ethnicities in that ZIP code. The naloxone distribution measure was calculated as a proxy measure of reported distribution to specific races/ethnicities at the ZIP code level.

For example, if in one ZIP code there is only one program that reports distributing no-cost naloxone “often” to NH Black populations, and NH White populations, and “never” to AI/AN, Asian, NH/PI, or Hispanic populations, the naloxone distribution score for NH White in this ZIP code would be calculated by dividing the ordinal value for NH White (4, for “often”) by the sum of the ordinal values of distribution for all other populations (4 for Black, 1 for each population reported as “never”). This calculation was performed only in ZIP codes in which each population resides according to ACS data.

Descriptive statistics, means, and standard deviations were used to summarize the naloxone distribution scores by race/ethnicity group. Further, for each race/ethnicity and ZIP code where race/ethnicity resided, the data were tabulated to determine the percentage of ZIP codes where no-cost naloxone was distributed.

Generalized estimating equations (GEE) linear regression model was used to assess the association between race/ethnicity and no-cost naloxone distribution. The outcome variable was race/ethnicity-specific naloxone distribution score in each ZIP code. Model was adjusted for the race/ethnicity ZIP code population. GEE was used to account for the correlation arising from multiple race/ethnicity-specific observations within the same ZIP code, as each ZIP code contributed separate naloxone distribution scores for multiple racial/ethnic groups.

Results

The study team contacted a total of 548 contacts from 415 potential no-cost naloxone-distributing programs. A total of 315 contacts responded and over three-quarters (76.5%; 241/315) of responding contacts reported that their program distributed no-cost naloxone to community members. The 241 contacts represented a total of 317 different programs that distributed naloxone at no-cost. Most programs reported distributing no-cost naloxone to NH White and NH Black or African American populations most frequently (Table 1). Programs reported distributing no-cost naloxone less frequently to NH AI/AN, Asian, and Native Hawaiian or Pacific Islander populations (Table 1).

Table 1.

Reported frequency of distribution of no-cost naloxone to races and ethnicities by program type (N = 317)

Program type	Populations to whom naloxone was distributed in 2023						P-valuea
	NH White mean (SD)	NH Black or African American mean (SD)	Hispanic or Latino Mean (SD)	NH American Indian or Alaska Native mean (SD)	NH Asian Mean (SD)	NH Native Hawaiian or Pacific Islander mean (SD)
Total:	3.58 (0.72)	3.17 (0.83)	2.67 (0.88)	2.14 (0.91)	1.87 (0.80)	1.74 (0.79)	< 0.001
Health department (N = 81)	3.47 (0.70)	3.07 (0.84)	2.64 (0.91)	1.97 (0.88)	1.78 (0.78)	1.68 (0.75)	< 0.001
Treatment program/center (N = 59)	3.67 (0.69)	3.40 (0.73)	2.81 (0.88)	2.35 (0.92)	2.11 (0.82)	1.91 (0.91)	< 0.001
SSP (N = 41)	3.85 (0.43)	3.25 (0.74)	2.85 (0.83)	2.42 (0.93)	1.83 (0.78)	1.75 (0.81)	< 0.001
EMS/community PARAMEDICINE/ PORT (N = 33)	3.70 (0.59)	3.12 (0.89)	2.61 (0.83)	2.06 (0.90)	1.79 (0.78)	1.61 (0.75)	< 0.001
Health care provider (N = 20)	3.70 (0.66)	3.25 (0.79)	2.70 (0.98)	1.90 (0.72)	1.75 (0.64)	1.65 (0.75)	< 0.001
Justice-involved program (N = 17)	3.35 (1.0)	3.24 (1.09)	2.53 (1.01)	2.06 (1.03)	1.94 (0.97)	1.88 (0.93)	< 0.001
Other harm reduction organizations (N = 24)	3.83 (.75)	3.50 (0.88)	2.92 (0.88)	2.46 (0.88)	2.13 (0.80)	1.96 (0.75)	< 0.001
Community health center (N = 12)	3.92 (0.29)	3.25 (0.62)	3.17 (0.72)	2.00 (0.85)	1.92 (0.90)	1.75 (0.97)	< 0.001
Behavioral health (N = 10)	3.90 (0.32)	3.60 (0.52)	2.8 (1.03)	2.20 (1.14)	1.80 (0.63)	1.6 (0.52)	< 0.001
Government/ Community Coalition (N = 9)	4.00 (.00)	3.44 (.88)	3.00 (0.50)	2.33 (.87)	1.89 (.60)	1.89 (.60)	< 0.001
Other* (N = 5)	2.60 (1.52)	2.60 (1.14)	2.20 (1.10)	2.00 (1.23)	1.60 (0.55)	1.60 (0.55)	***
LME-MCO (n = 4)	3.50 (0.71)	3.00 (0.00)	2.00 (1.41)	1.50 (0.71)	1.50 (0.71)	1.50 (0.71)	***
Tribal health organization (n = 2)	3.50 (0.71)	3.00 (1.41)	2.50 (2.12)	3.50 (0.71)	1.00 (0.00)	1.50 (0.71)	***
P-valueb	0.667	0.988	0.858	0.937	0.650	0.915

For each racial and ethnic group, respondents indicated the frequency with which their program dispensed no-cost naloxone in the past six months (1 = never, 4 = often). Higher mean scores indicate greater frequency of distributing no-cost naloxone to that demographic group

*Other program types included church ministries, and college-based distribution

***Sample size too small to conduct robust analysis. P-values^a based on Type III test of fixed effects for race from linear mixed models. P-values^b based on Type III test of fixed effect for program type from linear mixed models

Across all demographic groups, over 92% of the 761 NC ZIP codes had at least one program that reported distributing no-cost naloxone (Table 2). Among ZIP codes where NH White people reside, 96.7% have at least one entity distributing naloxone to NH White individuals. Among ZIP codes where NH Black individuals reside, 96.3% have at least one entity distributing naloxone to NH Black populations. Only 92.6% of ZIP codes in which NH Asian populations live have access to a no-cost naloxone distribution entity serving NH Asian populations.

Table 2.

Frequency of ZIP codes where no-cost naloxone is distributed (by at least one program) by race/ethnicity

Race/ethnicity	N of ZIP codes where race/ethnicity population > 0	N (%) of ZIP codes where no-cost Naloxone is distributed
NH White	761	736 (96.7)
Hispanic	756	726 (96.0)
NH Black	747	719 (96.3)
NH Native Hawaiian/Pacific Islander	485	461 (95.1)
NH American Indian/Alaska Native	718	670 (93.3)
NH Asian	705	653 (92.6)

Total number of NC ZIP codes = 761. Percentages in column 3 were calculated as the number of ZIP codes with at least one program reporting no-cost naloxone distribution divided by the number of ZIP codes with a population greater than zero for the specified racial/ethnic group

Naloxone distribution score

The NH White group had the highest mean naloxone distribution score of 30.2 (SD = 8.9) across all ZIP codes, suggesting that no-cost naloxone was distributed most frequently to this demographic group when compared to other demographic groups in that area (Table 3). NH Native Hawaiian/Pacific Islander had the lowest mean naloxone distribution score of 8.5 (SD = 3.9), indicating less distribution to this demographic group. Median naloxone distribution scores followed the same pattern as the mean naloxone distribution scores (Table 3).

Table 3.

Summary statistics for no-cost naloxone distribution score by race/ethnicity

Race/ethnicity	N of ZIP codes	Mean naloxone distribution score(SD)	Median naloxone distribution score (IQR)
NH White	736	30.2 (8.9)	27.9 (25.0–32.5)
NH Black	722	24.9 (4.9)	24.1 (21.7–27.3)
Hispanic	731	18.8 (5.6)	18.4 (15.5–21.6)
NH American Indian/Alaska Native	696	13.6 (6.3)	13.5 (9.4–17.1)
NH Asian	686	8.9 (3.7)	9.3 (6.7–11.5)
NH Native Hawaiian/Pacific Islander	480	8.5 (3.9)	8.3 (5.9–11.3)

Naloxone Distribution Score: A composite measure of no-cost naloxone availability across ZIP codes for specific race/ethnicity group, calculated for ZIP codes where at least one program reported distribution of naloxone and where population of given race/ethnicity is greater than zero; higher values indicate greater distribution coverage

After adjusting for the size of the race/ethnicity population in GEE models, all minority race/ethnicity groups had significantly lower naloxone distribution than the NH White group. The largest differences in no-cost naloxone distribution were between NH Native Hawaiian/Pacific Islander and NH White groups (β = − 17.6, 95% CI: − 19.1, − 16.0), and NH Asian and NH White groups (β = − 16.5, 95% CI: − 18.1, − 15.0) (Table 4). NH Black and Hispanic groups also experienced significantly lower distribution, though the magnitude is smaller for NH Black.

Table 4.

Generalized estimating equations assessing the association between race/ethnicity and no-cost naloxone distribution to that group, as reported by service organizations, N = 736

Parameter	β	95% CI	P-value
Intercept	25.0	(23.5, 26.5)	< .0001
Population	0.08	(0.06, 0.10)	< .0001
Race/ethnicity
NH White	Ref
NH Black	− 1.6	(− 2.9, − 0.3)	0.0164
Hispanic	− 7.0	(− 8.4, − 5.5)	< .0001
NH American Indian/Alaska Native	− 11.7	(− 13.4, − 10.1)	< .0001
NH Asian	− 16.5	(− 18.1, − 15.0)	< .0001
NH Native Hawaiian/Pacific Islander	− 17.6	(− 19.1, − 16.0)	< .0001

Discussion

In this study we identified differences in no-cost naloxone distribution between demographic groups in NC. This study found that while over 92% of ZIP codes with AI/AN, Asian, Black, Hispanic, or Native Hawaiian/ Pacific Islander populations had at least one program distributing no-cost naloxone, these communities appear to have been distributed no-cost naloxone less frequently throughout NC when compared to NH White populations. Even after adjusting for population size in rural areas, Hispanic, AI/AN, Asian, and Native Hawaiian/ Pacific Islander NH/ PI populations had lower naloxone distribution scores than their NH White counterparts, indicating less frequent distribution of no-cost naloxone to these groups. These findings suggest reduced distribution of no-cost naloxone for these populations, highlighting a need for tailored and targeted interventions to improve access to naloxone in these communities.

In NC, these findings are especially urgent. The state has documented large racial disparities in overdose mortality during the ongoing fentanyl crisis. Specifically, between 2019 and 2021, overdose deaths increased by 139% among Black North Carolinians and 117% among AI/AN residents, outpacing an increase of 53% in the NH White populations during the same time period [25]. Our findings show that these same populations may be systematically underserved by naloxone distribution efforts, which can contribute to opioid overdose deaths.

Although our survey found that no-cost naloxone was available through health departments, syringe service programs (SSPs), and from emergency medical services, these resources were not equitably distributed across racial and ethnic groups or ZIP codes. AI/AN populations often reside in rural regions, which frequently lack access to opioid treatment programs [29], pharmacies that stock naloxone [30], and consistent harm reduction services [31, 32]. Even where naloxone-distributing programs are present, stigma, geographic isolation, underfunding, and lack of culturally appropriate services may inhibit actual access and use [33]. As community-based naloxone distribution programs have been shown to reduce both opioid overdose deaths and healthcare costs, continued and sustainable funding is needed to strengthen these efforts [17, 18]. Accordingly, naloxone distribution through community-based programs like SSPs should be expanded and better resourced with targeted strategies to ensure that programs can more effectively reach AI/AN, Asian, and Native Hawaiian/ Pacific Islander communities.

Similar to a previous study in Massachusetts, distribution to non-White populations was reported as lower than that to NH White populations in NC [21]. NC is home to over 130,000 AI/AN people, with eight recognized tribes [24, 34]; yet, AI/AN populations had significantly less distribution when compared to their NH White counterparts. Efforts to improve access to no-cost naloxone in AI/AN communities should be community-informed to maximize acceptance. For example, the Eastern Band of Cherokee Indians has installed emergency overdose response boxes in 40 locations across the Qualla Boundary, and ten vending machines that make naloxone available for tribal members in these areas [35]. Beyond having naloxone available, tailoring distribution to indigenous populations may include efforts to improve secondary distribution that leverages social connectedness to reach individuals who would not access overdose prevention resource services on their own [23, 36, 37].

This study contributes to a growing evidence base suggesting that current naloxone distribution may be suboptimal and leave populations without an available source of naloxone. Naloxone distribution should be prioritized for PWUD, using strategies tailored to communities with higher rates of opioid overdose, such as AI/AN populations. These approaches should be developed with meaningful community input and implemented in ways that ensure access and support lasting reductions in opioid overdoses among disproportionately affected groups.

Limitations

Our study has several limitations. First, some respondents were unreachable or did not respond to our survey, so the dataset does not encompass all no-cost naloxone availability points in NC. Our score of naloxone distribution reflects self-reported data and may not accurately reflect the true community-level distribution of naloxone. Respondents were instructed to complete the survey separately for each program with which they were affiliated; however, some respondents completed a single survey on behalf of multiple programs. As the survey did not include fields to indicate the specific program to which each response applied, we assumed that the reported responses reflected all programs with which the respondent identified in the “select all that apply” program type question. As in any survey study, our data may suffer from recall bias, especially in relation to questions regarding the distribution of naloxone in previous years, geographical locations where naloxone was distributed, as well as to whom programs reported distributing no-cost naloxone. Specifically, the frequency of naloxone distribution to racial and ethnic groups may be subject to recall bias and misclassification of race/ethnicity since many programs do not systematically collect self-reported demographic information from naloxone recipients. Our study did not attempt to measure the number of naloxone kits distributed in the community; however, this study was the first attempt to identify all sources of no-cost naloxone, and report where and to whom the distribution occurred. Further, we did not have ZIP code level race/ethnicity data on populations at risk of overdose to accurately adjust for “need” in our model. Finally, our data were collected in North Carolina; findings may not generalize to other locations.

If given the opportunity to conduct this survey again, we would change several aspects of our study design. Recommendations for survey replication are reported in a separate paper describing the specific involvement of the community advisory board [26]; however, other recommendations include: (1) using a closed-ended question for reporting program type and (2) collecting estimated number of naloxone kits distributed to each race/ethnicity group. Since some respondents selected more than one program type in the survey, analysis was based on the number of responses, not the number of selected program types to prevent over-estimation of naloxone distribution Originally, we chose not to ask respondents about the quantity of naloxone kits their program distributed to mitigate survey abandonment. Future survey efforts should ask respondents to report the quantity of naloxone distributed by their program, by race/ethnicity, to better estimate overall naloxone availability and differences in distribution across demographic populations.

Conclusion

Our findings suggest that no-cost naloxone distribution could be optimized to reach populations with higher overdose risk. These results highlight the urgent need to align naloxone distribution with racial and geographic patterns of overdose risk, ensuring that communities historically excluded from harm reduction efforts receive the resources necessary to prevent avoidable deaths. Naloxone distribution that meets the needs of all populations is not just a public health imperative; it is a moral and structural mandate for opioid overdose prevention.

Author contributions

Grace Marley: Data Curation; Investigation; Writing- Original Draft preparation—Caroline Shubel: Data Curation; Investigation; Writing—Review & Editing—Izabela Annis: Formal Analysis; Writing- Review & Editing—Bhoomi Parikh: Writing—Review & Editing—Paul Delamater: Writing—Review & Editing—Carolyn Thorpe: Writing—Review & Editing—Bayla Ostrach: Conceptualization, Data Curation; Supervision; Writing—Review & Editing—Delesha Carpenter: Conceptualization, Data Curation; Funding acquisition; Writing—Review & Editing.

Funding

Research reported in this publication was supported by the National Institute on Minority Health and Health Disparities of the National Institutes of Health under Award Number R01MD018085. The entire project was funded with Federal money. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed by the Institutional Review Board at the University of North Carolina-Chapel Hill and determined to be non-human subjects research (IRB: 22–2664).

Competing interests

The authors declare no competing interests.