“With a little help from my friends” - collaborating with community experts to improve sampling and obtain a high response rate for a statewide naloxone availability survey

Abstract

Background

Research and advocacy emphasize the importance of naloxone saturation not just in terms of how much naloxone is distributed but also to whom, by whom, and how it is distributed. Given the nuanced ways in which naloxone is distributed, community experts involved with naloxone distribution and harm reduction work can help researchers better assess naloxone availability. The objective of this paper is to describe collaborating with community experts to obtain a diverse sample and high response rate of a statewide survey investigating community-based naloxone availability.

Methods

This survey is part of a larger study that aims to develop and disseminate visual tools, such as a website with coverage maps to show naloxone sources, availability, and associated cost by county. Community experts, consisting of advisory board members and experts on the study team, represented the different constituencies from which study data would be sampled and for their expertise on community harm reduction work. Surveys were sent to individuals if they were identified as a representative of a program believed to potentially distribute naloxone at no-cost to community members in NC by the study team or community advisory board. Data collection occurred over the course of 8 months (June 2023-January 2024) and consisted of emails, phone calls, and mailers.

Results

The overall response rate was 81.4%. In collaboration with community experts, the study team surveyed eight additional program types from the original sample list. Community experts helped identify contact information for program types and participated in triangulation during data cleaning. Four out of five program types with response rates above 80% had received personalized emails from a community expert.

Conclusions

Community experts informed decision-making on survey design, sampling, survey distribution, and data cleaning processes for a survey designed to identify community-based sources of no-cost naloxone across NC. The immense contributions of the community experts for this statewide survey, which we are confident improved our sample, also likely increased the survey response rate. This underscores the importance of engaging community members in every step of research to produce relevant and meaningful outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-026-27526-7.

Background

There were over 75,000 drug overdose deaths involving opioids in the United States (US) in 2023 [1]. In North Carolina (NC), nearly 12 people died each day from a drug overdose in 2022 [2]. While opioid overdose deaths persist across the country and in NC, disparities in opioid overdose deaths are growing. For example, overdose deaths in NC increased 72% from 2019 to 2021; however, American Indian/Indigenous populations experienced an increase of 117%, and Black individuals experienced an increase of 139% during that same time period [3]. A key harm reduction strategy to reduce opioid overdose deaths is the distribution of naloxone [4, 5]. Naloxone is an opioid antagonist medication that effectively reverses opioid overdose and saves lives [6]. Though the value of widespread naloxone access and distribution are well-recognized [4, 7–10], a growing body of research and advocacy also emphasize the importance of naloxone saturation not just in terms of how much naloxone is distributed but also to whom, by whom, and how it is distributed [11, 12].

Targeted naloxone distribution

Despite the emergence of over-the-counter (OTC) naloxone in community pharmacies, the high cost and varied availability of OTC naloxone underscore the need for continued no-cost naloxone distribution [13–15]. Targeted naloxone distribution, particularly to those who use drugs and are at high risk of overdose, is the most effective way to prevent overdose deaths [16, 17]. In addition to prioritizing naloxone distribution among people who use drugs (PWUD), providing naloxone to other potential witnesses of overdose, including friends and family of people at risk, can also prevent deaths [16, 18, 19]. Targeted distribution has also been recommended for those leaving drug treatment programs, people discharged from hospitals following an overdose event or with a history of opioid use, people recently released from incarceration, and unhoused populations [20–27].

In the US, Syringe Service Programs (SSPs), which distribute supplies for safer drug consumption and to prevent drug use-related infections and overdose deaths, have provided free naloxone since the 1990s [28], giving out close to 1 million doses per year directly to PWUD [29]. Providing naloxone through SSPs offers targeted distribution to people at greatest risk of overdose and is thus considered the most evidence-based approach to reduce overdose deaths [4, 8, 12, 19]. As of mid-2023 fifty SSPs were registered with the state of North Carolina; these programs collectively distributed nearly 110,000 naloxone kits in the 2022–2023 reporting year [30]. Even so, research suggests that SSPs are underutilized due to factors such as lack of access, stigma, and public perceptions of who SSPs serve, among others [31–38]. For example, people who consume non-prescribed substances by means other than injection may underestimate their overdose risk and/or not perceive SSPs as offering safer consumption supplies relevant to them [39].

Thus, while SSPs distribute a large amount of naloxone to the most at-risk populations, distribution by other means is also necessary to reach people who may not have access to or may not directly interface with SSPs [35, 39–41]. Local health departments (LHDs) are known to provide no-cost naloxone to community members [42, 43]. Post Overdose Response Teams (PORTs) are another widely used model that, ideally, provide naloxone to people who have recently experienced a non-fatal overdose, and, if following the most evidence-based protocols, offer referrals to harm reduction services, other social services, and linkage to care [44, 45]. However, law enforcement involvement in these programs may limit their impact and appeal [46, 47]. Community paramedicine teams have been suggested as another method to reach PWUD following a non-fatal overdose, as they function to provide services and resources to underserved populations and bridge healthcare gaps [48–50]. Other sources of no-cost naloxone distribution in NC include patients treated at registered opioid treatment programs (OTPs) and through jail-based naloxone vending machines (accessible at the time of release), which each increase access for specific populations of people at known increased risk for overdose [51, 52].

Despite the demonstrated need for and effectiveness of community-based SSPs and other community programs distributing no-cost naloxone, funding for such naloxone is precarious and sometimes difficult to acquire especially for community-based organizations (CBOs) [53]. This is especially alarming as community-based SSPs have been found to provide more harm reduction services than those managed by public health departments, and increased funding for community-based SSPs is associated with greater naloxone distribution [54, 55]. Since some CBOs distribute no-cost naloxone without government funding, it follows that naloxone distribution programs may not be comprehensively captured on state-maintained naloxone distribution lists [55].

Nationally, organizations that distribute naloxone to populations at greatest risk for overdose may join Remedy Alliance (RA), a 501(c) [3] non-profit organization. Members of RA may purchase naloxone in various formulations at reduced cost, or request a subsidized rate based on their funding status. Within NC, from August 2022 to August 2024 RA provided over 380,000 doses to member organizations to distribute naloxone [56]. Organizations may also submit requests to the state Department of Health and Human Services (NCDHHS) for no-cost naloxone, with priority given to those actively distributing to “high-risk” populations. Requests to NCDHHS are fulfilled on a rolling basis and neither fulfillment nor quantities are guaranteed [57].

Community-based harm reductionists and affiliated researchers emphasize that not only the amount of naloxone distributed but also accompanying education on how to recognize signs of overdose and effectively administer naloxone are necessary to effectively reduce harm [11]. These nuances regarding how much naloxone is distributed, to whom, by whom, and how, speak to a need to engage affected community partners about their naloxone needs [5].

Community expert engagement

Community-engaged research, through key informant and community member input, has long been recognized as key to enhancing research validity and facilitating the application of research to patient-and-community-centered outcomes [58–62]. Involving diverse community members representing constituencies likely affected by the focus of the research is widely recognized to increase the likelihood that findings can be applied to address barriers and disparities especially in public health and health policy research [59, 63–67]. One way community members can be engaged in research is as members of study teams and community advisory boards [58, 59]. Although several studies have reported on the value of community advisory boards in study recruitment [68–70], to our knowledge, few research methodology papers have focused on how community-engaged research experiences can inform sampling and survey distribution processes. Recent literature suggests the average response rate for online or email-based surveys varies from 44 to 68% across social science and education research [71, 72]. Given the importance of high response rates to ensure validity and increase generalizability [73, 74], collaboration with community experts in sampling and survey distribution can be a valuable tool to enhance response rates. Incorporating such community expertise can be especially impactful with potentially challenging or ambiguous research circumstances, such as when collecting data at an organizational level.

Aim

The aim of this paper is thus to describe how community experts on the study team and community advisory board helped expand the sample and support survey distribution for a cross-sectional survey of naloxone distributors across the state. We describe in detail the composition of the community advisory board and study team, including representatives from tribal communities, CBOs, state and local health departments, and community pharmacies, and how they contributed to the various stages of the survey. This survey is part of a larger study that aims to develop and disseminate visual tools, such as a website with coverage maps to show naloxone sources, availability, and associated cost by county. The larger study will combine primary data collected from a secret shopper study of community pharmacies [14, 15, 75] and the survey described herein with secondary data from public and private insurance claims. In this paper, we describe the process of working with community experts to identify the appropriate survey sampling and distribution approach, and the resulting diverse sample and high response rate.

Methods

Ethics

The study protocol for the larger study of which the survey was a subset was reviewed by the Institutional Review Board at the University of North Carolina-Chapel Hill and found to be Not Human Subjects Research. Although the survey was determined Not Human Subjects Research, informed consent was nevertheless obtained through a statement of consent that preceded the survey for all participants.

Community expert design

Community expert involvement consisted of a community advisory board and several study team members. Criteria for community advisory board members included a diversity of geographic representation from across NC; a diversity of racial/ethnic representation; and purposive recruitment of members to represent the different constituencies from which study data would be sampled. In addition, investigators purposively recruited key informants familiar with the landscape of naloxone distribution in NC. The Principal Investigator (DC) directly recruited the majority of community advisory board members; in a few cases, other study team members approached prospective members they knew from other settings. Community advisory board members consisted of representatives from Health and Human Services with the Eastern Band of Cherokee Indians, a representative from the North Carolina Department of Health and Human Services (NCDHHS) involved in naloxone distribution planning, several local health department employees working on substance use-related programming, a former NCDHHS liaison for justice-involved programs, a community pharmacist, a director of Emergency Medical Services (EMS) programs, and a community-based harm reduction program client who had personally used naloxone to reverse more than 100 overdoses. Community advisory board members were compensated for participation at a rate of $75.00 (USD) per half-hour meeting.

Several study team members were recruited by the principal investigator due to their expertise and professional and/or personal involvement in substance use disorder (SUD) treatment work, harm reduction efforts, and related topics. For example, study team members included a long-time provider of medications for opioid use disorder (MOUD) with extensive experience advocating for naloxone access in primary care and SUD treatment settings (BF); an enrolled member of the Lumbee tribe and researcher on how substance use impacts tribal communities (RB); and a community-based researcher with extensive harm reduction expertise (BO) who led the Aim 1 survey group tasked with survey distribution and data interpretation.

Positionality

As this topic is of interest to us not only scientifically and professionally, we consider it worthwhile to briefly discuss our related positionalities. We all feel personally invested in improving availability of naloxone and ensuring it is strategically accessible to populations at greatest risk for overdose death. At least four authors of this paper have personally lost loved ones to overdose. Two authors of this paper were members of the community advisory board described above. At least one author of this paper is also a participant of a rural SSP. Authors include two pharmacists, a Medications for Opioid Use Disorder (MOUD) provider, four public health professionals and researchers, and a community-based harm reduction researcher who works closely with many of the programs sampled.

Sampling

Our goal was to survey all programs that distribute no-cost naloxone in NC. Individuals were considered eligible for survey participation if they were identified (by community experts on the study team or advisory board) as a representative of a program believed to distribute naloxone at no-cost to community members in NC. We followed an open-ended sampling approach without a specific target sample size and added additional categories of prospective respondents as we refined the sample.

Data collection

Following The Dillman Tailored Design Method [76] for survey distribution, a subset of the larger study team, hereafter referred to as the Aim 1 survey group (led by a community expert on the study team), conducted data collection over the course of eight months (June 2023-January 2024) consisting of a three-step process with multiple phases and modes of distribution. The Aim 1 survey group began survey distribution with emails; followed by phone calls; and finally distributed mailers to offer ample opportunity for survey completion. Survey links were created in Qualtrics, an online survey software (Qualtrics, Provo, UT). Recruitment included a monetary incentive (Amazon e-gift card worth $20) for programs that reported no-cost naloxone distribution. A monetary incentive has shown to increase the response rate for online surveys [77]. Individuals who indicated that their program did not distribute naloxone at no-cost were directed to the end of the survey and were not eligible to receive a gift card. Each contact in the study sample was sent a unique link, so participation could be tracked.

Survey distribution step 1: email distribution

Survey links with introductory emails were sent in batches by program type (i.e., SSP, LHD, PORT, etc.) to allow for bulk reminders to be sent at intervals. Email distribution occurred from June to August of 2023. Follow-up emails were sent to non-respondents at 2 weeks and 4 weeks post-initial email. A fourth email was sent prior to calling non-respondents to ensure survey recipients had an active survey link.

Each email mentioned the recipient’s program type (i.e., “registered syringe service program,” “local health department,” “tribal organization,” etc.) and the specific program name (e.g., “[Harm Reduction] SSP,” or “[Example] County Health Department”). The email also explained the survey’s purpose, estimated completion time, and incentive.

Survey distribution step 2: phone distribution

No sooner than four weeks after all non-responding programs had received up to four emails, phone call follow-up was initiated. Two Aim 1 survey group members conducted phone calls from September to December 2023 using a standardized script. If reached, contacts were given the opportunity to fill out the survey over the phone or to use the survey link previously sent to their email. Most opted to fill out the online survey independently after the call and requested the survey be resent. If contacts did not answer the phone, messages were left either on voicemail or with a receptionist. Voicemails included a callback number and information about the last email contact (date and from whom the email was sent).

At least one week in between calls was given to allow people the opportunity to complete the survey. A call was considered “complete” only if the potential participant was directly spoken to, left a voicemail, or left a message with a receptionist or secretary. Aim 1 survey group members made up to three “complete” calls to each contact. Several contacts, when called, informed the Aim 1 survey group of additional programs distributing naloxone that had not been identified in the original study sample. These four programs were sent a survey link, triggering the beginning of The Tailored Design Method [76].

Survey distribution step 3: mail distribution

No sooner than two weeks after the final phone call was made to all non-responding programs, a survey mailer was sent in mid-December 2023. The mailer contained a flyer briefly re-explaining the study and invited them to complete the survey using a QR code that linked to the survey. Individuals were given until the end of January 2024 to complete the survey, at which time data collection ceased. Programs in which representatives did not respond to any survey request method were considered non-respondents.

Results

Survey design contributions

The survey used in this study is an original survey developed in collaboration with community experts. Survey design is described in more detail elsewhere; however, community experts were involved in providing feedback on early drafts and subsequent iterations of the survey itself [78]. This included offering the study team suggestions for question and response option wording; input on whether to add a question related to over-the-counter naloxone (a new development at the time the survey was drafted); and generally ensuring survey wording would be as accessible as possible to the various constituencies and sectors community experts represented. A pdf version of the survey is provided in additional file 1 in Supplementary Material. The community expert contributions to the sampling and survey distribution approach, in which this article is focused, are described in more detail below and in Table 1.

Table 1.

Community expert contributions to sampling and survey distribution approach

Study Phase	Community Expert Contributions
Sample Identification	● Identified 8 additional program types to include in sample ● Helped procure contact information for many respondents in different program type categories
Data Collection	● Identified best contacts for personalized introductions (warm hand-offs) to precede survey link emails ● Provided ‘warm hand-offs’ for emails ● Helped identify alternate contacts for bounced contacts
Data Cleaning	● Participated in triangulation to determine if a survey response represented multiple programs within one organization

Sample identification contributions

The study team initially identified two program types that likely distribute naloxone at no-cost to community members in NC (syringe service programs (SSPs) registered with the state Department of Health and Human Services and recipients of no-cost, state-subsidized naloxone from NCDHHS). Upon iterative review with the community advisory board and study team, eight additional program types were identified and added to the target sample list. The complete array of sampled program types were wide-ranging: syringe service programs (SSPs) registered with the state Department of Health and Human Services; federally and/or state-recognized tribal organizations; community paramedicine teams; Post-Overdose Response Teams (PORT); local health departments (LHDs); Local Management Entity/Managed Care Organizations (LME-MCOs) that in NC manage behavioral health/substance use disorders treatment funds for Medicaid-eligible residents at the regional level; a variety of programs all characterized as “justice-involved” that included law enforcement diversion, drug and treatment courts, jails and prisons that provide naloxone upon release, and jail-based naloxone vending machines; a philanthropic organization; a medical relief organization; other community organizations that did not fall into any of the aforementioned categories; and organizations that receive no-cost naloxone for community distribution from either NCDHHS or Remedy Alliance and were not on any of the preceding lists.

Contact information for representatives of programs in each of the above categories was identified through input from community advisory board members and other study team members, as well as key informants representing the category/ies. In some cases, different contact information was provided through referrals from initially identified contacts. For most contacts, the original contact information lists acquired through the community advisory board and study team included the individual’s phone number and mailing address. Aim 1 survey group members used an online search engine to identify otherwise missing phone numbers and mailing addresses.

Data collection contributions

In collaboration with the community advisory board, the Aim 1 survey group identified an ideal contact person affiliated with each program type to provide a “warm hand-off”, defined as a personalized note at the top of the email and an email signature from that person. In addition, the email distribution feature in Qualtrics allows a “from” name to be added to the email, such that recipients would see the contact’s name affiliated with their program type upon receiving the email. For example, for recipients of state-provided naloxone, the state NCDHHS representative with whom they placed their naloxone orders introduced the email, signed the email, and had their name listed in the email’s ‘from’ line. Community advisory board members and community expert study team members provided personalized notes for five program types. An example personalized email is included in additional file 2 in Supplementary Material.

There were 31 email addresses in the original sample that could not be delivered. To reconcile these undeliverable (bounced) contacts, the Aim 1 survey group first checked if there was another contact surveyed from the same program. To increase the chance of receiving a survey response from every program, if there were multiple contact people at each program, each was sent a survey link. If no other program contact was available for bounced emails, then alternate contacts were sought through input from the community advisory board and study team, or failing that, through further online searching. Community experts suggested over half of the alternate contacts identified and most of these contacts responded to the survey.

Data cleaning and response rate contributions

Prior to survey distribution, the Aim 1 survey group reviewed and compared contact lists obtained by community experts to identify and reconcile duplicate contacts. Therefore, the response rate broken down by the sampling lists community experts suggested and helped procure (Table 2) does not necessarily reflect the total number of programs within each program type, but rather a simplified list that excluded known duplicates. Moreover, for programs with multiple survey responses, we retained the most complete response using a standardized process. Through iterative comparisons the Aim 1 survey group reached consensus on which response to retain among duplicate survey responses for 38 programs.

Table 2.

Program types & response rates

Program Type	Number of Programs	Number of Responses (response rate for program type)
Remedy Alliance - NC Recipients a	7	7 (100%)
Registered SSPs (as of 2023) a	38	37 (97.4%)
Local Health Departments (LHDs)	80	73 (91.3%)
Tribal Organizations a	8	7 (87.5%)
NC Department of Health and Human Services - Recipients a	104	85 (81.7%)
Community Paramedicine programs	34	27 (79.4%)
Other Community Programs (including add-on programs)	9	7 (77.8%)
Post-Overdose Response Teams (PORTs)	39	29 (74.4%)
Justice-Involved Programs a	90	63 (70%)
Local Management Entity/ Managed Care Organizations (LME-MCOs)	5	2 (40%)
All Programs	414	337 (81.4%)

^a Program types that received a personalized email from a community expert on the study team or advisory board

The Aim 1 survey group identified an additional 23 program responses during data cleaning when it became apparent that some respondents completed surveys that represented more than one program type. For example, there were circumstances in which one person was the contact for multiple programs based at one organization (e.g., an SSP housed within an LHD; or a PORT based at an LHD or within EMS). This could lead to one survey response representing no-cost naloxone distribution by multiple programs. Such determinations were made through iterative consultations with the community advisory board and study team, when the Aim 1 survey group felt triangulation with a relevant member or members most familiar with the relevant program type would be helpful. For instance, triangulation with a member of the community advisory board that represents the Eastern Band of Cherokee Indians verified that the survey response filled out for the Tribal Organization also served as the response for the registered SSP affiliated with the Tribal Organization’s LHD.

The overall program response rate reached 81.4%. This response rate was calculated by the number of distinct programs represented in the sample. Accounting for the removal of duplicates (n = 38) and inclusion of additional program responses captured by an existing survey response (n = 23), there were a total of 337 responding programs. Thirteen programs were removed from the original sample size (n = 427) because they did not meet inclusion criteria (i.e. non-existent, unreachable, did not distribute naloxone directly to people, or did not distribute naloxone in North Carolina), yielding a final sample size of 414 programs. The flowchart (Fig. 1) demonstrates the data collection process and resulting survey responses.

Fig. 1.

Survey data collection process

Table 2 demonstrates the response rates broken down by the sampling lists community experts suggested and helped procure. Four out of five program types with response rates above 80% had received personalized emails from a community advisory board member or community expert study team member who represents or works closely with those program types.

Discussion

Our diverse sample and high survey response rate are likely the result of the contributions of the community experts on the advisory board and within the study team. Though the iterative process of identifying and refining the survey sample and then gathering and updating contact information for respondents was time-consuming and often repetitive, by conferring with community experts on the advisory board and study team, the Aim 1 survey group was increasingly confident that we were including a more comprehensive sample consisting of more programs and more program types than we would have otherwise identified. This process exemplifies a way in which community-engaged research can enhance survey methods, such as the Tailored Design Method, which suggests using multiple sampling frames to reduce coverage limitations [76].

Participation of individuals directly involved or familiar with community naloxone distribution yielded a more comprehensive and diverse sample grounded in local knowledge, while previous studies with similar aims used only state-managed lists or state-provided distribution data in their assessment of naloxone access [79–81]. Likewise, direct input from those working in the field likely yielded the most relevant contacts for survey distribution. Finally, having community experts willing to provide ‘warm hand-offs’ likely contributed to our high response rate. These findings reinforce existing literature that personalized survey invitations can help improve survey response rates and aligns with The Tailored Design Methods’ suggestion for increasing response rates by building upon previously established relationships [76, 82]. Together these approaches demonstrated the specific methodological value of recognizing community members as experts, rather than relying only on ‘official’ data sources.

Contact information for programs were often outdated, incomplete, or included inaccuracies, evident in part by the 31 bounced contacts from the original sample. Staff turnover and the nuanced settings in which naloxone is distributed, as well as the multi-faceted program design(s) that can result as a function of funding availability, contribute to this problem [19, 44, 48, 49, 51–53, 55]. Resulting complexities in the sampling and survey distribution process emphasizes the importance of involving community experts not only in study design but also in data analysis and interpretation; so that those most involved in program work can help fill in gaps. We are confident the methodological findings described in this paper hold relevance beyond the immediate topic on which our survey gathered data; to wit, other researchers seeking to collect program-level data, especially by surveying programs for which there is not a comprehensive contact list, can greatly benefit from working with community experts at every stage of research.

Limitations

The meeting schedule and format for the community advisory board was not easily accessible for some recruited members. Due to the statewide nature of the advisory board, it met virtually, requiring internet access and a device that could connect to the internet. A community advisory board member recruited for their lived experience as a member of a syringe service program had difficulty remaining engaged with the advisory board because of these requirements, particularly when struggling to maintain access to a phone. In addition, meetings were also held during traditional work hours. Although advisory board members were asked what days and times were most convenient for them, those with traditional 9am-5pm jobs were less consistently involved in attending meetings. Future studies focused on reducing overdose death risks should consider how best to engage PWUD and other community experts in more precarious situations in study design and interpretation. For example, providing community advisory board members with prepaid cell phones or codes to refill minutes on prepaid phones might have increased access for the initial SSP client who encountered difficulty remaining involved. For those who find large Zoom meetings intimidating, having smaller, regional, in-person advisory board meetings, and/or grouped by area of expertise, might make participation more comfortable.

Data cleaning challenges potentially affected our response rate calculation, leading to over-or-underestimates of survey responses. After reconciling such responses, often in consultation with the larger study team and/or community advisory board, the Aim 1 survey group adjusted program response rates accordingly. In another such study, confirming contact information prior to data collection can improve the quality of contact lists and ease the data cleaning process.

Conclusion

Given the limitations encountered during sample compilation, sampling, survey distribution, and data cleaning, researchers seeking to conduct statewide surveys of public health initiatives should consider confirming contact information for programs prior to the start of data collection. This additional step can improve data quality and response rates by ensuring the best contact person for any given program receives the survey link. Beyond this, it would be helpful to confirm which organizations have multiple, distinct programs (PORT, SSP, etc.) that may distribute naloxone to separate or even overlapping populations of interest, and if so, verify contacts for each program within the organization to reduce duplicate responses and ensure all programs’ data are obtained. Moreover, study teams should allow enough time for identifying a comprehensive sample and compiling accurate contact lists on the front end; and for reconciling and triangulating responses following data collection to determine an accurate response rate. A full list of recommendations for survey replication and community expert involvement can be found in Table 3. Ultimately, given the many complexities of survey research, producing high-quality sample and contact lists in conjunction with community experts can enhance future research on naloxone availability. Likewise, these recommendations can be broadly applied to individuals conducting survey research with organizations.

Table 3.

Recommendations for Survey Replication

Recommendation	Community Expert Involvement
Confirm contact information for programs prior to the start of data collection	● Gathering contact lists ● Identifying alternate contacts for any contacts that prove to be out of date ● Help contact programs to confirm information before starting data collection
Confirm organizations that appear to have multiple, distinct programs	● Review contacts lists to flag potential redundancies
Allow enough time for identifying a comprehensive and accurate sample	● Participate in iterative review of sample ● Ensure funding proposals include long enough window for community advisory board compensation
Allow enough time for reconciling and triangulating responses following data collection	● Participate in triangulation

Community advisory board members and community expert study team members informed decision-making on survey design, sampling, survey distribution, and data cleaning processes for a survey designed to identify community-based sources of no-cost naloxone across North Carolina. Community experts provided contact lists for several program types, suggested multiple programs distributing free naloxone that were not previously identified and provided specific contacts in some cases, provided personalized notes for warm hand-offs through email distribution of survey links, and participated in triangulation discussions to reach consensus during data cleaning. The immense contributions of the community experts for this statewide survey, which we are confident improved the scope and quality of our sample, also likely increased the survey response rate. This underscores the necessity to engage community members in every step of research to produce relevant and meaningful outcomes.

Abbreviations

CDC

Centers for Disease Control and Prevention

US

United States

NC

North Carolina

PWUD

People Who Use Drugs

SSP

Syringe Services Program

LHD

Local Health Department

PORT

Post Overdose Response Team

OTP

Opioid Treatment Program

CBO

Community Based Organization

RA

Remedy Alliance

NCDHHS

North Carolina Department of Health and Human Services

EMS

Emergency Medical Services

MOUD

Medications for Opioid Use Disorder

LME-MCO

Local Management Entity/Managed Care Organization

Authors’ contributions

All authors (CS, GM, BO, DC, AI, RB, BF, MB) contributed to survey design and contact list development for survey distribution. CS, GM, BO participated in data collection and cleaning. CS, GM, BO conceptualized, drafted, and finalized the manuscript. DC obtained funding for and conceptualized the study. All authors offered revisions (CS, GM, BO, DC, AI, RB, BF, MB). The finalized manuscript was read and approved by all authors (CS, GM, BO, DC, AI, RB, BF, MB).

Funding

The 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 content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Data availability

The datasets generated during the current study are not publicly available due to confidentiality concerns, but an anonymized version of the data can be made available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was reviewed and determined Not Human Subjects Research by the Institutional Review Board at the University of North Carolina-Chapel Hill. Although the survey was determined Not Human Subjects Research, informed consent was nevertheless obtained through a statement of consent that preceded the survey for all participants. The study procedures adhered to the tenets of the Declaration of Helsinki.

Consent for publication

N/A.

Competing interests

The authors declare no competing interests.