‘A Way to Know’ Your Stress: Acceptability and Meaning of Hair Sample Collection for Cortisol Analysis From African-American and Latina Women in a Digital Storytelling Study

Abstract

Obtaining biological samples for research purposes from members of marginalized communities demands careful social and ethical consideration. This paper aims to discuss how feasibility and acceptability translate into methodological and ethical considerations for collecting hair samples in a study comprised of young, pregnant Latinx and African-American women participating in a digital storytelling intervention (DST). Transcripts from two focus groups with Healthy Families home visitor staff (N =10) and follow-up interviews with DST participants (N = 8) were analyzed through qualitative content analysis for hair sampling acceptability. Responses regarding the feasibility of obtaining a hair sample were generally positive amongst Healthy Families home visitor staff participating in the two focus groups, though responses overall were more favorable with the pregnant women participants who were interviewed post-DST intervention. Home visitors emphasized clear communication to participants around the reasoning for obtaining biological samples. Social and ethical considerations for communication of biomarker data to participants, as well as the intersection of lay knowledge of stress and participants' perspectives on providing a hair sample for cortisol analysis, is discussed.

Keywords

community health workers focus groups biomarker postpartum depression Latina African-American

Introduction

Obtaining biological data from research participants belonging to marginalized ethnic minorities yields a vast body of ethical questions and recruitment barriers.¹ These challenges originate from a legacy of participant mistrust due to exploitation of marginalized persons within the medical system.² One of the most prominent cases in the contemporary discourse on the bioethics of human biological samples collection is that of Henrietta Lacks, an African-American woman whose cells obtained from cancerous cervical tissue would go on to be the first immortalized cell line.³ Occurring before the advent of Institutional Review Boards and human subjects protections, Lacks received no compensation, nor provided informed consent for her tissue to be obtained and used for research purposes. Despite the necessity of informed consent today, Henrietta Lacks is a reminder for investigators that participant skepticism and mistrust in institutions doing research has profound historical roots, especially for individuals belonging to marginalized groups or identities. These considerations are especially important in community-based research, where investigators endeavor to establish trust with the community, but may still face barriers or resistance to recruiting members of marginalized communities if and when biological data collection is a part of the research protocol. Community-based participatory research (CBPR) provides a framework for addressing these barriers and allowing investigators to build a cohesive foundation of trust with participants belonging to marginalized communities.⁴

These considerations guided the early planning of a mixed-methods, community-based pilot study that examined the feasibility and acceptability of using a digital storytelling (DST) intervention to decrease depressive symptoms in young pregnant Latina and African-American women in the Springfield, MA metropolitan area. The study also incorporated psychosocial and biological measures of stress, notably hair samples to analyse cortisol (Table 1), to preliminarily assess the effect of the intervention on study participants.

Table 1.

Study Timeline, With Hair Sample Collection Times Highlighted in Bold.

Time point	Purpose	Measures used/obtained at TP
1	Recruitment/screening	Screening questionnaire
2	DST workshop	Consent form Surveys Demographics questionnaire Adverse Childhood Experience (ACE) Questionnaire Edinburgh Postnatal Depression Scale (EPDS) Perceived Stress Scale (PSS) Generalized Anxiety Disorder 7-item Scale (GAD-7) Self-Esteem Questionnaire General Self-Efficiency Scale (GSE) Perceived Social Support Relationship Characteristics Questionnaire Hair sample
3	Third-trimester data collection	Same surveys as used in T2 Semi-structured interviews Hair sample
4	Post-delivery data collection	Short survey on pregnancy outcomes
5	3-month postpartum data collection	Same surveys as used in T2 Breastfeeding outcome Maternal-Infant Bonding Scale 3-Month Hair sample
6	6-month postpartum data collection	Same surveys as used in T2 Breastfeeding outcome Maternal-Infant Bonding Scale Hair sample

DST is a community-based participatory research (CBPR) method that prioritizes the tradition of oral storytelling,^5,6 and is shown to increase self-efficacy and self-esteem and decrease anxiety.^7,8 In a DST intervention, each participant produces a first-person narrative approximately 1-3 minutes long, addressing a theme or important moment in the participant’s life. Planning and production take place over a three-day workshop with other program participants, where feedback is encouraged and a peer-support environment is created. DST interventions are also shown to increase peer support in pregnant and parenting young women.⁹

While psychometric instruments are frequently used in health-based participatory methodologies, including in DST interventions, results are often inconsistent and point to the need to more effectively integrate these instruments with rich field research and rigorous data collection.¹⁰ While self-report data might involve challenges related to bias or cultural legibility,¹¹ the collection of biological data must contend with the historical legacy discussed earlier in tandem with the physicality of obtaining a biological sample. In our study, these challenges are further confounded by the cultural values around women’s hair, including religious and spiritual proscriptions about the cutting of hair,¹² as well as the potentially coercive use of hair samples in state surveillance to determine substance use in Latina and African-American communities.¹³

Despite these challenges, hair samples provide several benefits over the traditional biological sampling methods for cortisol. Cortisol is a diurnal steroid hormone implicated in HPA-Axis regulation, a key physiological pathway in stress physiology. Cortisol can be analysed from several biological matrixes, including blood serum, saliva, and hair, making it a readily accessible biospecimen for researchers to examine. While serum and salivary measures of cortisol have been well-validated and frequently used in clinical and biomedical research, barriers exist to these sampling measures that are often especially challenging in community-based maternal health studies. For both serum and saliva, a fasting (12-hour) measure is most ideal, which can be especially challenging with pregnant women. Blood draws, which are necessary for obtaining plasma serum, are invasive and require special training that is often cost-prohibitive and inaccessible in community-based projects operating outside of a hospital setting.¹⁴

The ease of obtaining a hair sample could make it a compelling, highly accessible method for community-based health researchers who wish to characterize and identify chronic stress in research participants. To obtain a hair sample, a small portion of hair—approximately a pencil thickness’ diameter of hair strands—is cut just below the crown of the head, as close to the scalp as possible. In contrast to saliva collection, which requires several samples taken over the course of a day, a single hair sample provides an effective measure of the peripheral cortisol accumulated in a person’s system over the span of three months’ time.^15,16 Thus, cortisol measures from hair samples can illustrate whether an individual is in a physiological state of chronic stress, with less biological material than would be needed to obtain the same picture via serum or saliva.¹⁷

The purpose of this article is to (1) discuss our preliminary findings related to the feasibility and acceptability of hair sampling for cortisol analysis in a community-based study involving young pregnant Latina and African-American women, and (2) provide recommendations for community-based researchers wishing to integrate methods of biological data collection into their studies. For investigators who are more familiar with qualitative research methods, and are literate in the ethical and historical contexts of biological sample collection, collecting this sort of data from participants may seem daunting. This is especially challenging when existing literature does little to characterize the acceptability of certain biological sample collection methods, like hair sample collection, from Latina and African-American women. Acceptability is integral to ethical, community-based biological sample collection, as it ensures that biological samples are being given on the terms of the participant, and not solely to benefit researchers’ questions about marginalized groups or drive forward their own research. This praxis is key to devising community-based, mixed-methods projects that see community members as more than passive research participants.

Methods

Data for this study were obtained from a study entitled “Digital Storytelling for Postpartum Depression Prevention Pilot Study,” which sought to assess the feasibility of using a DST intervention with young (ages 15–24) Latina and African-American women with a singleton pregnancy living in and around the Springfield, MA metropolitan area. DST consists of participation in a 3–day workshop, where participants participate in talking and writing activities with fellow group members to craft a story script that serves as the basis for their digital story. After finalizing the stories, audio-recording their script as a voiceover, and choosing images to align with the voiceover, participants were led through a digital editing tutorial to learn how to piece their digital stories together. The workshop culminates with a digital story screening, where each participant’s story is viewed with discussion following. Each participant receives a copy of their digital story to keep and share.

Transcripts resulting from two focus groups held with Healthy Families Home Visitors (N = 5 participants per group, 10 total)—conducted well ahead of the DST workshop, findings from which were used to inform the larger study design—and from follow-up interviews held with DST intervention participants (N = 8 participants) were analysed using qualitative content analysis. Healthy Families Massachusetts is a free and voluntary support program for first-time parents across the state. Home visitors are family-support staff who provide personalized services to new families (www.healthyfamiliesma.org). In focus groups, home visitors were asked about the feasibility and acceptability of collecting hair samples from young women similar to those they serve in the community.

DST participant interviews were semi-structured, occurring after the intervention, and ranged in length from 35-80 minutes. Participants were asked general questions about the hair sample collection process, as well as their perceptions on the feasibility of utilizing hair-sampling methods.

Study instruments collected in the study are listed in Table 1 along with hair sample time points. Corresponding means and proportions shown in Table 2 for the purpose of characterizing study participants. Instruments were managed using Research Electronic Data Capture (REDCap) tools hosted free of charge at Tufts University. REDCap is a secure, web-based application designed to support data capture for research studies. IRB approval to conduct this study was obtained for all activities at all affiliated institutions.

Table 2.

Characteristic and Descriptive Data From Participants.

Data	Mean (SD) or proportion (N%)
Age	21 (2)
Race/ethnicity
African-American	2 (25%)
Latinx/Hispanic	6 (75%)
Highest education level
Less than HS	1 (12.5%)
High school or GED	5 (63%)
Trade school	1 (12.5%)
2-Year college	1 (12.5%)
Frequency of mental health care
Weekly	2 (25%)
Less than monthly	5 (62.5%)
>2 ACEs reported	5 (62%)
<2 ACEs reported	3 (38%)
EPDS score	15^a (3)
PSS score	22^b (4)
GAD-7 scale	12 (5)

^aAt risk for Mild or Major Depression is defined as a score ≥10 on the EPDS.

^bScores ranging from 14-26 are considered moderate stress on the PSS.

Results

“‘Who’s Gonna Get That, What Are They Gonna Do With That?’” Home Visitor Perspectives on Hair Sample Collection

Conducting a focus group with home visitors was an important first step in assessing the feasibility of obtaining hair samples. A total of 10 home visitors participated in the focus groups. Because home visitors offer the perspective of extensive first-hand engagement with members of the study community, we believed that they offered unique insight into the study community and how they might feel about this form of data collection. Generally, home visitor responses related to the feasibility of obtaining hair samples were encouraging. They emphasized the need for transparency, suggesting that the research team should provide clear information about the purpose for hair sample collection. One home visitor’s remark—“Who’s ‘gonna get that. What are they gonna do with that?”—in response to our feasibility question was illuminating and signalled the need to clearly explicate to participants the purpose of the hair sample collection. Several home visitors acknowledged possible ethno-cultural and spiritual-based barriers to hair sampling, as indicated in the following dialogue between two home visitors:

Home Visitor 1: A lot of Hispanic people might be uncomfortable with the hair because of culture-wise, they’ll think witchcraft. They’re gonna—

Home Visitor 2: “No one is getting a piece of my hair.”

Home Visitor 1: And it’s not only a moral of their culture. Their religion—

Home Visitor 2: Yeah, can’t cut their hair.

Two cultural groups—women practicing the Pentecostal faith or who were practitioners or believers in Brujería (witchcraft)—were acknowledged by the home visitors as those who might prefer providing a saliva rather than hair sample for cortisol measures. For Pentecostal adherents, this may relate to religious proscriptions against hair cutting. For those with cultural knowledge or practice in Brujería, this likely relates to the potential use of hair samples for mal-intended purposes. Instead of the structural tensions our group was principally concerned with, such as the concern about perceived surveillance given the similarity between providing a hair sample for cortisol analysis and providing a hair sample for drug testing, the home visitors indicated that the cultural value given to hair by different ethnic groups would be a firmer boundary for potential participants.

“I Do Want to Know My Stress Levels, Because I Have Stress Every Day”: Participant Perspectives on Hair Sample Collection

A total of eight (N = 8) women participated in the DST Intervention in July 2018. Table 2 presents descriptive data for our participants, obtained at the start of the workshop. To be eligible for study participation, women were required to score at a 10 or above on the EPDS. The scale is commonly used both clinically and for research purposes and has good sensitivity and specificity for detecting perinatal depression,¹⁸ including in young women.¹⁹ For the purpose of participant characterization, mean EPDS score at the time of DST Intervention implementation (T2 as indicated in Table 1; mean score: 15, SD = 2) indicates that our participants were at high risk for developing Postpartum Depression. In a systematic review of 37 validation studies of the EPDS for postpartum depression, a cut-off score of 9/10 provided “strong evidence” for minor and/or major depression.²⁰ Our participants also reported several adverse childhood experiences, mild reported stress, and moderate risk for Generalized Anxiety Disorder based on GAD-7 scores. These scores tell us that our participants demonstrated various dimensions of vulnerability to emotional distress during pregnancy.

Hair samples were willingly provided by each of the participants at the DST time point, as well as by seven participants at all three follow-up time points in the study. In follow-up interviews with participants after the intervention was completed, feedback on hair sample collection was generally favourable, with most participants expressing comfort with providing hair samples, especially when they saw how little was obtained for a successful sample. Even participants who were initially hesitant about providing hair samples emphasized that the actual amount of hair obtained from the procedure was not enough to warrant their continued anxiety. This sentiment is exemplified in the following statement from a participant during the interview:

Investigator: … is there anything you think we can do to make other women more comfortable with the hair sample? Like, things to say, things to do?

Participant: … I feel like, regardless … I was freaking out, there are going to be people who freak out, but … once they see that you don't know where it really came from, I don’t think they're going to be like, “Oh, that's a big deal.”

During follow-up interviews, participants consistently articulated their interest in learning the results from collecting their hair samples. For many, the understanding of why we were collecting hair samples—as a means of measuring cortisol—translated into the idea that these results could indicate the degree to which they were stressed on a daily level:

Participant 1: I would, I would never usually do that [provide a hair sample]. But I'm pretty interested in knowing the stress level…[i]t was a way to actually kind of know.

Participant 2: I mean, I do want to know my stress levels, because I feel like I have stress every day. [Laughs]

Participants regularly projected their own understandings onto the hair sample collection, which transformed their hair into an immediately quantifiable variable—a stress level. This translated into the understanding that they might be able to pinpoint “how stressed” they were based on each individual hair sample collected.

Biology, Collaboration, and Meaning

This study was built from a foundation of previous work done in collaboration with community organizations and stakeholders in the Springfield Metropolitan Area^10,21 and is principally concerned with centering the needs of community members as communicated by these stakeholders. While biological data can provide meaningful information on the efficacy of community-based interventions, the collection of this data can be complicated by ethical and socio-cultural considerations. We sought to address these barriers with intervention participants and home visit service providers alike in our pilot study. By refocusing the discussion around biospecimen data to the positionalities and experiences of the bodies from which the samples are obtained, we believe meaningful ethical and methodological considerations emerged for other investigators who wish to incorporate biological data in building holistic community-based interventions.

Home visitors addressed several of our considerations, emphasizing the need for comprehensive communication regarding the purpose of obtaining hair samples, along with the possibility that specific spiritual and cultural groups in the community may prefer another form of data collection due to the value and beliefs associated with hair. Participants were supportive of providing hair samples. One reason for participant enthusiasm and willingness to provide hair samples may have been related to the collaborative and trust-building nature of the DST intervention. By nature, DST workshops are intensive, and inherently promote vulnerability between the participant and researcher. As a result, DST investigators serve the role of facilitators-researchers: they collaborate with participants to establish community norms during workshops, provide feedback on written material, troubleshoot media and technological details of constructing the digital stories, and most importantly, support participants through the emotional and technical triumphs and vulnerabilities inherent to the digital story construction process. The facilitator-researcher role creates a foundation of trust and opens up lines of communication by creating a structured space where both parties come together to create a meaningful end product (a digital story). As a result, a considerable amount of participant-researcher trust is developed in a short amount of time. Our findings indicate that using CBPR methods that utilize intensive, generative processes, and how these processes might carry through into participants’ willingness to participate in longitudinal follow-up.

Within the landscape of the DST workshop, data collection can be organically weaved into the structure of the intervention to best leverage the foundation of trust between researcher and participant. In our study, participants noted that they appreciated the collection of hair samples later on in the first day of the DST workshop, as this afforded a sense of ease and trust before providing their hair samples to the researchers. Hair samples also afforded a greater flexibility of sample collection time than saliva, which requires several samples taken over the course of a day to gain an accurate representation of participants’ diurnal curves.

Intervention participants also consistently attached meaning to their samples as biological data, and, in knowing that we were measuring cortisol because of its relationship to stress physiology, were interested in what their samples might illuminate about their individual stress levels through each sample collection. Given that four hair samples were collected over the course of the study, it was important to emphasize to participants that we would not be able to provide stress level data on each sample, but rather that, analyzed together, the samples could map out their stress levels over time (i.e. provide data on chronic, rather than discrete, levels of stress). The desire for biological meaning at discrete time points may be heightened by the focus given to the body-as-biological during pregnancy.²² This search for biological meaning by participants has immense social scaffolding via lay health data collection that occurs via apps like those used by our participants,²³ which they regularly employed and findings from which they shared over the course of the DST intervention. Additionally, the idea that our research could tell participants something “about their stress” resembles other findings in CBPR research regarding the conflation of participant biological data analysis as providing a clinical service.² These findings indicate that investigators should not only have a very clear understanding of why they choose to collect certain forms of biological data: they must also think carefully about how they will share this data with participants to account for the cultural phenomenon of personal health data and the risk of participant conflation.

Our participants, not nearly as concerned with ideas of power imbalance or surveillance as we anticipated, were instead intrigued by what the data we collected could say about their state of mind or well-being. Together, our focus groups with home visitors and our engagement with participants challenge the conventional portrayal of participants as “passive” members in the biological sample collection and analysis process.

Conclusion

This paper aimed to characterize the feasibility and acceptability of obtaining hair samples for biospecimen analysis from Latina and African-American women in a mixed-methods study. While our participants were not as skeptical of the collection of biological data as our team predicted, there is still room for improvement and reflection in how researchers communicate and contextualize biological data collection to research participants. Given our findings, we suggest that researchers consider the following in excess of IRB benchmarks:

Creatively incorporate explanations of how and why biological samples will be collected into the processes of building trust. For instance, participants noted that they appreciated the collection of hair samples later on in the first day of the DST workshop, as this afforded a sense of comfort in providing their hair samples to the researchers.

Continually emphasize the difference between biological data collected for research purposes, and biological data collected for clinical or diagnostic purposes. In vulnerable populations, where stress is not just a biological measure, but also a daily reality, the findings obtained from analysis of biological measures may have significant meaning for participants. Researchers should be prepared to continually discuss the limitations of what researchers can convey to participants about their overall health from the collection of biological data.

Consideration of innovative models for biological data storage, communication of results, and resource referral. Yassin et al.²⁴ propose an ethical framework for biospecimen research rooted in custodianship, where researchers function as intermediaries towards knowledge, and participants are understood as stakeholders from the point of providing biological data to the eventual write-up of findings.²⁴ In this spirit, community resources for referral, from social services, to programs in housing, education and employment assistance should be meaningfully integrated into the process of data collection. These resources should be made accessible to all participants regardless of findings.

We present these suggestions as possible considerations for researchers who wish to carefully, and intentionally, integrate the collection of biological samples and biospecimen data, especially those minimally invasive measures like hair sample collection, as part of their research. While we recognize the utility of biology in understanding health, we are also cognizant of a complex history that crosses science, research, and those communities which are most marginalized. The findings in this study indicate that the path forward in community-based, mixed-methods research can be an equitable one when investigators thoughtfully integrate and communicate biological data in their work, and commit to operating from a place of participant empowerment and autonomy.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support forthe research, authorship, and/or publication of this article: (1) the Developmental Science Seed Grant Initiative, 2017-2018, for project entitled “Pilot Testing a Community-Based Digital Storytelling Intervention for PPD Prevention in Adolescents,” University of Massachusetts – Amherst, (2) The FPR-HC Culture, Brain, and Development Program, Hampshire College, 2017-2018; and (3) The Denice O’Neill Award, Hampshire College, 2017-2018.

ORCID iDs

Hannah Davidson

A. C. Gubrium

Author Biographies

Hannah Davidson is a Postbaccalaureate Intramural Research Training Awardee at the National Institutes of Health. Her research interests include perinatal health, patient-provider interactions, and health communication.

A. C. Gubrium is a professor in Health Promotion and Policy in the School of Public Health and Health Sciences at the University of Massachusetts Amherst. She conducts research on critical narrative intervention and reproductive and sexual health inequities with marginalized women, youth, and families.

References

Javitt

Why not take all of me? Reflections on the immortal life of Henrietta Lacks and the status of participants in research using human specimens. Minnesota J Law Sci Technol 2010; 11: 713.

Hohl

Gonzalez

Carosso

, et al. “I did it for us and i will do it again”: perspectives of rural latinos on providing biospecimens for research. Am J Public Health 2014; 104: 911–916.

Truog

Kesselheim

Joffe

Paying patients for their tissue: the legacy of Henrietta Lacks. Science 2012; 337: 37–38.

Christopher

Watts

McCormick

, et al. Building and maintaining trust in a community-based participatory research partnership. Am J Public Health 2008; 98: 1398–1406.

Banks-Wallace

Talk that talk: storytelling and analysis rooted in African American oral tradition.

Qual Health Res 2002; 12: 410–426.

Cueva

Kuhnley

Revels

, et al. Bridging storytelling traditions with digital technology. Int J Circumpolar Health 2013; 207-217: 72.

DiFulvio

Gubrium

Fiddian-Green

, et al. Digital storytelling as a narrative health promotion process: evaluation of a pilot study. Int Q Commun Health Educ 2016; 36: 157–164.

Zainuddin

Kadir

Wijayanti

, et al. Effect of digital storytelling on self-efficacy of patients with a stoma: preliminary study. Enferm Clin 2020; 30: 229–233.

Sehrawat

Jones

Orlando

, et al. Digital storytelling: a tool for social connectedness. Gerontechnology 2017; 16: 56–61.

10.

Gubrium

Fiddian-Green

Lowe

, et al. Measuring down: evaluating digital storytelling as a process for narrative health promotion. Qual Health Res 2016; 26: 1787–1801.

11.

Bloch

Methodological challenges for national and multi-sited comparative survey research. J Refugee Stud 2007; 20: 230–247.

12.

Mercer

K. Black hair/style politics. In: Owusu K.

Black British culture and society: a reader. London: Routledge, 2000, pp. 111–121.

13.

Strang

Black

Marsh

, et al. Hair analysis for drugs: technological breakthrough or ethical quagmire? Addiction 1993; 88: 163–166.

14.

O’Brien

Tronick

Moore

CL.

Relationship between hair cortisol and perceived chronic stress in a diverse sample.

Stress Health 2013; 29: 337–344.

15.

D’Anna-Hernandez

Ross

Natvig

, et al. Hair cortisol levels as a retrospective marker of hypothalamic-pituitary axis activity throughout pregnancy: comparison to salivary cortisol. Physiol Behav 2011; 104: 348–353.

16.

Meyer

Novak

Hamel

, et al. Extraction and analysis of cortisol from human and monkey hair. J Visual Exp 2014; 83: e50882.

17.

Binz

Rietschel

Streit

, et al. Endogenous cortisol in keratinized matrices: systematic determination of baseline cortisol levels in hair and the influence of sex, age and hair color. Foren Sci Int 2018; 284: 33–38.

18.

Clare

Yeh

Postpartum depression in special populations: a review.

Obstet Gynecol Surv 2012; 67: 313–323.

19.

Logsdon

Birkimer

Simpson

, et al. Postpartum depression and social support in adolescents. J Obstet Gynecol Neonatal Nurs 2005; 34: 46–54.

20.

Gibson

McKenzie-McHarg

Shakespeare

, et al. A systematic review of studies validating the Edinburgh Postnatal Depression Scale in antepartum and postpartum women. Acta Psychiatr Scand 2009; 119: 350–364.

21.

Gubrium

Krause

Lucey

‘ Doing your life’: narrative intervention with young mothers as storytellers. Hum Org 2018; 77: 214–227.

22.

Martin

The woman in the body: a cultural analysis of reproduction. Boston: Beacon Press, 2001.

23.

Thomas

Lupton

Threats and thrills: pregnancy apps, risk and consumption. Health Risk Soc 2016; 17: 495–509.

24.

Yassin

Lockhart

Del Riego

, et al. Custodianship as an ethical framework for biospecimen-based research. Cancer Epidemiol Prev Biomark 2010; 19: 1012–1015.