Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Aug 18.
Published in final edited form as: J Subst Use Addict Treat. 2026 Jul 2;190:210063. doi: 10.1016/j.josat.2026.210063

Leveraging Barbershops for Community-Based Addiction Research: A Pilot Study Using Hair Samples to Detect Fentanyl and Stress Biomarkers

Nicholaus J Christian 1, Lisa A Farinelli 1, Meg Jordan 2, Shelley N Jackson 1, Lindsay A Kryszak 1, Garrick Sherman 1, Stephanie T Weiss 1, Stephen B Thomas 2,*, Lorenzo Leggio 1,3,4,5,6,*
PMCID: PMC13480290  NIHMSID: NIHMS2193786  PMID: 42392315

Abstract

Introduction.

The ongoing opioid overdose crisis disproportionately affects minority communities, highlighting the urgent need for community-based addiction research to better understand and address this public health issue. Community-academic partnerships are a key strategy to engage these communities in addiction research. Here we describe the Health Advocates In-Reach and Research (HAIR) initiative, a collaboration between the University of Maryland and the National Institute on Drug Abuse Intramural Research Program. We tested whether barbershops, a trusted community venue, can support hair collection for fentanyl toxicology and a stress biomarker (cortisol) to inform equitable overdose prevention.

Methods.

Within the HAIR initiative, community barbers/stylists and academic partners co-designed an anonymized, barbershop-based hair collection pilot project. Prespecified objectives were to (1) build trust between researchers and community members, (2) evaluate feasibility (successful collection and laboratory assay completion) and acceptability, (3) quantify fentanyl in hair, (4) quantify hair cortisol, and (5) explore associations between fentanyl detection and cortisol. Liquid chromatography–mass spectrometry (LC/MS) assays for quantitative analysis were developed and applied. Forty samples were collected from predominantly Black (85%), non-Hispanic (77.5%), male (95%) patrons.

Results.

Cortisol concentrations were successfully measured in 36 of the 40 samples, ranging from 3.36 to 104.00 pg/mg. Fentanyl was detected in 13 of the 40 samples, specifically four in quantifiable amounts (10.20 – 171.00 pg/mg) and nine samples containing trace amounts below the limit of quantification. An ordinary least squares regression model suggested that fentanyl presence may be predictive of increased cortisol after controlling for demographic variables. This association did not reach statistical significance (p = 0.065), likely due to limited statistical power from the small sample size.

Conclusions.

This pilot study presents a promising community-based translational research model to advance addiction implementation science. Barbershops represent a feasible, acceptable setting for community-based hair toxicology and stress biomarker collection, enabling the identification of otherwise unrecognized exposure and stress signals within high-burden populations. This practice-ready workflow could augment local overdose strategies, linking on-site findings to naloxone distribution and rapid Medications for Opioid Use Disorder (MOUD) navigation, and merits evaluation in larger hybrid implementation science studies.

Keywords: community-based addiction research, fentanyl detection, cortisol, hair toxicology, fentanyl toxicology, minority populations, opioid overdose crisis

1. Introduction

Provisional data from the Centers for Disease Control estimate that 107,543 individuals died from opioid overdose in 2023, marking a 3% decline in estimated deaths from 2022 (Centers for Disease Control and Prevention, 2024). While this is the first annual decrease in nearly a decade, overdose remains a leading cause of injury-related mortality in the United States with disproportionate impacts on minorities, particularly American Indian/Alaska Native and Black communities (Zhu, 2024). Overdose deaths among Black men have increased fourfold since 2015(Harris, 2023), and non-Hispanic Black men over the age of 65 now experience the highest overdose death rates (Kramarow & Tejada-Vera, 2022). Furthermore, Black individuals are significantly less likely to receive medications for the treatment of addiction (Hollander et al., 2021), underscoring the need to address health disparities in the context of substance use disorders.

Barbershops and beauty salons have long served as trusted institutions within the Black community. Historically, barbers and beauticians have played multifaceted roles, not only as entrepreneurs but as confidants, educators, advocates, and even social media influencers, providing culturally grounded support and services. These unique attributes make barbershops and salons highly effective platforms for delivering health promotion and disease prevention interventions (Khosla et al., 2024). Over the past decade, the University of Maryland (UMD) School of Public Health has advanced this potential through the Health Advocates In-Reach and Research (HAIR) Network initiative, which engages Black barbershops and beauty salons in biomedical research and public health education. The HAIR Network’s mission is to braid biomedical science, public health education, and medical services into trusted community spaces (Linnan et al., 2014; Maryland, 2024). Rooted in the principles of community-based participatory research, the model has proven effective in addressing critical health issues such as hypertension, diabetes, and vaccine hesitancy, underscoring its ability to build trust and address health issues within Black communities (Manns et al., 2024). For instance, during the COVID-19 pandemic, HAIR delivered vaccines in barbershops and salons across Prince George’s County, Montgomery County, and Baltimore City, and was recognized as a model for the nationwide “Shots at the Shop” campaign (Thomas et al., 2022). Building on this success, the UMD team and HAIR initiative partnered with the National Institute on Drug Abuse (NIDA) to advance addiction research and implementation science related to the opioid overdose crisis, primarily driven by the significant increase in fentanyl use (Volkow & Dye, 2025). The overarching framework of this approach emphasizes meeting individuals where they are, designing collaborative interventions, protecting privacy, providing timely benefits and conducting rigorous research that meaningfully includes those most affected.

Here we describe the UMD HAIR Network’s collaboration with NIDA - a partnership emphasizing early community engagement with barbers and beauticians playing a pivotal role as co-creators in research design. We then present findings from a pilot study co-developed with the community to explore the intersection of fentanyl use and stress and investigate the feasibility, acceptability, and implementation fidelity of collecting hair as a biospecimen in barbershop settings. Specifically, prespecified objectives were to (1) build trust between researchers and community members, (2) evaluate feasibility (successful collection and laboratory assay completion) and acceptability, (3) quantify fentanyl in hair, (4) quantify cortisol in hair, and (5) explore associations between fentanyl detection and cortisol. From a broader perspective, this study seeks to demonstrate how meaningful community engagement can drive addiction research and implementation science, yield actionable insights, and foster trust in populations historically underrepresented in biomedical research.

2. Methods

2.1. Partnership Development

In February 2022, amid the late-Omicron phase of the COVID-19 pandemic, the National Institute on Drug Abuse Intramural Research Program (NIDA IRP) and the University of Maryland (UMD) School of Public Health formalized a partnership to promote an equity-minded, community-based response to the overdose crisis. The COVID crisis had already broken down silos, with public health officials, clinical researchers, local businesses, and community leaders were collaborating on testing, vaccination, and trusted outreach in barbershops and salons. Building on that momentum, the agreement spelled out shared roles for education, trust-building, and evidence-based research focused on Baltimore/Washington D.C.’s historically underrepresented communities. Because UMD’s HAIR Network had earned credibility during the pandemic, UMD and NIDA clinical scientists entered shops not as outsiders but as co-designers, able to pilot novel methods (including on-site hair biospecimen collection) that would have been far harder to launch before COVID forced cross-sector innovation and accelerated trust-by-design practices. Please see Figure 1 for a study timeline of community engagement activities.

Figure 1. Study Activity Timeline.

Figure 1.

Open in a new tab

Outlines the implementation and community engagement activities of the project from the launch of the HAIR initiative at UMD in 2015 through the collection of hair samples and analysis of data in 2023.

2.2. Research Protocol Development

A silver lining of the pandemic was the greater use of virtual communication platforms, which opened our doors wider to two-way dialogue. The UMD and NIDA teams held virtual monthly meetings for 1.5 years to discuss shared research, clinical and public health interests, key public health issues in the local community, and ways to build synergistic collaborations. Furthermore, NIDA investigators were invited onto UMB-led HAIR Network’s live Zoom show, The Cutting Edge: All Things COVID-19 and Beyond, for recurring discussions. By showing up, taking questions from barbers, beauticians, and Wellness Warrior leaders, and returning week after week, the group built relationships and earned trust in full public view. Those sessions shaped the science: community members endorsed hair samples as a noninvasive, privacy-preserving option, and asked for an examination of the intersection of chronic stress (cortisol) and overdose risk (fentanyl), areas with little prior research in minority communities.

Discussions with community members focused on the benefits of using hair samples as a non-invasive method to assess substance exposure and stress. The decision to focus on measuring cortisol and fentanyl in hair samples to better understand the intersection of stress and substance use was supported by direct feedback from barbers/beauticians and community members through a series of Zoom meetings. Notably, there is a dearth of research investigating hair biomarkers in minority groups (Rippe et al., 2016). Hair as a biospecimen is unique in that it can detect the presence of metabolites for months post-exposure, although application is limited due to inability to detect recent use (takes 1–2 weeks post-exposure for detection) and increased costs compared to urine, blood or saliva testing (Palamar & Salomone, 2023). To address privacy concerns, foster trust and enhance recruitment, the study was designed to preserve participant anonymity with deidentified data collection. This pilot study received approval from the University of Maryland Institutional Review Board.

2.3. Participant Recruitment

We recruited adults (≥18 years) from a HAIR Network barbershop in Hyattsville, Maryland, selected a priori as a sentinel site based on its longstanding partnership with UMD and its neighborhood reach. Using convenience sampling and prespecified feasibility targets (≥80% consent; ≥90% successful collection; ≥85% laboratory sufficiency), trained research staff from UMD approached patrons during a scheduled “Pop-Up: Health is Wealth” event. To protect voluntariness and privacy, invitations were delivered by research staff, not barbers or stylists of record, consent occurred in a private area, participation did not affect services, and incentives were modest and uniform. Barbers and stylists completed a standardized 30-minute training on study talking points and referral procedures; fidelity was monitored via spot checks. Importantly, NIDA personnel were not involved in recruitment or hair-sample collection. Staff from NIDA and UMD attended Pop-Up Clinics in HAIR Network barbershops to offer screening and health counseling on cardiometabolic risk factors such as vital signs, blood glucose, and body mass index, as well as naloxone training. These activities were operationally distinct from research activities to avoid therapeutic misconception and served to strengthen the partnership by providing an opportunity to share early iterations of the study and demonstrate a commitment to transparency and inclusivity.

2.4. Human Hair Sample Collection, Field Processing, and Shipment to the NIDA Laboratory

Under a UMD IRB–approved protocol with a waiver of written consent, trained UMD research staff obtained documented verbal consent from adults who voluntarily agreed to participate and donate approximately 50–100 mg of hair during their visit at the barbershop/hair salon. No code key linking samples to identities was retained; the only participant information collected was age, sex, race, and ethnicity. Participants received $25 for time and inconvenience, and a $100 facility stipend compensated the barbershop or salon owner for space and time only; owners were not involved in identifying participants.

Hair collection followed a standardized, streamlined procedure. Barbers, hairdressers, stylists and beauticians received training from research staff on sample collection techniques to ensure consistency and, when involved, assisted only with cutting the hair. A small rubber band was placed at the scalp end of the sample, and hair was cut above the band from the posterior vertex. Stainless steel scissors were cleaned with 70% isopropanol between participants. Research staff immediately secured and marked the root end (band and foil notch), wrapped samples in aluminum foil, and placed them into paper envelopes with silica gel, sealed with tamper-evident tape, labeled with a barcode identifier, and logged onto a chain-of-custody form. Field blanks and 10% blind duplicates were included for quality assurance and quality control. Samples were stored dry in a desiccator at room temperature on the day of collection and shipped within 72 hours to the NIDA IRP Translational Analytical Core (TAC) laboratory. No field washing was performed; the laboratory applied a standard decontamination wash, and analytes were quantified using validated methods with predefined lower and upper limits of quantification (LLOQ/ULOQ; see below). This workflow minimized handling time, preserved sample integrity, and supported feasibility in a community-based setting while protecting participant privacy and voluntariness.

2.5. Hair Laboratory Analysis

Hair, composed primarily of α-keratin, is a protein filament that grows at an average rate of approximately 0.5 inches (1.27 cm) per month and can detect substance use biomarkers out to months or years depending on the length of the hair sample (Barone et al., 2024). Unlike serum or saliva, hair offers a more integrated measure of long-term exposure, reflecting biomarker accumulations over weeks or months rather than the short-term fluctuations characteristic of diurnal rhythms (Sauvé et al., 2007). The feasibility of collecting hair samples and running analyses in ethnically diverse populations has been described in few studies (Michaud et al., 2024; Vargas et al., 2023; Wosu et al., 2015).

The NIDA IRP TAC laboratory conducted analytical assays for the quantification of cortisol, fentanyl and norfentanyl hair concentrations. Ingested fentanyl is metabolized in the body to norfentanyl; therefore, the presence and concentrations of norfentanyl can help differentiate actual fentanyl ingestion from environmental contamination of hair samples. Specifically, a person whose hair was contaminated by environmental fentanyl would only test positive for the fentanyl parent compound and not for the norfentanyl metabolite. The laboratory testing methodologies were designed to maximize the efficiency of biomarker recovery from hair while ensuring high analytical accuracy and reproducibility. All samples were analyzed prior to July 2023.

Cortisol Laboratory Protocol.

Hair samples were finely chopped and shaken with the solvent, methanol (MeOH) at 55°C for 16 hours to extract cortisol effectively from the keratin matrix. After incubation, the resulting supernatant was evaporated, and the resulting residue was reconstituted in enzyme-linked immunosorbent assay (ELISA) buffer for analysis. The commercially available Salimetrics cortisol ELISA kit was employed to quantify cortisol concentrations. To enhance the reliability and accuracy of the assay, three matrix control standards were included, ensuring the robustness of cortisol measurement across samples. The cortisol ELISA buffer contains a pH indicator that changes color under acidic conditions. A subset of samples exhibited a yellow discoloration upon buffer addition, consistent with low pH. These samples yielded cortisol concentrations that were implausibly elevated and not biologically consistent. Therefore, they were excluded from analysis and not reported. Although the specific source of interference was not identified, we hypothesize that exogenous compounds (e.g., acidic constituents of hair or skin styling products) may have contributed to the observed pH alteration.

Fentanyl and Norfentanyl Laboratory Protocol.

Fentanyl and norfentanyl analyses were conducted using a liquid-liquid extraction method optimized for high sensitivity and specificity. Hair samples were first subjected to digestion with 1N sodium hydroxide (NaOH) at 80°C while agitated at 500 RPM for one hour in the presence of deuterated internal standards to facilitate the release of fentanyl embedded in the hair matrix. Following digestion, a liquid-liquid extraction was performed using a 9:1 mixture of hexane and ethyl acetate. The organic layer was subsequently evaporated and reconstituted in a solvent compatible with liquid chromatography-mass spectrometry (LC-MS) analysis. This method allowed for precise quantification of fentanyl concentrations and presence of trace concentrations in the hair samples.

2.6. Data Analysis

The data were collected and stored in an encrypted Microsoft Excel spreadsheet. Descriptive statistics were used to summarize the age, sex, race, and ethnicity of participants as well as to record cortisol, fentanyl, and norfentanyl quantification. Fentanyl positivity was coded as a binary variable (negative or positive) to assess its association with cortisol concentrations. Because prior research has established a negative correlation between age and mean cortisol concentrations (Sauvé et al., 2007), bivariate correlation between age and cortisol concentration was assessed. An ordinary least squares (OLS) regression model was estimated to assess the relationship between fentanyl and cortisol concentrations, controlling for age, race, and ethnicity. Sex was excluded as a covariate due to the extremely low number of female participants.

3. Results

We collected 40 hair samples from one barbershop and one hair salon; 38 (95%) were from men. Participants were predominantly Black (85%), and 20% identified as Hispanic (race and ethnicity not mutually exclusive). The mean age was 41.3 years (SD=14.7, range=18–79).

A total of 13 samples (32.5%) were positive for fentanyl. Specifically, 4 out of 40 samples (10%) had quantifiable concentrations of fentanyl, with a range of detection between 10.2–171.0 pg/mg. Furthermore, 9 of 40 samples (22.5%) contained trace amounts of fentanyl below the limit of quantification (LOQ=2 pg/mg) but above background noise. All individuals with quantifiable amounts of fentanyl were 35 years of age or younger. Norfentanyl was detected in two samples, both of which also had high concentrations of fentanyl detected. Figure 2 portrays the results of fentanyl testing.

Figure 2. Fentanyl Test Results.

Figure 2.

Open in a new tab

Fentanyl was detected in 13/40 samples (32.5%), including 4/40 (10.0%) with quantifiable concentrations (10.2–171.0 pg/mg); norfentanyl was detected in two samples, both with quantifiable fentanyl. The unit of measurement (pg/mg) indicates picograms of analyte per milligram of hair sample.

Cortisol was successfully measured in 36 of 40 samples, ranging from 3.36–104.00 pg/mg. Cortisol concentration in one sample was below the LOQ, and three samples were not compatible with the assay due to the presence of interfering compounds not eliminated by the laboratory decontamination wash. The average cortisol concentration among all participants was 21.00 (SD=20.93). Age was not significantly correlated with cortisol in our sample (r = −.066, p = .702). Cortisol concentrations grouped by age are shown in Figure 3.

Figure 3. Cortisol Test Results by Age.

Figure 3.

Open in a new tab

Cortisol concentrations (pg/mg; log-transformed y-axis) did not differ significantly by age group (F(4, 31) = 0.128, p = 0.971). Cortisol was quantified in 36/40 samples (range, 3.36–104.00; mean [SD], 21.00 [20.93]). One sample was BLQ and three were assay incompatible due to interfering compounds. BLQ: below the limit of quantification; SD: standard deviation.

An OLS linear regression model was estimated predicting cortisol concentrations based on fentanyl presence with covariates including age, race, and ethnicity. Cortisol values were tested for normality and found to be significantly right-skewed (Shapiro-Wilk W = .721, p < .001); we therefore natural log-transformed the outcome before analysis and report back-transformed effects. This model yielded a significant effect of fentanyl (β = 2.451, 95% CI [1.31, 4.60], p = .007), indicating an almost 2.5 pg/mg increase in cortisol concentration when fentanyl was present. Although a trend was seen, the model was not significant overall (F(4, 28) = 2.636, p = .055). Table 1 reports the individual data of fentanyl, norfentanyl, and cortisol results in the hair samples, and Figure 4 demonstrates the cortisol concentrations by presence of fentanyl.

Table 1.

Individual results of the fentanyl, norfentanyl, and cortisol hair samples

Sample Fentanyl (pg/mg) Norfentanyl (pg/mg) Cortisol (pg/mg) Cortisol CV (%)
1 ND ND 29.1 1.02
2 ND ND 57.0 2.66
3 ND ND No result
4 ND ND 24.0 1.02
5 ND ND 8.05 18.6
6 ND ND 32.0 6.42
7 ND ND 6.40 12.0
8 Trace ND 14.6 1.66
9 Trace ND 15.6 1.88
10 Trace ND 20.4 2.57
11 Trace ND 18.4 1.57
12 Trace ND 10.6 13.2
13 ND ND 13.0 9.06
14 45.3 5.51 31.3 6.92
15 Trace ND 77.1 2.40
16 ND ND 4.46 7.10
17 10.2 ND 8.49 2.06
18 ND ND 3.95 7.13
19 ND ND 12.7 7.75
20 ND ND No result
21 171 8.41 38.1 6.34
22 ND ND 6.98 6.84
23 ND ND 8.78 2.99
24 ND ND 15.0 9.24
25 ND ND 15.1 1.09
26 ND ND 34.8 7.78
27 ND ND 27.8 1.78
28 15.0 ND 4.00 12.5
29 ND ND 16.6 2.13
30 Trace ND 104 0.80
31 ND ND 5.76 1.70
32 ND ND 26.3 9.63
33 ND ND 3.63 5.34
34 ND ND No result
35 ND ND 13.1 15.8
36 Trace ND 21.1 7.88
37 ND ND 11.9 0.58
38 ND ND 9.85 9.88
39 Trace ND BLQ
40 ND ND 5.67 5.66
Open in a new tab

For fentanyl/norfentanyl: ND, not detected; Trace, detected below the lower limit of quantification, but above background noise

For cortisol: BLQ, below the limit of quantification; No result, assay incompatible.

Figure 4. Cortisol Test Results by Fentanyl Presence.

Figure 4.

Open in a new tab

Cortisol concentrations (pg/mg; log-transformed y-axis) exhibited a statistically significant association with fentanyl presence (Detected = trace or quantified, None = ND; p = .007). However, a test of overall model fit was only significant at a trend level (F(4, 28) = 2.636, p = .055). BLQ: below the limit of quantification; ND: not detected.

4. Discussion

This study demonstrates how community-academic-government partnerships can be imbedded within community spaces such as Black barbershops and salons to build trust and enhance inclusion of this population in addiction research and implementation science. The pilot study, developed in partnership with community barbers and hairdressers, analyzed de-identified hair samples collected from a barbershop and hair salon in the Baltimore/Washington D.C. region, and found a high rate of fentanyl detection with 32.5% testing positive or trace positive.

According to the National Survey on Drug Use and Health (NSDUH), 24.9% of people endorsed any illicit drug use, and 3.1% of people endorsed misusing opioids in 2023 (Substance Abuse and Mental Health Services Administration, 2024). Consistent with NSDUH, the majority of fentanyl-positive cases in our sample were in individuals younger than 35 years of age. Norfentanyl was detected in only two samples. The low number of samples with norfentanyl detected is most likely attributable to norfentanyl concentrations that are below the limit of laboratory detection as found in another study (Salomone et al., 2020), although it could also result from external contamination of the hair rather than ingestion, potentially through direct contact or unintentional or secondhand exposure to fentanyl. Given that fentanyl is widespread in the drug supply, including in stimulants and counterfeit pressed pills, it is possible that individuals testing positive for fentanyl in our study may have been exposed to lower amounts of fentanyl than previous studies, thus decreasing our detection of norfentanyl.

The median hair cortisol concentration in this study was 14.8, which is much lower than the median of 46.1 pg/mg found in a previous study (Sauvé et al., 2007). Previous studies have noted that labs frequently use different hair cortisol assay techniques and thus obtain different yet correlated values (Russell et al., 2015). Similar to previous studies, increased cortisol was found in younger individuals. Although not definitive, the linear regression model trended toward a positive correlation between cortisol and fentanyl after accounting for age, race, and ethnicity, which suggests that the enrolled cohort may have been underpowered to detect a true effect due to the small sample size. In addition, previous studies have demonstrated that acute or chronic opioid exposure can reduce or increase cortisol concentrations in individuals depending on the context (Abs et al., 2000; Oltmanns et al., 2005; Palm et al., 1997; Zis et al., 1984). It is important to note that our results don’t establish causality, and it is also plausible that increased stress may lead to increased opioid use, suggesting the possibility of bidirectional effects between cortisol concentrations and opioid use.

One key finding from this study is the feasibility of integrating hair biospecimen collection into a culturally relevant, community-based research model. Given that overdoses with stimulants and opioids disproportionately affect men (Butelman et al., 2023), barbershops may represent an ideal setting for implementing overdose prevention interventions. For instance, given that hair can provide detection of use throughout the lifespan of the hair follicle, often 90 days or longer, hair testing could be a tool to detect population-level trends in substance use and/or stress over time. However, given that hair does not detect recent use and is relatively costly, it is unlikely to surpass urine, blood, or saliva in terms of individual-level clinical utility.

By leveraging hair as a biospecimen, this study demonstrates how community-based research can incorporate innovative, non-invasive methodologies among historically underrepresented patient populations. While the COVID-19 pandemic brought unprecedented challenges, it also served as a catalyst for collaborations between basic and social science researchers, prompting them to move beyond disciplinary silos and engage more directly with communities. This initiative exemplifies how such collaborations have extended beyond the COVID public health crisis, evolving into sustained partnerships that foster integrated community-academic-government engagement for other community health initiatives. These efforts strengthen trust in biomedical research and address chronic conditions such as substance use disorders, which disproportionately affect minority populations and are associated with high morbidity and mortality.

Additionally, this study provides a framework for how academic and governmental institutions can invest in community research and development. Through this effort, barbers and hairdressers were introduced to the research process, increasing awareness of their roles as trusted educators and health messengers within their communities. The study also actively engaged community members in participant recruitment, study participation, and educational outreach. This approach empowered trusted community stakeholders to act as bridges between biomedical researchers and participants, fostering a shared sense of ownership in the project. By integrating community leadership into the research process, the study generated valuable scientific insights while advancing a model for equitable and culturally responsive addiction research. The “Health is Wealth” event supported early recruitment by blending cultural relevance, personal connection, and celebration to deepen community engagement.

To honor the community’s generous contribution to the research, dissemination of findings and educational outreach extended beyond the completion of study enrollment. This ongoing exchange underscores a bidirectional model of reciprocity between biomedical researchers and community members. For example, participating barbers and hairdressers were invited to tour the NIH facilities located at the NIDA IRP (Bayview campus) in Baltimore, MD, where they observed how the hair samples were received, processed and analyzed. A customized postcard commemorating this event was produced and distributed to all patrons of the participating barbershop and hair salon serving as a “thank you” to the community for their contribution to advancing addiction science. These events went beyond simply reporting results - they facilitated dialogue, deepened understanding, built trust, and reaffirmed the community’s role as an equal partner in the research process. This transparent and pioneering approach strengthened trust and laid the foundation for future innovative collaborations that extend well beyond the scope of this initial pilot project.

Limitations of this study include that findings may not be generalizable to the general public given that the sample was predominantly black and male. Additionally, the study was underpowered to conclusively determine whether a relationship exists between fentanyl and cortisol concentrations. Working with hair as a biospecimen proved relatively simple in collection, although it introduced unique technical processing challenges during laboratory analysis. Additionally, addressing community concerns and misconceptions about the use of biospecimens required careful, ongoing communication to maintain stakeholder investment. These challenges highlight the importance of designing research methodologies that are not only scientifically robust, but also culturally and contextually appropriate.

Future research could build upon this study by including larger and more diverse populations to further examine the relationship among substance use, stress, and health outcomes related to chronic diseases such as substance use disorders. Including self-reported substance use and stress levels would strengthen the data and provide a clearer understanding of the role of hair toxicology testing in the community. Moreover, scaling up this approach could provide a broader framework for integrating community-based participatory research into addiction research and implementation science and other public health domains. The sustained engagement of researchers with the community in the post-pandemic era models a feasible method to enhance minority representation within biomedical research.

Conclusion

In this pilot feasibility study, we found that hair samples collected from barbershops demonstrated higher rates of fentanyl exposure compared to that found in national samples (Substance Abuse and Mental Health Services Administration, 2024). Concentrations of cortisol and fentanyl appeared to be positively correlated, although the sample size was likely too small to establish a statistically significant relationship.

This pandemic-era pilot was a feasibility and acceptability test, not a prevalence estimate or causal study. Given the single-site setting, small sample, and the possibility that human hair can reflect inadvertent ingestion or environmental contamination, these findings should be read as proof-of-practice. The larger lesson is methodological. The HAIR Network converts everyday businesses into standing, privacy-preserving access nodes for addiction research, offering evidence that trust is a bridge to science. Community-academic-government partnerships are a promising method to engage communities in addiction research. For generations, barbershops and hair salons have been trusted spaces within Black communities. Their welcoming, familiar atmosphere makes them ideal spaces to engage community members in research on stress and substance use - key factors underlying substance use disorders and other chronic diseases that disproportionately affect minority populations.

  1. 40 hair samples were collected from predominantly Black, non-Hispanic, males

  2. Fentanyl was detected in 13 of the 40 samples

  3. Cortisol levels were successfully measured in 36/40 samples

  4. Barbershops can be leveraged for community-based addiction research

Acknowledgements:

The authors would like to thank all the study participants who made this work possible.

The authors would also like to thank the several colleagues and staff who provided input, suggestions and/or technical or administrative support during the execution of this project, especially: (1) From the NIDA IRP: Christine Brannock, Dr. Brenda Curtis, Max Dennis, Susan Eldridge, Menbere Haile, and (2) From UMD: Mike (DynaMike) Brown, Maggie Daly, Surayyah Khan, Dr. Robin Puett. The authors would also like to thank Dr. Dushanka Kleinman (UMD) for her support on this project and to the broader partnership between UMD and the NIDA IRP.

The authors would like to thank Ms. Lauren Brick (Visual Media, NIDA IRP) for graphical design, Dr. Teri Franklin for editing and bibliographic assistance, and Dr. Alberto Salomone (University of Turin, Italy) for providing fentanyl-positive hair samples for use in assay development.

Funding/sources of support:

This work was supported by (1) the University of Maryland (UMD) and (2) the National Institutes of Health Intramural Research Program (NIH IRP) via the National Institute on Drug Abuse (NIDA) IRP, specifically the joint NIDA/NIAAA Clinical Psychoneuroendocrinology and Neuropsychopharmacology Section (ZIA-DA000635), the NIDA IRP Office of the Clinical Director (ZIA-DA000535) and the NIDA IRP Translational Analytical Core. The contributions of the NIH authors are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Abs R, Verhelst J, Maeyaert J, Van Buyten JP, Opsomer F, Adriaensen H,…Van Acker K. (2000). Endocrine consequences of long-term intrathecal administration of opioids. J Clin Endocrinol Metab, 85(6), 2215–2222. 10.1210/jcem.85.6.6615 [DOI] [PubMed] [Google Scholar]
  2. Barone R, Pelletti G, Giorgetti A, Mohamed S, Pascali JP, Sablone S,…Pelotti S. (2024). Validation and application of a method for the quantification of 137 drugs of abuse and new psychoactive substances in hair. Journal of Pharmaceutical and Biomedical Analysis, 243, 116054. 10.1016/j.jpba.2024.116054 [DOI] [PubMed] [Google Scholar]
  3. Butelman ER, Huang Y, Epstein DH, Shaham Y, Goldstein RZ, Volkow ND, & Alia-Klein N (2023). Overdose mortality rates for opioids and stimulant drugs are substantially higher in men than in women: state-level analysis. Neuropsychopharmacology, 48(11), 1639–1647. 10.1038/s41386-023-01601-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Centers for Disease Control and Prevention, N. C. f. H. S. (2024). U.S. Overdose Deaths Decrease in 2023, First Time Since 2018. Retrieved Nov 5 2024 from https://www.cdc.gov/nchs/pressroom/nchs_press_releases/2024/20240515.htm [Google Scholar]
  5. Harris RA (2023). Drug Overdose Deaths Among Non-Hispanic Black Men in the U.S.: Age-Specific Projections Through 2025. AJPM Focus, 2(1). 10.1016/j.focus.2022.100063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Hollander MAG, Chang CH, Douaihy AB, Hulsey E, & Donohue JM (2021). Racial inequity in medication treatment for opioid use disorder: Exploring potential facilitators and barriers to use. Drug Alcohol Depend, 227, 108927. 10.1016/j.drugalcdep.2021.108927 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Khosla P, Amin K, & Doshi R (2024). Combating Chronic Disease with Barbershop Health Interventions: A Review of Current Knowledge and Potential for Big Data. Yale J Biol Med, 97(2), 239–245. 10.59249/otfp5065 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Kramarow EA, & Tejada-Vera B (2022). Drug Overdose Deaths in Adults Aged 65 and Over: United States, 2000–2020. NCHS Data Brief(455), 1–8. [PubMed] [Google Scholar]
  9. Linnan LA, D'Angelo H, & Harrington CB (2014). A literature synthesis of health promotion research in salons and barbershops. Am J Prev Med, 47(1), 77–85. 10.1016/j.amepre.2014.02.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Manns BJ, Thomas S, Farinu O, Woolfork M, & Walker CL (2024). Hyperlocal lessons from the COVID-19 pandemic: Toward an equity-centered implementation science approach. Soc Sci Humanit Open, 9. 10.1016/j.ssaho.2024.100844 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Maryland, U. o. (2024). The Health Advocates In-Reach and Research Campaign (HAIR): Maryland Center for Health Equity. Retrieved Nov 5 2024 from https://sph.umd.edu/hair [Google Scholar]
  12. Michaud M, Evans M, Mendez R, Zapanta J, Trochez A, Mehta KM,…Parangan-Smith A. (2024). Investigating the Impacts of a Modified Mindfulness Practice on Minoritized College Students' Chronic Stress. Integr Med Rep, 3(1), 102–110. 10.1089/imr.2024.0009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Oltmanns KM, Fehm HL, & Peters A (2005). Chronic fentanyl application induces adrenocortical insufficiency. J Intern Med, 257(5), 478–480. 10.1111/j.1365-2796.2005.01483.x [DOI] [PubMed] [Google Scholar]
  14. Palamar JJ, & Salomone A (2023). On the challenges of hair testing to detect underreported substance use in research settings. Am J Drug Alcohol Abuse, 49(1), 1–4. 10.1080/00952990.2023.2166414 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Palm S, Moenig H, & Maier C (1997). Effects of oral treatment with sustained release morphine tablets on hypothalamic-pituitary-adrenal axis. Methods Find Exp Clin Pharmacol, 19(4), 269–273. [PubMed] [Google Scholar]
  16. Rippe RCA, Noppe G, Windhorst DA, Tiemeier H, van Rossum EFC, Jaddoe VWV,…van den Akker ELT. (2016). Splitting hair for cortisol? Associations of socioeconomic status, ethnicity, hair color, gender and other child characteristics with hair cortisol and cortisone. Psychoneuroendocrinology, 66, 56–64. 10.1016/j.psyneuen.2015.12.016 [DOI] [PubMed] [Google Scholar]
  17. Russell E, Kirschbaum C, Laudenslager ML, Stalder T, de Rijke Y, van Rossum EF,…Koren G. (2015). Toward standardization of hair cortisol measurement: results of the first international interlaboratory round robin. Ther Drug Monit, 37(1), 71–75. 10.1097/FTD.0000000000000148 [DOI] [PubMed] [Google Scholar]
  18. Salomone A, Bigiarini R, Palamar JJ, McKnight C, Vinsick L, Amante E,…Vincenti M. (2020). Toward the Interpretation of Positive Testing for Fentanyl and Its Analogs in Real Hair Samples: Preliminary Considerations. J Anal Toxicol, 44(4), 362–369. 10.1093/jat/bkz102 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Sauvé B, Koren G, Walsh G, Tokmakejian S, & Van Uum SH (2007). Measurement of cortisol in human hair as a biomarker of systemic exposure. Clin Invest Med, 30(5), E183–191. 10.25011/cim.v30i5.2894 [DOI] [PubMed] [Google Scholar]
  20. Substance Abuse and Mental Health Services Administration. (2024). 2023 Companion infographic report: Results from the 2021, 2022, and 2023 National Surveys on Drug Use and Health (SAMHSA Publication No. PEP24-07-020). https://www.samhsa.gov/data/report/2021-2022-2023-nsduh-infographic [Google Scholar]
  21. Thomas SB, Quinn SC, & Jordan M (2022). Strengthening Health Promotion Through Sustained Hyperlocal Community Engagement. [Google Scholar]
  22. Vargas SM, Parra LA, Rivas WA, Payat S, Mistry R, Williams CR,…Miranda J. (2023). Recruitment and Feasibility of Hair Cortisol Collection in a Sample of Ethnically and Sexually Diverse, Low-Income Adults: A Qualitative Community-Partnered Participatory Research Study. J Health Care Poor Underserved, 34(1), 74–101. 10.1353/hpu.2023.0006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Volkow N, & Dye LR (2025). Groundbreaking Research from NIDA Addressing the Challenges of the Opioid Epidemic. J Med Toxicol, 21(1), 69–77. 10.1007/s13181-024-01041-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Wosu AC, Gelaye B, Valdimarsdottir U, Kirschbaum C, Stalder T, Shields AE, & Williams MA (2015). Hair cortisol in relation to sociodemographic and lifestyle characteristics in a multiethnic US sample. Ann Epidemiol, 25(2), 90–95, 95 e91–92. 10.1016/j.annepidem.2014.11.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Zhu DT (2024). Racial and Ethnic Disparities in Fentanyl and Polysubstance Overdose Deaths. JAMA Psychiatry. 10.1001/jamapsychiatry.2024.3435 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Zis AP, Haskett RF, Albala AA, & Carroll BJ (1984). Morphine inhibits cortisol and stimulates prolactin secretion in man. Psychoneuroendocrinology, 9(4), 423–427. 10.1016/0306-4530(84)90050-7 [DOI] [PubMed] [Google Scholar]

RESOURCES