Differences in Perceptions of Community Stigma Towards Opioid Use Disorder Between Community Substance Use Coalition Members and the General Public

Abstract

Introduction:

To examine differences in perceptions about community stigma towards individuals with opioid use disorder (OUD) between community members involved in the opioid response (i.e., coalition members) and the general public, and how community geography may moderate this relationship.

Methods:

This study administered identical cross-sectional surveys about perceived community opioid-related stigma to two distinct populations in 66 communities participating in the HEALing Communities Study prior to the intervention period (i.e., coalition members, November 2019–January 2020; residents, March–April 2020). Linear-mixed models compared survey responses of populations, including the moderating effect of community rural/ urban location.

Results:

A total of 826 coalition members and 1,131 residents completed the surveys. The study found no differences between the coalition members and residents for general perceived community opioid-related stigma. In both urban and rural communities, coalition members reported greater perceived community stigma than residents reported towards medication for opioid use disorder (MOUD), naloxone, and drug treatment as an alternative to incarceration.

Conclusion:

Our findings suggest similar perceived community opioid-related stigma between coalition members and residents, yet differences emerge related to evidence-based practices (i.e., MOUD, naloxone, and drug treatment as an alternative to incarceration) to reduce opioid overdose deaths.

Trial Registration:

ClinicalTrials.gov Identifier: NCT04111939

1. INTRODUCTION

The opioid epidemic is a pernicious public health problem, contributing to 75,673 overdose deaths in the U.S. in 2021 (Centers for Disease, 2021). As part of the response, federal, state, and local agencies have invested significant resources to increase the accessibility of evidence-based practices (EBPs) for opioid prevention and treatment such as increasing access to naloxone and creating coordinated systems of care to increase the number of people receiving treatment (Ghertner, 2019; Guy, 2019). However, widespread stigma towards people with OUD is a central barrier to improving the accessibility and uptake of prevention treatment and services (Luoma, 2011). Additionally, those with OUD may fall into more than one stigmatized group such as race and face multiple challenges (Bowleg, 2021; Friedman et al., 2022). These unmet needs aggravated by stigma for opioid use disorder (OUD) highlight the need to understand how perceptions of community stigma may impact decisions around prevention and treatment services (Freeman et al., 2018).

Stigma is a complex phenomenon occurring at multiple levels, including individual, community, and structural stigma (Luoma et al., 2010a; Tsai et al., 2019). Within and across these levels are various manifestations of stigma, including experienced, perceived, witnessed, anticipated and internalized stigma (Nyblade et al., 2021). Individual stigma involves a person’s negative perceptions towards a specific group, and when the public endorses these negative views, it is then known as community stigma (Beachler et al., 2021). These community held views can then transform into structural stigma when they are codified into laws or policies that systemically diminish opportunities (Beachler et al., 2021; Corrigan et al., 2005; Tsai et al., 2019). Additionally, community stigma can become anticipated or internalized in the targeted group reinforcing their marginalization and over time negatively impact help seeking behaviors (Vogel et al., 2013). Community stigma can be directed towards medication/treatment or health tools used by stigmatized groups (i.e., intervention stigma) and manifest as anticipated or experienced stigma that inhibits people with OUD from accessing evidence-based treatments or medications for opioid use disorder (MOUD) (Madden, 2019; Madden et al., 2021). Alternatively, evidence suggests that perceptions of stigma may differ based on treatment status, with stigma towards untreated OUD greater than treated OUD (McGinty et al., 2015).

In this study, we focus on perceptions of community stigma, as this level of stigma is particularly relevant to understand the community context that impacts acceptance of, and decisions about, the local opioid response (Doll et al., 2012). Perceptions of community stigma surrounding OUD, refers to the awareness of stigmatizing beliefs and attitudes that exist in the community such as not believing OUD is a medical condition, seeing OUD as a personal choice, and higher perceived levels of dangerousness of those with OUD (Tankard & Paluck, 2016). Prior research shows perceived community stigma to be associated with decreased healthcare utilization, suboptimal care, and poorer physical and mental health outcomes (Stringer & Baker, 2018; van Boekel et al., 2013). Perceived community stigma can also contribute to the acceptance of discriminatory practices, such as stronger public support for denying housing and employment for those with OUD (Kennedy-Hendricks et al., 2017a; Ledford et al., 2021).

Stigma towards people with OUD may (or may not) differ between the residents and community members who are involved in the local opioid response (‘local experts’; e.g., healthcare and behavioral health service providers, criminal justice representatives, individuals with lived experience) (van Boekel et al., 2015). Research into familiarity suggests that community stigma may exist on a U-shaped curve, where those with relatively less familiarity associated with individuals with a given condition (e.g., no professional experience, co-workers) as well as those with more familiarity (e.g., service providers, nuclear family) exhibit greater community stigma (Corrigan & Nieweglowski, 2019). Conversely, individuals with moderate familiarity (e.g., extended family members) demonstrate the lowest levels of community stigma. For instance, some prior research finds that healthcare providers’ perceptions of people who use opioids were less stigmatizing than that of the public, while others found similar levels of stigmatizing attitudes between the two populations (Aronowitz & Meisel, 2022; Mackey et al., 2020; Stone et al., 2021).

Prior research suggests differences in opioid-related stigma, or substance use disorders more broadly, between those who live in rural and urban communities (Miller et al., 2022). For instance, qualitative studies in rural areas describe considerable stigma towards persons who use drugs (Ezell et al., 2021). Additionally, rural attitudes expressing opposition to OUD EBPs such as OUD treatment and MOUD (Browne et al., 2016; Richard et al., 2020), and the use of naloxone (Ezell et al., 2021), reflect that intervention stigma is still prevalent. Survey findings comparing stigma between rural and urban residents show mixed results. For instance, one study found greater stigma towards substance use disorders than mental disorders among rural residents (Corrigan et al., 2009). Yet, one survey of a US national sample found that rural respondents reported significantly lower stigma specific to prescription drug misuse than urban respondents (Miller et al., 2022) while another found no difference in perceived stigma between rural and urban respondents (Dschaak & Juntunen, 2018).

Research has not yet thoroughly investigated the levels of perceived community stigma among local experts directly involved in opioid overdose response and how these align or diverge from residents in the same communities. We explore differences in perceptions of community stigma towards people with OUD and intervention stigma against EBPs between two groups: experts involved in the local opioid response characterized as having high familiarity and engagement in the community’s response; and the residents in those same communities who have varying levels of familiarity and less engagement in the community’s response. We also examine how these differences may be moderated by whether a community is classified as rural or urban. Examining differences in the perceptions of community stigma among experts overseeing the response to the opioid epidemic compared to residents’ perception of community stigma could reveal insight into experts’ misconceptions about the community’s level of support for specific evidence-based interventions.

2. METHODS

2.1. Study Design and Setting

The HEALing (Helping to End Addiction Long-term^SM) Communities Study (HCS) is a multisite, wait-listed, community-level cluster randomized trial that seeks to test the effectiveness of a community-level intervention on reducing opioid overdose deaths in 67 communities heavily impacted by the opioid epidemic across Kentucky, Massachusetts, New York and Ohio (Walsh et al., 2020). HCS tests the Communities That Heal (CTH) intervention which deploys a community-engaged, data-driven process and tailored communication campaigns to assist community coalitions in adopting and implementing EBPs to reduce opioid fatalities (Walsh et al., 2020). The study sample for this analysis includes 66 of 67 communities in the HCS (see Supplemental Material -Appendix Table 1 for community demographic characteristics). One community withdrew from the study after randomization and before the CTH implementation began. Therefore, we did not include it in the data collection or analysis reported here.

This present analysis compares pre-intervention data collected from two surveys: the Community Coalition Member/ Key Stakeholder Baseline Survey (CBS) and the Campaign Evaluation Questionnaire (CEQ). The Institutional Review Board of Advarra Inc., the HCS single IRB, approved all study procedures (Pro00038088).

2.2. Community Coalition Member/ Key Stakeholder Baseline Survey Sample Selection, Recruitment, and Data Collection

The CBS captured perspectives of experts involved in their community response to the opioid epidemic either as members of community coalitions or as key stakeholders (referred to as ‘coalition members’ for concision). CBS data collection occurred between November 2019 and January 2020, before any intervention activities had begun in each community. We identified participants in the CBS through membership or attendance rosters provided by leaders of existing substance and tobacco use community coalitions in 55 of the HCS communities. Coalition members typically worked for various organizations including substance use prevention organizations; substance use and mental health treatment centers; hospitals; community health centers; public health departments; jails; sheriff’s offices; courts; local government; local businesses; and religious organizations. Coalition members also could represent consumers (e.g., person in recovery, person who uses drug) or families (e.g., family member uses drugs or has experienced an overdose). Some coalition members volunteered to participate in the coalition while other were appointed based on their role in a member organization; however not all coalitions were open to all community members. For those communities that did not have existing coalitions (n=11), public health and substance use treatment contacts identified key stakeholders. The number of participants per community varied based on the size of the community and the size of the community coalition. In total, 3,592 coalition members and key stakeholders were eligible for the CBS, of which 3,203 had valid email addresses or were still participating in their coalition at the time of data collection.

In each community, a survey champion who was either a local or state government official in the community, or the chairperson of the coalition, introduced the survey to potential participants either via e-mail or during in-person coalition meetings. Subsequently, the study managed a four-week recruitment process in each state with a Research Electronic Data Capture (REDCap) website sending an email invitation with instructions regarding how to access and complete the CBS.

Participants completed surveys through one of three modes: through REDCap; on paper copies distributed at coalition meetings and returned by mail; or by a structured telephone interview. All surveys were in English. For individuals self-administering the survey via REDCap or on paper, they first signed an informed consent either digitally or on paper. Participants who completed the survey through a telephone interview first gave verbal informed consent which the interviewer documented in writing. Research staff trained by site leads conducted telephone interviews at the four sites. In three of the four states in which the communities were located, the study offered incentives to complete the survey unless the participant declined or was unable to accept the compensation. A total of 1,055 coalition members and key stakeholders responded to the CBS (32.9% of the 3,203 individuals invited), with 826 (78.3% of respondents) providing complete data used in this analysis.

2.3. Community Evaluation Questionnaire Sample Selection, Recruitment, and Data Collection

The CEQ captured perspectives of the general public (i.e., residents), and assessed specific components of stigma toward individuals with opioid use disorder, and the awareness and acceptability of naloxone and MOUD treatment in the same communities as the CBS. CEQ data collection occurred between March and April 2020. Communities in the 1^st Wave of the intervention (n=33) had begun receiving the CTH intervention in February 2020, but at the time of the CEQ data collection all intervention activities were focused on coalition development and no activities had begun that involved the public. Potential participants were recruited via a series of Facebook/Instagram advertisements that targeted zip codes corresponding to the 66 HCS communities. Community residents older than 18 who resided in an HCS community were eligible to participate in data collection. The study pre-established a completion goal for each of the four states in which the communities were located to not oversample a specific state, and stopped recruitment within a community at n=20 to avoid oversampling more populated (urban) areas.

The study directed individuals who authorized Facebook to collect information to check that they had a real Facebook account and to screen out bots to a brief screening instrument to validate their place of residence (i.e., provide their zip code), age (i.e., insert date-of-birth), and verify their email address. Participants entered all email addresses on a separate form so they would not be linked to the screener or CEQ responses. Eligible participants were then routed to the CEQ survey instrument hosted on REDCap where they signed a digital informed consent. Upon completion of the CEQ, each participant had the opportunity to enter a drawing to win a $100 Amazon electronic gift card. If they elected to do so, they provided their name and email address at the end of the survey so the study could contact them if they won the gift card. At the close of the CEQ, the study distributed an electronic gift card to one randomly selected winner in each community. A total of 1,353 community residents responded to the survey (14.1% of the 9,591 individuals who clicked on the recruitment advertisement), with 1,131 complete cases for analysis (83.6% of survey respondents).

2.4. Key Independent Variables

The study examined independent and interaction effects for two independent variables: (1) respondent role categorized as either a community coalition member/ key stakeholder (‘coalition member’) or a community resident (‘resident’) based on whether the respondent completed the CBS or CEQ survey; and (2) urban/ rural status - whether the respondent was from a rural or urban community. Due to practical considerations, the study defined communities in three states by county and utilized the National Center for Health Statistics criteria where “metropolitan” counties were considered urban (non-rural) and “nonmetropolitan” counties were considered rural (Ingram & Franco, 2014). Communities in one state were not defined by county lines, and the study assigned a ‘rural’ designation to those with a population density of less than 500 per square mile (Walsh et al., 2020).

2.5. Main Outcome

The outcomes of interest for this study were the 11 survey questions about perceived community opioid-related stigma answered by participants on both the CBS and CEQ (see Table 1). The study adapted eight of the questions from the perceived stigma scale developed by Louma et al. (Luoma et al., 2010b); and developed the remaining three questions (i.e., MOUD, Naloxone, Incarceration) for this study (Knudsen et al., 2020). The study piloted the survey with 15 individuals in non-HCS communities and refined it prior to deployment in HCS communities (see Knudsen et al., 2020). Each question has a scale of 1–7: strongly disagree (1), disagree, somewhat disagree, neither agree or disagree, somewhat agree, agree, strongly agree (7). The survey also gave respondents an option to not answer the question. The study reverse coded four items (i.e., Think Less, Pass Over Applicant, MOUD, Naloxone) such that greater agreement (i.e., higher scores) corresponds to more perceived community stigma; all other items are interpreted as greater agreement (i.e., higher scores) corresponding to lower perceived community stigma. The study evaluated each item individually rather than as a scale in order to identify specific rather than general differences in measured components of stigma.

Table 1:

Perceived community opioid-related stigma questions on both the community coalition member/ key stakeholder baseline survey and campaign evaluation questionnaire survey.¹

Survey Text	Question
	Label
Most people would willingly accept someone who has been treated for opioid use disorder as a close friend.	Close Friend
Most people in my community believe that someone who has been treated for opioid use disorder is just as trustworthy as the average citizen.	Trustworthy
Most people in my community would accept someone who has been treated for opioid use disorder as a teacher of young children in a public school.	Teacher
Most people in my community would hire someone who has been treated for opioid use disorder to take care of their children	Child Caretaker
Most people in my community think less of a person who has been in treatment for opioid use disorder.2	Think Less
Most employers in my community will hire someone who has been treated for opioid use disorder if he or she is qualified for the job.	Hire
Most employers in my community will pass over the application of someone who has been treated for opioid use disorder in favor of another applicant.2	Pass Over Applicant
Most people in my community would be willing to date someone who has been treated for opioid use disorder.	Date
Most people in my community believe that medications for opioid use disorder such as methadone and buprenorphine are just replacement drugs and not real treatment.2	MOUD
Most people in my community believe that if you provide naloxone to reverse an overdose to someone that it will encourage them to continue using opioids in the future.2	Naloxone
Most people in my community believe that it is better to offer drug treatment as an alternative to incarceration for people with opioid use disorder who are charged with non-violent crimes.	Incarceration

¹Each question is rated on a scale of 1–7: strongly disagree (1), disagree, somewhat disagree, neither agree or disagree, somewhat agree, agree, strongly agree (7). Respondents were also given an option to not answer the question.

²Reverse coded item, such that greater agreement (higher score) corresponds to more perceived community stigma.

2.6. Covariates

Both the CBS and CEQ collected demographic information from respondents. This information included race/ ethnicity (coded as non-Hispanic White, non-Hispanic Black, Hispanic, non-Hispanic or Other); Gender (i.e., male, female, different identity); Age (i.e., 18–34, 35–49, 50–64; 65–74, 75 and over); and Education (i.e., high-school or less, some college/ Associate degree; Bachelor degree; Graduate degree). The analysis included an indicator of state and community in models.

2.7. Statistical Analysis

The analysis only used survey responses with a valid response to at least one stigma question in the analysis. Linear-mixed models estimated differences between the coalition members and residents on the stigma items. The analysis estimated each survey item using a separate model and included interaction terms in modeling to test for differential effects of urban/rural status on stigma responses among coalition members versus residents. The study reports generalized least squares mean differences among coalition members versus residents, and within urban/ rural strata. With several models being run (n=11 for coalition member and resident comparisons, n=22 for coalition members and resident responses by rural and urban status), we controlled for the false discovery rate by applying Benjamini-Hochberg adjustments to the p-values (Benjamini & Hochberg, 1995). The false discovery rate controls the expected proportion of incorrectly rejected null hypotheses (i.e. Type 1 errors) in a list of rejected hypotheses, thereby allowing for conservative identification of hypotheses that are statistically significant after adjustment. Listwise deletion handled missing data (see Supplemental Material -Appendix Table 2). The study conducted analysis using SAS v9.4 (Cary, NC).

3. RESULTS

Respondent characteristics for the CBS and the CEQ are in Table 2. Among the 826 respondents to the CBS survey with complete data, 468 (56%) were from urban communities and 358 (44%) were from rural communities. The majority of coalition members were ages 35–49 (36%) or 50–64 (40%); identified as non-Hispanic White (92%); were female (67%); and had a graduate degree (55%).

Table 2.

Demographic characteristics of respondents to the community coalition/ key stakeholder baseline survey and campaign evaluation questionnaire, by urban/ rural status.

Characteristic, n(%)	Coalition Members			Residents
	Urban (N=468)	Rural (N=358)	Overall (N=826)	Urban (N=947)	Rural (N=184)	Overall (N=1131)
State
Kentucky	91 (19.4%)	63 (17.6%)	154 (18.6%)	290 (30.6%)	65 (35.3%)	355 (31.4%)
Massachusetts	133 (28.4%)	67 (18.7%)	200 (24.2%)	152 (16.1%)	21 (11.4%)	173 (15.3%)
New York	144 (30.8%)	125 (34.9%)	269 (32.6%)	148 (15.6%)	74 (40.2%)	222 (19.6%)
Ohio	100 (21.4%)	103 (28.8%)	203 (24.6%)	357 (37.7%)	24 (13.0%)	381 (33.7%)
Age
18–34 Years	76 (16.2%)	39 (10.9%)	115 (13.9%)	373 (39.4%)	37 (20.1%)	410 (36.3%)
35–49 Years	157 (33.5%)	141 (39.4%)	298 (36.1%)	251 (26.5%)	55 (29.9%)	306 (27.1%)
50–64 Years	187 (40.0%)	140 (39.1%)	327 (39.6%)	240 (25.3%)	62 (33.7%)	302 (26.7%)
65+ years	48 (10.3%)	38 (10.6%)	86 (10.4%)	83 (8.8%)	30 (16.3%)	113 (10.0%)
Race/Ethnicity
NH White	420 (89.7%)	339 (94.7%)	759 (91.9%)	771 (81.4%)	173 (94.0%)	944 (83.5%)
NH Black	18 (3.8%)	12 (3.4%)	30 (3.6%)	75 (7.9%)	1 (0.5%)	76 (6.7%)
Hispanic	21 (4.5%)	4 (1.1%)	25 (3.0%)	51 (5.4%)	2 (1.1%)	53 (4.7%)
Other	9 (1.9%)	3 (0.8%)	12 (1.5%)	50 (5.3%)	8 (4.3%)	58 (5.1%)
Gender Identity
Male	145 (31.0%)	121 (33.8%)	266 (32.2%)	251 (26.5%)	39 (21.2%)	290 (25.6%)
Female	321 (68.6%)	235 (65.6%)	556 (67.3%)	674 (71.2%)	143 (77.7%)	817 (72.2%)
Other	2 (0.4%)	2 (0.6%)	4 (0.5%)	22 (2.3%)	2 (1.1%)	24 (2.1%)
Education
High School or Less	12 (2.6%)	7 (2.0%)	19 (2.3%)	166 (17.5%)	46 (25.0%)	212 (18.7%)
Some College/Associate Degree	69 (14.7%)	44 (12.3%)	113 (13.7%)	364 (38.4%)	92 (50.0%)	456 (40.3%)
Bachelor Degree	125 (26.7%)	113 (31.6%)	238 (28.8%)	248 (26.2%)	26 (14.1%)	274 (24.2%)
Graduate Degree	262 (56.0%)	194 (54.2%)	456 (55.2%)	169 (17.8%)	20 (10.9%)	189 (16.7%)

%: Percentages may not add up to 100 due to rounding

Note: Abbreviations: HS=High School; NH=Non-Hispanic. Results restricted to respondents responding to all demographic items.

In total, 1,131 residents completed the CEQ survey, with 947 (84%) respondents located in urban communities and 184 (16%) located in rural communities. The greatest percentage of the urban residents were ages 18–34 (39%), whereas the greatest percentage of rural respondents were ages 50–64 (34%). The majority of resident respondents were non-Hispanic White (83%), female (72%), and had some college/Associate Degree (40%) or a Bachelor Degree (24%).

3.1. Stigma differences between coalition members and residents

As displayed in Table 3, aggregated across all communities, both coalition members and residents perceived community stigma towards people treated for OUD on all items except for more neutral responses to whether people would accept someone being treated for OUD as a close friend, that most employers would hire a qualified person who has been treated for OUD, or that people would be willing to date someone who has been treated for OUD. No differences between the CBS and CEQ were observed for the 8 general stigma items. However, coalition members perceived significantly greater levels of community stigma towards both MOUD (0.55 Least-Squares (LS) mean difference, p<0.0001) and naloxone (0.59 LS mean difference, p<0.0001) than residents. Coalition members also had significantly greater perceptions of community stigma towards drug treatment as an alternative to incarceration than residents (−0.40 LS mean difference, p<0.0001).

Table 3.

Difference in average responses to perceived opioid-related community stigma between coalition members and residents.

Question Label	Coalition Members (N=801)	Residents (N=1131)	LS Mean Difference (SE)	Adjusted P-Value1,2
	LS Mean (SD)	LS Mean (SD)
Close Friend	4.0 (0.14)	4.1 (0.13)	−0.10 (0.09)	0.40
Trustworthy	2.9 (0.13)	3.1 (0.12)	−0.16 (0.09)	0.16
Teacher	2.6 (0.13)	2.6 (0.12)	0.01 (0.08)	0.94
Child Caretaker	2.7 (0.13)	2.6 (0.12)	0.08 (0.08)	0.40
Think Less3	4.9 (0.12)	4.8 (0.11)	0.11 (0.07)	0.23
Hire	4.0 (0.13)	4.2 (0.12)	−0.18 (0.08)	0.06
Pass Over Applicant3	4.9 (0.12)	4.8 (0.11)	0.08 (0.08)	0.40
Date	3.8 (0.12)	3.8 (0.11)	0.03 (0.07)	0.74
MOUD3	5.1 (0.13)	4.6 (0.12)	0.55 (0.08)	<0001***
Naloxone3	4.6 (0.14)	4.0 (0.13)	0.59 (0.10)	<0001***
Incarceration	4.6 (0.13)	5.0 (0.12)	−0.40 (0.08)	<0001***

¹Linear mixed model including term for respondent role; covariates for site, age, race, gender, and education; and random effect for community.

²Adjusted for multiple comparison using Benjamini-Hochberg Adjustment.

³Reverse coded item, such that greater agreement (higher score) corresponds to more perceived community stigma.

^***p<0.001

Note: LS = Least Squares; MOUD = Medications for Opioid Use Disorder; SD = Standard Deviation; SE = Standard Error. Results restricted to respondents responding to all demographic items presented and having a valid response to at least 1 stigma question. Community members had to have completed and consented to questionnaire. Each question is rated on a scale of 1–7: strongly disagree (1), disagree, somewhat disagree, neither agree or disagree, somewhat agree, agree, strongly agree (7). Respondents were also given an option to not answer the question.

3.2. Stigma differences between coalition members and residents by urban/ rural status

Differences in perceived community stigma items between coalition members and residents by urban/rural status are in Table 4. Within urban communities, coalition members perceived less community stigma about hiring someone who has been treated for OUD to take care of their children (0.2 LS Mean Difference, p=0.05) than residents. However, urban coalition members perceived greater stigma about the use of MOUD (0.5 LS Mean Difference, p<0.0001) or naloxone (0.5 LS Mean Difference, p=0.0007), and offering drug treatment as an alternative to incarceration (−0.3 LS Mean Difference, p=0.0035) than did urban residents.

Table 4.

Differences in average responses to perceived opioid-related community stigma between coalition members and residents by urban/ rural status.

	Stratified Analysis								Full Model1
Question Label	Urban				Rural
	Coalition Members (N=451)	Residents (N=947)			Coalition Members (N=350)	Residents (N=184)
	LS Mean	LS Mean	LS Mean Difference (SE)	Adjusted P-Value1,2	LS Mean	LS Mean	LS Mean Difference (SE)	Adjusted P-Value1,2	Adjusted Interaction P-value1,2
Close Friend	4.2 (0.15)	4.1 (0.13)	0.1 (0.10)	0.62	3.8 (0.15)	4.2 (0.17)	−0.4 (0.15)	0.03*	0.03*
Trustworthy	3.1 (0.14)	3.1 (0.12)	0.0 (0.09)	0.98	2.8 (0.15)	3.3 (0.16)	−0.5 (0.14)	0.0035**	0.01*
Teacher	2.7 (0.13)	2.6 (0.12)	0.1 (0.09)	0.26	2.5 (0.14)	2.8 (0.16)	−0.2 (0.14)	0.16	0.06
Child Caretaker	2.8 (0.13)	2.6 (0.12)	0.2 (0.09)	0.03*	2.5 (0.14)	2.8 (0.16)	−0.3 (0.14)	0.10	0.01*
Think Less3	4.8 (0.13)	4.8 (0.11)	−0.0 (0.09)	0.92	5.1 (0.14)	4.8 (0.15)	0.3 (0.14)	0.06	0.08
Hire	4.1 (0.13)	4.2 (0.12)	−0.1 (0.09)	0.39	3.9 (0.14)	4.3 (0.16)	−0.4 (0.14)	0.03*	0.13
Pass Over Applicant3	4.9 (0.13)	4.9 (0.12)	−0.0 (0.09)	0.85	5.0 (0.14)	4.6 (0.15)	0.3 (0.14)	0.03*	0.05
Date	3.8 (0.13)	3.7 (0.11)	0.1 (0.09)	0.38	3.8 (0.13)	3.9 (0.15)	−0.2 (0.13)	0.24	0.10
MOUD3	5.0 (0.14)	4.6 (0.13)	0.5 (0.10)	<0001***	5.3 (0.15)	4.6 (0.16)	0.7 (0.14)	<0001***	0.18
Naloxone3	4.4 (0.15)	3.9 (0.14)	0.5 (0.11)	0.0007**	4.8 (0.16)	4.1 (0.18)	0.7 (0.16)	0.0002**	0.19
Incarceration	4.7 (0.14)	5.0 (0.12)	−0.3 (0.10)	0.0035**	4.4 (0.15)	4.8 (0.16)	−0.4 (0.14)	0.01*	0.51

¹Linear mixed model including interaction between respondent role and urban/ rural location; covariates for site, age, race, gender, and education; and random effect for community.

²Adjusted for multiple comparison using Benjamini-Hochberg Adjustment.

³Reverse coded item, such that greater agreement (higher score) corresponds to more perceived community stigma

^*p<0.05;

^**p<0.01;

^***p<0.001

Note: LS = Least Squares; MOUD = Medications for Opioid Use Disorder; SD = Standard Deviation; SE = Standard Error. Results restricted to respondents responding to all demographic items presented and having a valid response to at least 1 stigma question. Community members had to have completed and consented to questionnaire. Each question is rated on a scale of 1–7: strongly disagree (1), disagree, somewhat disagree, neither agree or disagree, somewhat agree, agree, strongly agree (7). Respondents were also given an option to not answer the question.

Rural residents perceived less community stigma than rural coalition members on many survey items. Specifically, among coalition members and residents in rural communities, there were significant differences in the perception of whether someone would accept a person being treated for OUD as a close friend (−0.4 LS Mean Difference, p=0.03), how trustworthy a person is who has been treated for OUD (−0.5 LS Mean Difference, p=0.0035), and whether an employer would hire (−0.4 LS Mean Difference, p=0.03) or pass over someone who had been treated for OUD (0.3 LS Mean Difference, p=0.03). The differences in perceptions of stigma towards MOUD (0.7 LS Mean Difference, p<0.0001), naloxone (0.7 LS Mean Difference, p=0.0002), and incarceration (−0.4 LS Mean Difference, p=0.01) were also significant and consistent with the results in Table 3.

There were three items for which the study detected significant interaction effects. Rural coalition members perceived less community stigma about people being willing to have a person treated for OUD as a close friend than urban coalition members (p=0.03). Rural coalition members also had more perceived stigma (relative to urban coalition members), and rural residents less perceived stigma (relative to urban residents), about people in their community believing that a person treated for OUD was trustworthy (p=0.01) and that people would hire someone treated for OUD as a child caretaker (p=0.01).

4. DISCUSSION

Acceptance and uptake of EBPs to prevent or reduce opioid-overdose deaths can be severely hampered by community stigma towards people with OUD and the EBPs to prevent overdoses and treat OUD (Kennedy-Hendricks et al., 2017b; Madden et al., 2021). Our analysis found that these general perceptions about community’s stigma did not differ between local experts directly involved in their community opioid response efforts and residents. However, our analysis did identify discordance between the two groups related to EBPs to address the opioid overdose crisis – MOUD, naloxone, and drug treatment as an alternative to incarceration. Coalition members consistently, and significantly, perceived greater community stigma towards these interventions than did residents in both rural and urban communities.

These differences in perceived stigma towards EBPs, to a certain extent, may be expected given that coalition members have high familiarity and engage with more specific referent groups (e.g., colleagues who also work with OUD issues and service providers), and as a result may have exposure to a different constellation of beliefs, attitudes, and norms than other community members. Compassion fatigue (negative emotions after helping others repeatedly) may also contribute to the perceptions of coalition members (Beaugard et al., 2022; Winstanley, 2020). Extant research suggests compassion fatigue as a factor that undermines community support for initiating, continuing, or expanding EBPs (Winstanley, 2020), yet the relationship between compassion fatigue and perceived stigma were not explored in this current study, and remains an important area for future research.

Another possible explanation for the differences in perceived stigma, especially the lower level of perceived stigma around EBPs by community residents is that residents may have varying levels of familiarity with OUD through acquaintances, extended family, coworkers, or nuclear family members (Corrigan & Nieweglowski, 2019). They may perceive less community stigma as they attribute OUD to inadequate treatments for pain, improper prescribing practices, and pharmaceutical marketing (Kennedy-Hendricks et al., 2017a). Residents may also have less familiarity with specific interventions such as naloxone or MOUD, and as a result may be unaware of the level of stigma that other segments of the community may have towards them.

4.1. Public Health Implications

Prior work has found that coalitions may create closed or ‘dense’ networks whereby coalition members primarily seek information from others in their coalition and may not engage many outside sources of information from the community (Valente et al., 2007). As a result of this density, the coalition may be less effective at making decisions and mobilizing the resources they need to adopt EBPs. To counter these forces, coalition members may need more deliberate effort to engage with a greater range of residents, including those with direct or indirect lived experience with OUD, as equals in planning and decision making (Committee on the Science of Changing Behavioral Health Social Norms et al., 2016). In their professional roles, coalition members may frequently interact with individuals with OUD (e.g., as treatment or emergency services providers) that have embedded hierarchies (Link & Phelan, 2001). Yet, these power dynamics may undermine coalition members’ abilities to create an inclusive environment for individuals with OUD and color their own perceptions of the utility of EBPs to address the opioid crisis in their community.

In addition to promoting representation and inclusivity of people with lived experience in coalitions and other public facing groups, having them as central figures in communications campaigns could enhance their reach and impact within a community. Additional educational efforts for the public are also needed to highlight OUD as a treatable chronic health condition (Kennedy-Hendricks et al., 2019). Directly addressing misinformation and negative attitudes among the public and local stakeholders about the benefits of naloxone as a life-saving medication and that medications (buprenorphine and methadone) are the most effective treatments for OUD are needed to reduce stigma, increase the acceptance of EBPs to reduce the toll of opioid overdoses, and promote public health policies.

4.2. Limitations

Our findings should be considered with key limitations in mind. First, our survey used measures of perceived community stigma, rather than objective measures of community or structural levels of stigma, and as a result, reports of perceived community norms may not directly reflect individual-level stigma. Therefore, it is not possible to validate if one population has a more accurate perception, nor if perceived community stigma is in line with enacted or other levels of stigma in these communities. Second, recruiting residents via social media limited these survey respondents to those who had access to internet, social media accounts, and logged on during the recruitment period (Forgasz et al., 2018). Social media recruitment is prone to over-representation of non-Hispanic White females (Darko et al., 2022; Whitaker et al., 2017) – a phenomenon observed in our sample as compared to the general population demographic characteristics (see Supplemental Material – Appendix Table 1). Third, researchers’ sampling methods to recruit respondents to either the CBS or CEQ surveys possibly resulted in biased perspectives regarding community opioid-related stigma from the population they are intended to represent and over-representation among highly motivated respondents. Similarly, both the CEQ and CBS samples skew younger, non-Hispanic White, and females in comparison to the general population statistics. Prior research identified differences in stigma perceptions for drug users based on both the race of the perceiver and the race of the drug user (Goodyear et al., 2022), though less evidence supports differences in perceived stigma by gender (Wood & Elliott, 2020). Nonetheless, it is possible that the bias in our sample could overestimate perceived community stigma in the residents group, ultimately resulting in conservative estimates of differences between the residents and the local experts. This phenomenon could contribute to our null findings for the OUD stigma items between the community stakeholders and the residents. As a result, these null findings should be interpreted cautiously, as more differences may exist between the two populations than identified in our sample. Fourth, respondents were not specifically queried on effective approaches to reduce perceived community stigma – the outcome assessed in our analysis. As a result, our findings underscore the importance of identifying, testing, and promoting effective approaches to reduce stigma.

Finally, the intersectionality of stigma with racial prejudices can compound to negatively impact communities of color (Bowleg, 2021) and the community responses to the opioid epidemic. Pyra et al. (2022), demonstrated that negative attitudes towards Black individuals and conservative political beliefs, in addition to stigma, are associated with less support for public-health responses to the opioid epidemic. While our findings show similar differences about specific interventions (i.e., MOUD, naloxone, incarceration) between coalition members and residents within urban and rural areas, we did not capture political or racial prejudice measures that would enable evaluation of how these intersectional forces may operate in other geographic regions with different prevailing beliefs about race and/ or politics. Understanding the intersectionality of community stigma towards OUD with racism remains an important area for future inquiry in order to advance health equity.

4.3. Conclusion

This study is the first to our knowledge demonstrating that local experts involved in the opioid response hold higher levels of perceived community stigma and intervention stigma as compared to residents in both rural and urban communities, especially with respect to MOUD, naloxone, and drug treatment as an alternative to incarceration. Our findings highlight a critical question concerning whether and how intervention stigma and perceptions of community stigma may relate to decisions about EBPs to address the opioid epidemic. Moving forward, continuing to reframe the narrative around OUD as a treatable chronic medical condition, and that effective treatments are known and available, may reduce stigma and help promote public support for these EBPs. However, perceptions of community stigma should not undermine evidence-informed approaches. Instead, implementation strategies should acknowledge that the opioid epidemic and substance use bring genuine harms to communities, and endorsing evidence-based approaches is the most effective, albeit imperfect pathway to ameliorating these harms and to promote recovery.

Highlights.

• Similar perceived opioid-related stigma between coalition members and residents.

• Differences related to evidence-based practices (i.e., MOUD, naloxone, and drug treatment as an alternative to incarceration) to reduce opioid overdose deaths.

• Need to reframe the narrative around OUD as a treatable chronic medical condition.