Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: Stress. 2019 Oct 22;23(3):265–274. doi: 10.1080/10253890.2019.1675628

Assessment of Feasibility and Outcomes of a Salivary Cortisol Collection Protocol in Five American Indian Communities

Melissa Walls a,1, Melinda Dertinger b, Michael Unzen b, Angie Forsberg a, Benjamin Aronson c, Stephanie Wille a, Mustafa al’Absi b; Gathering for Health Team*
PMCID: PMC7174135  NIHMSID: NIHMS1542791  PMID: 31578895

Abstract

We evaluated the feasibility and outcomes of administering a naturalistic saliva collection procedure and assessment in American Indian (Indigenous) communities. We focus on Indigenous adults living with type 2 diabetes given the “epidemic” of the disease disproportionately impacting many tribal groups. Data are from community-based participatory research (CBPR) involving 5 tribal communities. Participants were randomly selected from tribal clinic records. The sample includes 188 adults living with type 2 diabetes (56% female; age range = 18 - 77 years; M age = 46.3 years). Participants provided a total of 748 saliva samples, representing 4 samples/participant on a single day with instructions for collection at 4 time points: upon waking, 1 hour after waking, 2 hours after waking, and at 8 PM. Saliva sample times were recorded by participants on paper and electronically via placement in a Medication Event Monitoring System (MEMS®) bottle. Overall, 67% of samples were completed within 10 minutes of protocol instructions and 91% of participants provided at least one useable sample (79% provided four useable samples). Noncompliance, behavioral and environmental factors were not robustly associated with deviations in observed cortisol indices. Results suggest that home-based, community interviewer-involved protocols yields valid data with high compliance. The success of this study was facilitated by exemplary efforts of tribal community-based interviewers and our overall CBPR approach.

The harmful effects of social, psychological and physiological stress on health and well-being are widely documented (Williams et al., 2016; McEwen, 2008). Stress process models (Pearlin et al., 1981) cite socio-economic factors as antecedents of stress exposure (Aneshensel, 1992) wherein members of disadvantaged groups more frequently encounter stressors and therefore disproportionately bear burdens of subsequent ill-health. Stress process approaches are thus useful models from which to understand the unfolding of health inequities. To illustrate, American Indians (AIs) experience striking health disparities including chronic diseases like type 2 diabetes (T2D), which affects AIs at double the rate of the general U.S. population (Centers for Disease Control and Prevention [CDC], 2011; CDC, 2017). Many scholars working with AI communities consider stress in terms of its accumulation across generations; deeply rooted historically traumatic policies and experiences are formative pathways to contemporary stressors like poverty and discrimination (Burnette, 2015; Gracey et al., 2009). Stress has been implicated in the etiology of T2D for centuries, and psychosocial stressors are linked to poorer disease management and prognosis in diverse populations including AIs (Jiang et al., 2008; Walls et al., 2017). Thus, cumulative and disproportionate exposure to stressors is connected to the emergence of T2D as a major health inequity impacting AI communities.

Still, there are limits to our understanding of the role of stressors in disease processes like T2D. There are vast psychosocial stressor measurement possibilities ranging from chronic strains like unemployment, to severe trauma, to discrete life events including loss of a loved one, job changes, and relationship dissolution (Wheaton, 1994). Survey measures of stressors may thus underestimate stress burden (Turner, 2013), a problem amplified for racial/ethnic minority communities and AIs specifically, for whom culturally and contextually specific stressor measures (e.g., cultural loss, historical trauma; Whitbeck et al., 2004) remain underdeveloped. Stress biomarkers may afford a culturally neutral way of identifying the impacts of stress on disease progression and serve as a tool for measurement triangulation or cross-validation.

Cortisol is a classic and reliable marker of stress in healthy adults; the diurnal cycle of cortisol levels peaks in the early morning and decreases rapidly during the day (Born, Hansen, Marshall, Molle, & Fehm, 1999; Dallman, 1993). Cortisol release by the HPA axis is a result of a cascade that is initiated by the release of corticotropin-releasing factor (CRF) from neuronal cell bodies of the paraventricular nucleus. CRF acts on the corticotrope cells of the anterior pituitary, stimulating synthesis of pro-opiomelanocortin (POMS), and leading to the subsequent release of adrenocorticotropic hormone (ACTH) and endorphin into the systemic circulation. Upon reaching the adrenal cortex, ACTH leads to the synthesis and release of cortisol. The flattening of the slope of the daily decline of cortisol is a biomarker of impaired response to stress, and it has been associated with poor health outcomes in multiple conditions (Steptoe, Cropley, Griffith, & Kirschbaum, 2000; Joseph & Golden, 2017). This pattern is altered when high levels of psychological stress, physical stress or disease impair the stress response of the hypothalamic-pituitary-adrenal (HPA) axis (al’Absi, 2018). Medications also have the potential to influence HPA activity and salivary cortisol both directly and indirectly (Granger et al., 2009).

Salivary cortisol is a particularly appealing stress biomarker for community-based studies as a practical, low-risk, non-invasive method (Halpern et al., 2012). The ease and reliability of salivary sample collection has popularized the method for evaluating daily HPA functioning in field and lab studies (Nicolson, 2008; Saxbe, 2008). At-home, self-collection allows researchers to monitor cortisol under natural conditions and answer questions that cannot be examined in the laboratory for high ecological validity (Jacobs, Nicolson, Derom, Delespaul, van Os, & Myin-Germeys, 2005; Smyth, Thorn, Hucklebridge, Evans, &Clow, 2015). Self-collection requires monitoring of adherence to protocol, sample timing, and behaviors (e.g., eating, smoking, drinking, exercising) to avoid inaccurate depictions of actual diurnal rhythm of cortisol (Nicolson, 2008; Saxbe, 2008; Kudielka et al., 2003; Steptoe et al., 2006; Lovallo et al., 1996; Rudolph et al., 1998). Protocol adherence is especially important to monitor during waking and early morning hours, when cortisol levels are most variable and even moderate delays in sampling can result in inaccurate cortisol calculations (Nicolson, 2008; Smyth et al., 2015). Protocol adherence appears to be less of a concern for values taken later in the day and diurnal patterns (Jacobs et al., 2005). In sum, there is strong potential utility of self-collection methods with monitoring and checks of protocol compliance.

Limited applications of salivary cortisol collection with AI people has focused on laboratory, clinic or hybrid (e.g., lab and home) settings (Kaholokula et al., 2012; Laudenslager et al.,2009). We were unable to locate any exclusively home-based studies that examined adherence to and feasibility of administering self-collection saliva procedures with AI communities. On the one hand, heightened exposure to stressors and related diseases like T2D suggest special need for empirical attention to these issues for AIs who are historically underrepresented in health sciences and biomedical research (Hussain-Gambles et al., 2004). On the other, mistrust of researchers within many AI communities has resulted from unethical and in some cases harmful studies that have stigmatized tribes and Indigenous people (Boyer et al., 2011; Hodge, 2012). Notorious among the examples is the ethical misconduct case involving the Havasupai tribe, whose members’ blood samples were used to publish papers on unauthorized topics incongruent with cultural knowledge (Harmon, 2010; Tovar et al., 2014). Such exploitation is but one catalyst for heightened wariness of biomedical research for many AI communities (Pember, 2010), perhaps best exemplified by a 2002 moratorium on genetic research issued by the Navajo Nation, a ban that remains in place as of this writing.

Also critical to research in AI communities is respect for tribal sovereignty and local protocols for research, the latter of which vary tremendously across tribes. AI individuals report greater willingness to engage in research when studies are conducted by trusted entities, appropriately involve community, adequately address confidentiality and compensation, and focus on health issues of relevance to the community (Buchwald et al., 2006). Our community/university research team employed a community-based participatory research (CBPR) approach to gain an understanding of daily HPA activity among a sample of AIs living with a chronic illness, T2D. In this study, we evaluate the feasibility of administering a naturalistic saliva collection procedure and biological assessment in AI communities. Several research questions guide our analyses:

  1. What is the degree of participant adherence to unsupervised salivary collection protocols in a home-based, tribal community setting?

  2. Does level of compliance to protocol influence observed cortisol values?

  3. Do environmental and behavioral (e.g., smoking, eating, medications) factors occurring near the time of saliva collection influence observed cortisol values?

  4. Do demographic factors relate to observed cortisol values?

Methods

Data are from the Maawaji’ idi-oog Mino-ayaawin (Gathering for Health) project, a CBPR collaboration between the University of Minnesota and five Anishinaabe (Ojibwe) communities in Minnesota and Wisconsin. The project was established to examine the role of AI-specific stressors on T2D outcomes, and is supported by resolutions from each tribal government. Project procedures were reviewed and approved by the UMN and Indian Health Service National Institutional Review Boards. Community Research Councils (CRCs) comprised of an average of 6 members on each reservation are active partners in the research process and have participated in study planning, protocol development, and implementation to ensure cultural and local acceptability of study procedures. CRC members serve as co-authors and presenters on various data dissemination and translation activities for locally relevant programming and services. Salivary collection and disposal protocols were carefully developed with CRC members, two of whom toured assay units on campus to observe general assay and disposal procedures prior to formal data collection.

Sample

Medical/clinical staff at each participating community medical facility generated lists of possible participants using simple random sampling techniques. Inclusion criteria were a diagnosis of T2D documented in the medical record within 5 years of the sampling date, age 18 years or older, and self-identification as AI. A total of 194 participants enrolled in the study, representing a baseline response rate of 67%. Data for this report include responses from the 188 participants who provided baseline saliva samples.

Procedure

Clinic staff sent study invitation letters and brochures to residences of randomly selected patients. Mailings included initial options for study refusal including a pre-paid postcard and telephone number for clinic staff. Individuals who did not refuse study participation were contacted by trained community interviewers, screened for study eligibility, and formally invited to participate. Those who agreed to enroll in the study scheduled interviewer visits at a location of their choosing, at which time interviewers gathered signed informed consent and HIPAA authorization forms. Data were collected from three sources: 1) Computer-Assisted Personal Interview (CAPI) surveys assessing psychosocial stressors, coping factors, and diabetes outcomes; 2) Medical chart reviews to record recent lab results, health histories, and prescribed medications; and 3) Salivary Cortisol and Subjective State assessments (part 3 represents the data included in this report). Participants received a $50 incentive for survey completion plus $50 for adhering to saliva collection protocols.

Salivary Cortisol Collection Procedure and Materials

Interviewers were tribal members who lived and worked in participating tribal communities and attended an intensive, 3-day university hosted training on human subjects safety and study protocols followed by home-based practice interviews and a phone-based interviewer certification process assessed by a study coordinator. Full team (i.e., 5 sites plus university team) Booster Training sessions occurring every 6 months throughout the duration of data collection included additional training along with opportunities for interviewers to debrief, strategize, and support one another. Interviewers role-played salivary cortisol collection procedures by pairing up and completing samples and SSS forms (see below) following instructions from their partner. After training and certification, interviewers were approved to begin data collection. For salivary cortisol procedures, interviewers provided in-person instruction to participants to collect saliva samples four times during a single day: immediately upon waking, one hour after waking, two hours after waking, and in the evening at approximately 8:00 pm. Interviewers provided detailed written and scripted verbal instructions to participants on proper saliva collection and storage techniques. Reading level of all instructional materials was assessed and approved by CRC members. As part of these instructions, interviewers demonstrated proper collection techniques and helped participants practice collecting an “example sample.” Participants were advised not to brush their teeth, eat, or drink anything before the first sample and at least 15 minutes before each subsequent sample, to keep each Salivette™ in their mouth for at least two minutes until it was “very saturated,” and to collect samples on a “typical day.” Interviewers also instructed participants how to correctly use Medication Event Monitoring System (MEMS®, Aardex Ltd., Switzerland) bottles to record time of sample collection (e.g., to only open the bottle at time of sample collection and note any times the bottle was accidentally opened).

Participants were provided 4 Salivette™ tubes pre-labeled with the ID number, date, and sample number. The sampling points were chosen with two goals in mind; first, to capture significant changes in cortisol release during the first part of the day; second, to reduce burden on participants and increase adherence with the protocol. Participants recorded clock time during each collection and were instructed to place swabs in a MEMs bottles, which electronically record exact times the bottle is opened. Times were read and documented by trained interviewers upon sample pickup at participants’ homes. In the event of missing times, discrepancies between MEMS and participant reported times, and/or divergence from instructed times, interviewers offered a “second chance” and saliva collection procedures were repeated. Interviewers delivered saliva samples to a −20°C freezer held at each reservation clinic, where they were kept until they were transported in a cooler with dry ice to the UMN laboratory for processing and/or longer term storage in −80°C freezers.

Salivary Cortisol Assays

Salivettes were centrifuged at 2000rpm in an Eppendorf 5810R centrifuge using an A-4-62 rotor for 10 minutes at 4°C and the saliva was pipetted into a cryovial and stored at −20°C until sample was assayed. Cortisol levels were assayed in saliva using Cortisol Saliva Luminescence Immunoassay kit from IBL International (Cat. No. RE62111). All samples, standards, and controls were run in duplicate, with resulting inter-assay and intra-assay coefficients = 8.99% and 7.43%, respectively.

Measures

Each salivary sample was connected to three distinct times: participant self-reported, MEMS bottle times recorded by interviewers at time of pickup, and the MEMS bottle times recorded by University staff. A final time was selected for each sample to use in analyses. In order to determine the most accurate sample completion times for this variable, we carefully examined each source of data with consideration of any field notes from interviewers. All discrepancies were evaluated on a case-by-case basis. The MEMS bottle times recorded by University staff (UMN MEMS times) were generally considered most accurate due to verification checks we were able to do at the university and were selected as the final time variable when these times were available and for discrepancies <20 minutes or when consistent with interviewer-recorded MEMS times. In the event that UMN MEMS times were not available, interviewer recorded MEMS times were selected as the final time variable. Participant self-reported times were selected if no bottle times were recorded by the interviewer or university.

We calculated several cortisol measures to assess HPA functioning in this sample. Early morning cortisol was assessed by computing the difference in cortisol values between Sample 1 (awakening) and Sample 2 (1 hour later). Cortisol Change between morning (Sample 1) and evening (Sample 4) values were calculated by subtracting Sample 4 from Sample 1. A healthy diurnal cortisol pattern consists of higher values in the morning than evening; thus, a negative value for this variable may indicate dysregulation of the HPA axis and cortisol response (Nicolson, 2008). Diurnal Decline was calculated by subtracting Sample 4 (when values should be lowest) from Sample 2 (when values should be highest). Area Under the Curve with respect to ground (AUCG) was calculated as a measure of total hormonal output (Fekedulegn et al., 2007). In the current study, individual cortisol samples and time intervals between each sample were used to calculate AUCG using the trapezoidal technique suggested by Preussner et al. (2003). Single Sample Values were examined for possible measurement error and any outlier values that required deeper examination.

We also include several variables from surveys and chart reviews. We assessed participant use of medications reported in the medical chart reviews and/or survey self-reports. Medications were coded based on select indications (e.g., hypertensive medication, T2D medication) and pharmacologic categories (e.g., sulfonylurea, statin).

At the time of each saliva sample collection, participants were asked to complete the Subjective State Scale (SSS) to assess their environment, behavior, and subjective feelings of stress (Lundberg et al.,1980; al’Absi et al., 2004). Participants reported time, location, presence of other people, indication of significant events in the hour prior to the sample, and if so, how stressful the event was (Not stressful at all, Moderately stressful, or Severely stressful). They also indicated if they engaged in behaviors including smoking, exercise, or consumption of caffeine, alcohol, or food during the hour prior to each sample collection (response options for each behavior were Yes or No). The responses to these questions were considered when interpreting cortisol measures, as these variables have been shown to influence cortisol activity (Nicolson, 2008; Steptoe et al., 2006; Lovallo et al., 1996; Rudolph et al., 1998). Last, participants indicated subjective state factors by rating the degree to which descriptive words/symptoms applied to how they felt during the last 30 minutes on a scale from 0 (Not at all) to 7 (Strong). These items were coded into three subscales following previously used classifications (al’Absi et al., 2003; al’Absi et al., 2004; al’Absi et al., 2005; Huttlin et al., 2015) and results of exploratory factor analysis: positive affect (e.g., cheerful, happy, content; alpha at each sampling point ranged from .89 to .93), distress (e.g., anxious, irritable, sad; alpha .79 - .84), and physical symptoms of stress (e.g., headache, sweating, stomach problems, drowsiness; alpha .68 - .79).

Results

Sample Characteristics

A total of 188 American Indian participants (56% female) ranging in age from 18 - 77 years (M = 46.3) with a diagnosis of T2D provided cortisol samples for the study. The average length of time since T2D diagnosis was 1.56 years. Most participants (78%) were living on reservation land. Participant household per capita income ranged from $156 to $37,344 (M = $9,862). A majority of participants (89%) had at least a high school diploma or GED, with 14% holding a college or graduate degree. Most participants reported either being single (34%) or married/living with a partner (52%). A majority of participants (93%) were taking at least one prescribed medication.

Protocol Compliance

Saliva was collected at four different time increments throughout a single day resulting in a total of 748 individual baseline samples. Overall, 91% of participants provided at least one useable sample (i.e., a sample with enough volume to be assayed), and 79% provided useable samples for all four sampling points. Protocol compliance was only examined with the data collected from the second chance attempt of participants who required a second attempt as data were not collected from participants’ first chance attempts.

We examined the times that participants provided each saliva sample to determine protocol compliance. Self-reported times were within 10 minutes of MEMS times for most participants with available time data, ranging from 71% at Sample 4 to 83% at Sample 1. For each sample, MEMS times were selected as the final times for approximately 66% of samples. MEMS times were generally considered most accurate but were not selected as the final times for various reasons – most often when these times were not recorded, but also occasionally for other reasons (e.g., participant noted accidental bottle opening, time was mislabeled as PM instead of AM, bottle was only opened 3 times during the day). In the event of missing UMN MEMS times (~26% of samples), we selected bottle times recorded by the interviewer at time of sample pickup as the final time. Interviewer-recorded and University-recorded MEMS times were exact matches for 76% of the samples with available time data. Mean time durations between samples 1 to 2 (Interval 1) and samples 2 to 3 (Interval 2) were close to the required hour (Table 1).

Table 1.

Instructed and Observed Salivary Cortisol Collection Timing

Protocol Instructions Mean (SD) Observed Sample Time/Interval
Sample 1 Immediately upon waking 7:57 AM (1:55)
Sample 2 1 hour after waking 9:06 AM (2:02)
Sample 3 2 hours after waking 10:24 AM (2:22)
Sample 4 8:00 PM 8:10 PM (1:40)
Interval 1 (S1 to S2) 60 minutes 68.55 (32.48) minutes
Interval 2 (S2 to S3) 60 minutes 79.35 (47.29) minutes
Open in a new tab

Following (Golden et al., 2014), we calculated adherence to protocol within 10 minutes of instructed time intervals; 67% of all samples collected throughout the day were provided within this window. Deviations in sampling times up to 60 minutes from instructions are generally considered acceptable with the exception of cortisol awakening response (CAR), which was not assessed in the current study (Nicolson, 2008). Thus, we also calculated protocol compliance using the 60-minute criteria. A majority of all sample times (75%) were within this 60-minute window on either side of the protocol instructions at all three time points (Interval 1, Interval 2, and the evening sampling time). We examined differences between participants within or beyond 120 minutes (i.e., 60 minutes or less deviation on either side of timing instructions) during Interval 1, Interval 2, and/or at Sample 4; n = 44 participants were not compliant based on this criteria. There were no significant differences across cortisol indices between the more compliant and less compliant participants at each sampling interval. We also performed separate t-tests to examine the influence of the largest discrepancies (i.e., outliers and extreme cases) on cortisol indices. Only the morning response was significantly influenced by outliers of Interval 1; specifically, participants with extreme time discrepancies between samples 1 and 2 displayed negative early morning response values (mean = −5.12 nmol/L), whereas more compliant participants displayed positive early morning response values (mean = .61 nmol/L), t(157) = 2.67, p < .01. However, these differences were non-significant when examining the influence of outliers on percent change from sample 1 to 2 (rather than the change in raw values; p = .45).

We also examined adherence to instructions not to eat, drink, consume caffeine, or smoke before sample collection (Table 2). In general, most participants refrained from engaging in these behaviors for one hour before collecting each sample, despite instructions to refrain for only 15 minutes before samples 2 – 4. The overall compliance rate regarding instructions to abstain from eating, drinking, or at all sampling times was 74.3%. Participants adhered most to these instructions when providing sample 1, and least when providing sample 4. Consuming caffeine appeared most often during the ~1-2 hours after waking, eating was most common in the evening, and smoking became more common as the day went on.

Table 2.

Participant self-reported adherence behaviors

Participant adherence to refraining from: %
Sample 1 Smoking 85.6
Caffeine 87.6
Alcohol 99.4
Food 93.3
Sample 2 Smoking 67.0
Caffeine 56.4
Alcohol 98.9
Food 74.3
Sample 3 Smoking 63.6
Caffeine 49.1
Alcohol 99.4
Food 65.3
Sample 4 Smoking 50.9
Caffeine 59.9
Alcohol 95.2
Food 42.9
Open in a new tab

We also examined whether or not compliance to timing instructions was influenced by behavioral and environmental variables (i.e., positive affect, distress, experiencing a stressful event, smoking, eating, consuming caffeine). Compared to those who smoked early in the morning, participants who refrained from smoking before sample 1 were more likely to be within 10 minutes of the 60-minute interval during Interval 1, χ2 (1, N = 177) = 6.77, p < .01. None of the other behavioral or environmental variables tested were significantly associated with compliance to timing instructions (within either 10-minute or 60-minute deviations from protocol).

Cortisol Concentrations and Correlates

On average, we saw a normal diurnal cortisol pattern in the data: a rise after waking, a gradual decrease throughout the day, and lower values in the evening than upon waking. An average rise of 36.1% was seen in cortisol levels one hour after waking. A minority of participants (9.5%) demonstrated higher evening than waking values. This dysregulated pattern was not explained by differences in time of the first sample: there were no significant differences in time of sample 1 (or any of the following sampling points) between those who showed this dysregulated pattern and those who did not. Additionally, participants with a sample 1 time of at least 1 standard deviation from the mean did not display significantly different cortisol values at any of the sampling points than those within 1 standard deviation.

Medications

Among medications for T2DM, metformin was the most used medication (n = 134), followed by sulfonylureas (n = 24) dipeptidyl peptidase-4 inhibitors (n = 14), and 8 individuals taking other diabetes medications. Well over half of the participants were prescribed medication for blood pressure (n = 127). Other notable medication classes and categories included cholesterol medications (n = 81) and specific statin medications (n = 71), aspirin (n = 62), and medications for depression/anxiety (n = 48). Thirteen participants were taking no prescriptions. There were no significant differences between participants not taking medications and participants taking at least one medication. Separate t-tests revealed that participants taking a prescription blood pressure medication displayed significantly higher cortisol waking values (M = 8.86) than participants not taking blood pressure medication (M = 6.92), t(159.83) = −2.42, p < .05. After applying a Bonferroni correction to this analysis, however, this difference was no longer significant at the p < .05 level.

Self-report measures and Cortisol Responses

Table 3 lists correlations between reported behaviors, affective states, and symptoms from the SSS forms. Consuming caffeine before Sample 1 was associated with a decrease in the percent change from Sample 1 to Sample 2 (M = −22.59%, SD = 40.12%), whereas not consuming caffeine before Sample 1 resulted in a statistically significant increase in percent change from Samples 1 to 2 (M = 45.48%, SD = 156.51%; t(110.82) = 4.11, p < .001). Participants who smoked before Sample 1 displayed significantly higher waking cortisol values than did participants who did not, t(157) = −2.91, p< .01.

Table 3.

Associations (Pearson’s r) between SSS Form Responses and Cortisol Values

Sample 1 Sample 2 Sample 3 Sample 4
Experienced stressful event .01 .13 .03 −.05
Smoke .23** −.01 −.11 −.07
Caffeine .11 −.12 −.10 −.07
Alcohol -- -- -- .11
Eat .09 −.08 −.07 .07
Exercise .01 −.06 −.03 −.02
Positive affect .06 .14† .17* .04
Distress −.15† −.12 −.10 −.07
Physical symptoms −.16* −.09 −.06 −.06
Open in a new tab
†

p < .10;

*

p < .05;

**

p < .01;

***

p < .001

Notes: Correlation values are associations of cortisol values and SSS responses taken at the same sampling time. Correlation values were not reported for Alcohol at Samples 1, 2, and 3 due to a lack of participants who reported consuming alcohol at these times

Cortisol levels were also examined in terms of participants’ self-reported affect and symptoms on each SSS Form. Affective states and symptoms at each sampling time were grouped into three categories: positive affect, distress, and physical symptoms of stress Positive affect included feeling cheerful, content, calm, in control, interested, and happy. Distress included feeling anxious, irritable, impatient, sad, and angry. Physical symptoms of stress included headache, sweating, tremor, stomachache, drowsiness, fatigue, and coughing. Positive affect and reported physical symptoms were the only affective states significantly correlated with cortisol values, but in different directions and only at single sampling times (Sample 3 and Sample 1, respectively).

Cortisol Outcomes by Demographic Characteristics

We examined cortisol indices by demographic characteristics (Table 4). Waking values were significantly higher for men than women, t(162) = 2.49, p < .05. No other cortisol index differed by gender. Per capita annual income was related to total cortisol output throughout the day. Specifically, those in the lowest income quartile displayed significantly lower AUCG values than those in the highest income quartile, t(77) = −2.51, p < .05. Participant age was also significantly related to waking values, F(3, 160) = 3.14, p < .05. Post hoc comparisons using a Bonferroni correction indicated that participants in the youngest age quartile experienced significantly lower waking values than did participants in the oldest age quartile. Cortisol indices were unrelated to living on or off reservation lands during baseline data collection.

Table 4.

Mean Cortisol Values (in nmol/L) by Demographic Variables

Sample 1 (SD) Sample 2 (SD) Sample 3 (SD) Sample 4 (SD) AUCG (SD) Early Morning Response (SD) Change from Sample 1 to Sample 4 (SD) Diurnal Decline
TOTAL 8.15 (5.58) 8.78 (9.02) 5.31 (5.04) 2.57 (4.91) 1069.28 (848.02) 36.06% (146.48%) 5.93 (5.70) 6.33(7.11)
Mininum 0.63 0.68 0.16 0.11 77.47 −78.28% −7.58 −9.72
Maximum 38.99 80.44 50.60 57.23 7431.58 850.24% 35.37 38.62
Gender:
Men 9.42 (6.19)* 9.86(11.20) 5.33 (4.21) 3.25 (6.83) 1085.67 (707.15) 24.13% (153.37%) 6.88 (6.21)† 6.91 (7.75)
Women 7.26 (4.89)* 8.04 (6.93) 5.34 (5.66) 2.06 (2.56) 1070.16 (953.78) 44.18% (142.02%) 5.31 (5.21)† 6.00 (6.67)
Income:
Lower Quartile 7.52 (7.11) 6.67 (4.15)† 4.74 (3.79) 2.47 (2.10) 832.31 (419.48)* 18.21% (81.14%) 5.05 (7.09) 4.54 (5.28)
Upper Quartile 7.83 (5.09) 10.00 (7.56)† 4.94 (7.64) 2.73 (3.98) 1171.21 (808.04)* 70.12% (175.71%) 5.11 (5.32) 6.89 (7.76)
Age:
Quartile 1 6.68 (4.00)* 7.10 (5.30) 4.95 (3.18) 2.04 (1.83) 891.73 (469.23) 41.34% (130.68%) 4.64 (3.76) 5.25 (5.75)
Quartile 2 8.48 (4.00) 8.29 (7.13) 4.90 (2.61) 2.73 (4.07) 1177.25 (1234.33) 64.08% (207.21%) 5.74 (7.97) 5.96(7.41)
Quartile 3 8.10 (4.72) 8.53 (8.23) 5.65 (5.78) 1.87 (1.84) 985.33 (722.96) 33.37% (156.28%) 6.23 (5.25) 6.30 (7.48)
Quartile 4 9.73 (5.82)* 10.51 (8.15) 6.54 (7.59) 2.28 (2.00) 1279.29 (826.22) 22.42% (101.64%) 7.45 (4.92) 8.27 (7.61)
On Reservation:
Yes 8.24 (5.67) 8.16 (6.92) 5.51 (5.82) 2.20 (2.58) 1068.96 (899.48) 33.11% (144.09%) 6.04 (5.72) 6.07 (6.61)
No 8.52 (5.48) 10.58 (8.69) 5.74 (3.00) 2.33 (2.55) 1165.56 (700.34) 61.63% (173.50%) 6.19 (5.59) 7.41 (8.65)
Smoking:
Non-Smokers 8.19 (5.13) 8.14 (7.48) 6.23 (7.50) 1.96 (2.24) 1068.44 (820.56) 19.74% (119.50%) 6.23 (4.59) 6.91 (6.01)
Occasional Smokers 7.66 (4.46) 9.53 (9.34) 4.46 (2.38) 3.26 (3.52) 1288.17 (1524.86) 54.68% (168.36%) 4.40 (5.57) 6.29 (9.12)
Daily Smokers 8.79 (6.41) 8.96 (6.63) 5.29 (3.42) 2.13 (2.46) 1032.44 (513.86) 49.33% (169.62%) 6.66 (6.51) 6.65 (6.78)
Drinking:
Non-Drinkers 8.26 (4.59) 9.10 (7.77) 5.37 (4.27) 2.07 (1.97) 1141.12 (949.40) 34.97% (147.76%) 6.19 (4.46) 6.88 (7.27)
Drinking in Past Month 8.45 (7.04) 8.15 (6.73) 5.89 (6.77) 2.48 (3.34) 1017.31 (690.88) 48.26% (157.61%) 5.97 (7.30) 5.83 (6.94)
Open in a new tab
†

p < .10;

*

p < .05;

**

p < .01;

***

p < .001

Discussion

The purpose of this investigation was to examine the feasibility and outcomes of a home-based salivary cortisol collection protocol as part of a larger study evaluating the relationship between stress and T2D among AI adults. Indigenous communities disproportionately encounter psychosocial stressors that contribute to health inequities including T2D (American Psychological Association [APA], 2016). Efforts to more comprehensively understand the role of stressors and subsequent physiological stress on adverse health outcomes with AI communities is needed alongside recognition of possible wariness of research, particularly that involving biological samples.

The success of this study was facilitated by an essential process of creating trusting partnerships between communities and researchers, exemplary efforts of tribal community-based interviewers, and our overall CBPR approach. Interviewers exhibited high levels of professionalism and commitment to the project and implemented innovative compliance strategies above and beyond our study protocol. As examples, some interviewers called or sent reminder texts to participants on collection day, set alarms on participant smart phones to signal each collection time, used sticky notes in various places within participant homes as reminders, and set-up SSS forms and tubes on participants’ nightstands for ease of waking collection. Interviewer Melanie McMichael shared, “I would often write an example schedule to give them a visual of what the day’s collection should look like. I allowed the participant to pick the starting time of when they usually rise in the morning as the first time and wrote that on the cover sheet of the SSS forms, followed by writing the other three times. I also tried to make it clear that the times that the bottle is opened needs to correlate with the times written on the form.” Another interviewer, Tina Handeland, said that participants appreciated knowing that interviewers themselves were required to practice salivary cortisol collection procedures during training and that this experience was useful for explaining protocols. Interviewers worked very closely with university-based team members, especially the study coordinator. For instance, interviewers would often call from participants’ homes to talk through MEMS and SSS form discrepancies and determine the necessity of second chance samples. Several clinic-based team members and interviewers indicated that multiple participants inquired about the use of their saliva and/or DNA outside of the current study. Our clear and collaboratively defined salivary cortisol disposal plan, site-visits to the assay lab, and explicit indication in study brochures and consent forms that saliva would not be used for purposes beyond cortisol assessment assured participants of the safety of their biological specimens. Thus, future studies could enhance compliance by engaging community-based research workers and enhancing training to include some of these strategies employed by our team.

We strove for authentic collaboration with diligent attention to community outreach (e.g., newsletters about the project, reports to agencies and tribal councils, correspondence with study participants), regular dialogue with community leaders, and community-led decisions about treatment and disposal of collected samples (Hiratsuka et al., 2012). An important reflection of our CBPR orientation is the fundamental goal of using this research to identify strategies for improving community health. Written study materials and interviewer training emphasized informing participants of the importance of the study and implications for additional research and programs, and participants shared that they valued contributing to their community’s future health via study participation. Our journey together (we are indeed Gathering for Health) continues as we identify culturally relevant methods for identifying and intervening on stressors for T2D patients via community and family-based interventions (Blacksher et al., 2016).

Results suggest that home-based, community interviewer-involved protocols can result in valuable data and high participant compliance. Such naturalistic observation may enhance validity in that participants go about their daily lives with only brief, limited interruptions related to saliva collection/study protocols. In this study, 67% of all saliva samples were provided within 10 minutes of protocol instructions, 91% of participants provided at least one useable sample, and 79% provided four useable samples during a single day. These findings are comparable to those reported in prior research. For instance, in the Multi-Ethnic Study of Atherosclerosis (Golden et al., 2014), 57% of required samples were provided within 10 minutes of protocol instructions, and 86% of participants provided useable samples in National Study of Daily Experiences (Laudenslager et al., 2009). Our comparisons between more and less compliance to protocol suggests that noncompliance had limited impacts on observed cortisol indices in terms of extreme outliers or significantly different observed values. An exception was found in the case of outlier analyses: participants with outlier values on Interval 1 displayed significantly lower Early Morning Response values than did participants who adhered more closely to the protocol instructions. This finding makes sense considering that most variation in cortisol levels occurs in the minutes immediately after awakening, when protocol compliance is most important (Nicolson, 2008). These significant differences disappeared when examining percent change rather than change in raw values.

Participants also demonstrated good compliance to instructions not to eat, smoke, or drink close to sampling times (74.3% for all behaviors at all sampling times). Smoking and consuming caffeine prior to sample 1 were the only behaviors significantly related to corresponding cortisol indices (i.e., waking values, percent change from samples 1 to 2). Caffeine and nicotine consumption have been found to influence cortisol secretion in prior research (Steptoe et al., 2006; Lovallo et al., 1996); thus, we were somewhat surprised to find that these specific behaviors had limited relationships to the cortisol measures we generated. Additionally, the relationship between early morning smoking and waking values may be partially explained by delays in sampling, as participants who smoked before sample 1 appeared to be less compliant.

Participants taking prescription blood pressure medications had higher mean cortisol waking values than the rest of the study sample; however, when we applied a Bonferroni correction to this analysis the difference observed was no longer statistically significant. Furthermore, no other medication classes were significantly associated with cortisol values is notable: exclusion of participants taking medications would mean that nearly all (93.3%) of the AI adults in this study would be ineligible for participation. The AI “epidemic” of T2D is such that a representative study sample will include those living with comorbid conditions and taking multiple medications and thus would suffer from external invalidity, lack of feasibility, and outcomes of less relevance to real-world community conditions (Granger et al., 2009).

Our analyses also revealed several statistically significant trends in observed cortisol values by demographic statuses. First, the oldest participants displayed higher waking values than did the youngest participants. This finding suggests that cortisol values may increase with age – a finding that is generally consistent in existing literature (Nicolson, 2008). Our finding of higher waking values in men has been observed in some previous work, but our lack of significant differences in other cortisol indices (e.g., AUC) by gender are inconsistent with other work (Karlamangla et al., 2013) and signals need for additional research. Participants reporting the lowest income quartile for this sample displayed significantly lower AUCG values than those in the highest quartile, suggesting that lower income may lead to lower total cortisol output throughout the day. These findings are also difficult to compare to the existing literature, as findings from studies examining the role of SES on stress and cortisol have been mixed (Dowd et al., 2009).

These results should be interpreted with appropriate consideration of limitations and opportunities for future research. First, our results are based on a single day of sample collection and may have been influenced by noise or state factors that have potential to distort data (Thorn, Evans, Cannon, Hucklebridge, & Clow, 2011). Although we attempted to account for this possibility by measuring state factors at time of sample collection via the SSS form, a future goal includes examining within- and between-person variability in cortisol indices over time. Despite data triangulation to calculate protocol adherence (MEMS recorders, self-report, interviewer report), the reality of naturalistic observational studies means that unknown factors may influence actual compliance, sample collection, and participant reporting. We collected fewer samples than conventional in saliva cortisol research to reduce burden on participants. Future research may benefit from higher frequency of sampling during the first hour of the waking time. Additionally, because even moderate delays in sampling after waking may skew morning and CAR calculations (Smyth, Clow, Thorn, Hucklebridge, & Evans, 2013), morning cortisol values should be interpreted with caution as participants’ actual wake times were not monitored in the current study and some participants reported smoking and consuming caffeine before sample 1. Considering the high compliance exhibited by participants for other protocol requirements and the lack of significant differences in time of sample 1 collection, however, we believe that participants likely adhered well to the instructions to collect their first sample immediately upon awakening. Future research may benefit from utilizing electronic monitoring of wake times to identify sampling delays and examine compliance. Additionally, future studies should consider participant smoking status and caffeine use when creating protocol instructions and may benefit from including additional support for regular smokers and caffeine users (e.g., training interviewers to place collection materials near bedside for easy access in the morning). Protocol compliance results may be slightly inflated as first chance attempt data were not collected from the participants who were offered a second chance at sample collection (including the number of second chance attempts completed). It is reasonable to assume, however, that offering a second chance at collection resulted in increased compliance and accurate data, as participants who would normally need to be excluded due to noncompliance were instead offered a “do-over” to correctly follow protocol instructions. To achieve similar compliance in future studies, researchers should budget and account for second chance options. While the baseline response rate in this study was acceptable (67%), reasons for non-response or refusal are largely unknown and could include mistrust of biomedical research, a factor potentially impacting feasibility and compliance in this and other tribal settings. Participants in this study were given $50 for successful completion of cortisol collection procedures; thus, high adherence may be in part driven by this extrinsic motivator. Finally, our analyses revealed limited potential confounding influences on cortisol values; however, high protocol compliance may mean limited variability to detect how or if noncompliance to protocol timing or behaviors meaningfully influenced observed cortisol indices.

The use of naturalistic settings and methods to collect salivary cortisol allows for higher external validity and lower participant burden than lab settings, but often comes at the cost of lower internal validity. Even considering the limitations experienced in the current study, however, participants still demonstrated relatively high compliance to protocol. Our findings suggest that the protocol used here is a feasible method of obtaining cortisol data during a “typical day.” This protocol may benefit further from techniques to improve internal validity, particularly during early morning hours (e.g., electronic monitoring of wake time, training interviewers to highlight the importance of collecting the first sample immediately upon waking and before smoking, consuming caffeine, etc.). Stress biomarkers like cortisol may afford a culturally neutral way of identifying the impacts of stress on health and disease progression and serve as a tool for measurement triangulation or cross-validation to improve the state of stress measurement. Future goals are to examine within person variability/changes in HPA functioning over the course of our longitudinal study, observing how these patterns relate to T2D progression, and how cortisol relates to chronic and acute psychosocial stressors among AIs. This investigation of adherence to protocol, validity and possible confounding factors related to salivary cortisol values was a necessary and important step towards broader goals of understanding how psychosocial and physiological stressors impact T2D-related health for AIs. Ultimately, these advancements offer promise for culturally grounded, empirically informed community interventions to equip T2D patients and families with knowledge and tools for combatting the harmful effects of stress to achieve better health.

Lay Summary.

  • Authentic efforts for tribal community partnerships in research are critical to successfully implementing biological assessments with American Indians given legacies of research misconduct and mistrust

  • Our Community-Based Participatory Research with 5 tribes yielded high participant compliance to a home-based salivary cortisol collection protocol

  • Lack of compliance to salivary cortisol protocol and medication usage were not consistently associated with observed cortisol indices

Acknowledgements:

Whiteside Lab, Rowan Simonet

Funding Details

The Maawaji’ idi-oog Mino-ayaawin (Gathering for Health) project was funded by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Number DK091250 (M. Walls, PI). The contents of this manuscript are attributable to the authors and do not necessarily represent the viewpoints of the NIH.

Footnotes

Disclosure Statement

The authors have no conflicts of interest to report. This research involved human participants. All procedures were in accordance to the ethical standards of the University of Minnesota IRB, the National Indian Health Service IRB, and community approved protocols, as well as the 1964 Helsinki declaration and its later amendments. Informed consent was obtained from all participants in this study.

References

  1. al’Absi M (2018). Stress and addiction: When a robust stress response indicates resiliency. Psychosomatic Medicine 80, 2–16. DOI: 10.1097/PSY.0000000000000520 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. al’Absi M, Hatsukami D, & Davis G (2005). Attenuated adrenocorticotropic responses to psychological stress are associated with early smoking relapse. Psychopharmacology, 181(1), 107–117. DOI: 10.1007/s00213-005-2225-3 [DOI] [PubMed] [Google Scholar]
  3. al’Absi M, Hatsukami D, Davis G, & Wittmers L (2004). Prospective examination of effects of smoking abstinence on cortisol and withdrawal symptoms as predictors of early smoking relapse. Drug and Alcohol Dependence, 73(3), 267–278. DOI: 10.1016/j.drugalcdep.2003.10.014 [DOI] [PubMed] [Google Scholar]
  4. al’Absi M, Wittmers L, Erickson J, Hatsukami D, & Crouse B (2003). Attenuated adrenocortical and blood pressure responses to psychological stress in ad libitum and abstinent smokers. Pharmacology Biochemistry and Behavior, 74(2), 401–410. DOI: 10.1016/S0091-3057(02)01011-0 [DOI] [PubMed] [Google Scholar]
  5. American Psychological Association (2016). Stress in America: The impact of discrimination. Stress in America Survey. [Google Scholar]
  6. Aneshensel CS (1992). Social stress: Theory and research. Annual Review of Sociology, 18(1), 15–38. DOI: 10.1146/annurev.so.18.080192.000311 [DOI] [Google Scholar]
  7. Blacksher E, Nelson C, Van Dyke E, Echo-Hawk A, Bassett D, & Buchwald D (2016). Conversations about community-based participatory research and trust:” We are explorers together”. Progress in Community Partnerships, 10(2), 305–309. DOI: 10.1353/cpr.2016.0039 [DOI] [PubMed] [Google Scholar]
  8. Born J, Hansen K, Marshall L, Mölle M, & Fehm HL (1999). Timing the end of nocturnal sleep. Nature, 397, 29–30. [DOI] [PubMed] [Google Scholar]
  9. Boyer BB, Dillard D, Woodahl EL, Whitener R, Thummel KE, & Burke W (2011). Ethical issues in developing pharmacogenetic research partnerships with American Indigenous communities. Clinical Pharmacology & Therapeutics, 89(3), 343–345. DOI: 10.1038/clpt.2010.303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Buchwald D, Mendoza-Jenkins V, Croy C, McGough H, Bezdek M, & Spicer P (2006). Attitudes of urban American Indians and Alaska Natives regarding participation in research. Journal of General Internal Medicine, 21(6), 648–651. DOI: 10.1111/j.1525-1497.2006.00449.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Burnette CE (2015). Historical oppression and intimate partner violence experienced by Indigenous women in the US: Understanding connections. Social Service Review, 89(3), 531–563. DOI: 10.1086/683336 [DOI] [Google Scholar]
  12. Center for Disease Control and Prevention. Department of Health and Human Services. (2017). National Diabetes Statistics Report, 2017: Estimates of diabetes and its burden in the United States. Retrieved from: http://www.diabetes.org/assets/pdfs/basics/cdc-statistics-report-2017.pdf
  13. Centers for Disease Control and Prevention. Department of Health and Human Services. (2011). National diabetes fact sheet: National estimates and general information on diabetes and prediabetes in the United States. Atlanta, GA: U.S. [Google Scholar]
  14. Dallman MF. (1993). Adaptation of the hypothalamic-pituitary-adrenal axis to chronic stress. Trends in Endocrinology and Metabolism, 4, 62–69. [DOI] [PubMed] [Google Scholar]
  15. Dowd JB, Simanek AM, & Aiello AE (2009). Socio-economic status, cortisol and allostatic load: A review of the literature. International Journal of Epidemiology, 38(5), 1297–309. DOI: 10.1093/ije/dyp277 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fekedulegn D, Andrew M, Burchfiel C, Violanti J, Hartley T, Charles L, & Miller D (2007). Area under the curve and other summary indicators of repeated waking cortisol measurements. Psychosomatic Medicine, 69(7), 651–659. DOI: 10.1097/PSY.0b013e31814c405c [DOI] [PubMed] [Google Scholar]
  17. Golden SH, Sánchez BN, DeSantis AS, Wu M, Castro C, Seeman TE, … Diez Roux AV. (2014). Salivary cortisol protocol adherence and reliability by sociodemographic features: The Multi-Ethnic Study of Atherosclerosis. Psychoneuroendocrinology, 43, 30–40. DOI: 10.1016/j.psyneuen.2014.01.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gracey M, King M (2009). Indigenous health part 1: Determinants and disease patterns. Lancet, 374(9683), 65–75. DOI: 10.1016/s0140-6736(09)60914-4 [DOI] [PubMed] [Google Scholar]
  19. Granger D, Hibel L, Fortunato C, & Kapelewski C (2009). Medication effects on salivary cortisol: Tactics and strategy to minimize impact in behavioral and developmental science. Psychoneuroendocrinology, 34(10), 1437–1448. DOI: 10.1016/j.psyneuen.2009.06.017 [DOI] [PubMed] [Google Scholar]
  20. Halpern CT, Whitsel EA, Wagner B, & Harris KM (2012). Challenges of measuring diurnal cortisol concentrations in a large population-based field study. Psychoneuroendocrinology, 37(4), 499–508. DOI: 10.1016/j.psyneuen.2011.07.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Harmon A (April 21, 2010). Indian tribe wins fight to limit research on its DNA. New York Times. April 21, 2010. Retrieved from: http://www.nytimes.com/
  22. Hiratsuka V, Brown J, & Dillard D (2012). Views of biobanking research among Alaska native people: The role of community context. Progress in Community Partnerships, 6(2), 131–139. DOI: 10.1353/cpr.2012.0025 [DOI] [PubMed] [Google Scholar]
  23. Hodge FS (2012). No meaningful apology for American Indian unethical research abuses. Ethics & Behavior, 22(6), 431–444. DOI: 10.1080/10508422.2012.730788 [DOI] [Google Scholar]
  24. Hussain-Gambles M, Atkin K, & Leese B (2004). Why ethnic minority groups are underrepresented in clinical trials: A review of the literature. Health & Social Care in the Community, 12(5), 382–388. DOI: 10.1111/j.1365-2524.2004.00507.x [DOI] [PubMed] [Google Scholar]
  25. Huttlin E, Allen A, Tosun N, Allen S, & al’Abi M (2015). Associations between adrenocortical activity and nicotine response in female smokers by menstrual phase. Addictive Behaviors,. 50, 135–139. DOI: 10.1016/j.addbeh.2015.06.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jacobs N Nicolson NA, Derom C, Delespaul P. van Os J, & Myin-Germeys I (2005). Electronic monitoring of salivary cortisol sampling compliance in daily life. Life Sciences, 76, 2431–2443. [DOI] [PubMed] [Google Scholar]
  27. Jiang L, Beals J, Whitesell NR, Roubideaux Y, Manson S, & The AI-SUPERPFP Team (2008). Stress burden and diabetes in two American Indian reservation communities. Diabetes Care, 31(3), 427–429. DOI: 10.2337/dc07-2044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Joseph JJ & Golden SH (2017). Cortisol dysregulation: The bidirectional link between stress, depression, and type 2 diabetes mellitus. Annals of the New York Academy of Sciences, 1391, 20–34. DOI: 10.1111/nyas.13217 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kaholokula JK, Grandinetti A, Keller S, Nacapoy AH, & Mau MK (2012). Association between perceived racism and physiological stress indices in Native Hawaiians. Journal of Behavioral Medicine. 35(1), 27–37. DOI: 10.1007/s10865-011-9330-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Karlamangla AS, Friedman EM, Seeman TE, Stawksi RS, & Almeida DM (2013). Daytime trajectories of cortisol: Demographic and socioeconomic differences—findings from the National Study of Daily Experiences. Psychoneuroendocrinology, 38(11), 2585–2597. DOI: 10.1016/j.psyneuen.2013.06.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kirschbaum C, Kudielka BM, Gaab J, Schommer NC, & Hellhammer DH (1999). Impact of gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamus-pituitary-adrenal axis. Psychosomatic Medicine, 61(2), 154–162. DOI: 10.1097/00006842-199903000-00006 [DOI] [PubMed] [Google Scholar]
  32. Kudielka BM, Broderick JE, & Kirschbaum C (2003). Compliance with saliva sampling protocols: electronic monitoring reveals invalid cortisol daytime profiles in noncompliant subjects. Psychosomatic Medicine, 65(2), 313–319. DOI: 10.1097/01.PSY.0000058374.50240.BF [DOI] [PubMed] [Google Scholar]
  33. Laudenslager ML, Noonan C, Jacobsen C, Goldberg J, Buchwald D, Bremner JD, Vaccarino V, & Manson SM (2009). Salivary cortisol among American Indians with and without posttraumatic stress disorder (PTSD): gender and alcohol influences. Brain, Behavior, Immunity, 23(5), 658–662. DOI: 10.1016/j.bbi.2008.12.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Laudenslager ML, Noonan C, Jacobsen C, Goldberg J, Buchwald D, Bremner JD, … Manson SM (2009). Salivary cortisol among American Indians with and without posttraumatic stress disorder (PTSD): Gender and alcohol influences. Brain, Behavior, and Immunity, 23(5), 658–662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lovallo W, Al’Absi M, Blick K, Whitsett T, & Wilson M (1996). Stress-like adrenocorticotropin responses to caffeine in young healthy men. Pharmacology, Biochemistry, Behavior, 55(3), 365–369. DOI: 10.1016/S0091-3057(96)00105-0 [DOI] [PubMed] [Google Scholar]
  36. Lundberg U & Frankenhaeuser M (1980). Pituitary-adrenal and sympathetic-adrenal correlates of distress and effort. Journal of Psychosomatic Research, 24(3), 125–130. DOI: 10.1037/e576762009-001 [DOI] [PubMed] [Google Scholar]
  37. McEwen BS (2008). Central effects of stress hormones in health and disease: Understanding the protective and damaging effects of stress and stress mediators. European Journal of Pharmacology, 583(2-3), 174–185. DOI: 10.1016/j.ejphar.2007.11.071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Nicolson NA (2008). Measurement of cortisol In Luecken LJ, & Gallo LC (Eds.), Handbook of Physiological Research Methods in Health Psychology (pp. 37–74). California: Sage Publications. [Google Scholar]
  39. Pearlin LI, Menaghan EG, Lieberman MA, & Mullan JT (1981). The stress process. Journal of Health and Social Behavior, 22(4), 337–356. DOI: 10.1177/2156869315585388 [DOI] [PubMed] [Google Scholar]
  40. Pember M (2010). American Indians grow wary of genetics research. Diverse: Issues in Higher Education. Retrieved from: https://diverseeducation.com/article/13905/ [Google Scholar]
  41. Preussner J, Kirschbaum C, Meinlschmid G, & Hellhammer D (2003). Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change. Psychoneuroendocrinology, 28(7), 916–931. DOI: 10.1016/S0306-4530(02)00108-7 [DOI] [PubMed] [Google Scholar]
  42. Rudolph D, & McAuley E (1998). Cortisol and affective responses to exercise. Journal of Sports Sciences, 16(2), 121–128. DOI: 10.1080/026404198366830 [DOI] [PubMed] [Google Scholar]
  43. Saxbe D (2008) A field (researcher’s) guide to cortisol: tracking HPA axis functioning in everyday life. Health Psychology Review, 2(2), 163–190. DOI: 10.1080/17437190802530812 [DOI] [Google Scholar]
  44. Smyth N, Clow A, Thorn L, Hucklebridge F, & Evans P (2013). Delays of 5-15 min between awakening and the start of saliva sampling matter in assessment of the cortisol awakening response. Psychoneuroendocrinology, 38, 1476–1483. [DOI] [PubMed] [Google Scholar]
  45. Smyth N, Thorn L, Hucklebridge F, Evans P, & Clow A (2015). Detailed time course of the cortisol awakening response in healthy participants. Psychoneuroendocrinology, 62, 200–203. [DOI] [PubMed] [Google Scholar]
  46. Steptoe A, Cropley M, Griffith J, & Kirschbaum C (2000). Job strain and anger expression predict early morning elevations in salivary cortisol. Psychosomatic Medicine, 62, 286–292. [DOI] [PubMed] [Google Scholar]
  47. Steptoe A, & Ussher M (2006). Smoking, cortisol and nicotine. International Journal of Psychophysiology, 59(3), 228–235. DOI: 10.1016/j.ijpsycho.2005.10.011 [DOI] [PubMed] [Google Scholar]
  48. Thorn L, Evans P, Cannon A, Hucklebridge F, & Clow A (2011). Seasonal differences in the diurnal pattern of cortisol secretion in healthy participants and those with self-assessed seasonal affective disorder. Psychoneuroendocrinology, 36, 816–823. [DOI] [PubMed] [Google Scholar]
  49. Tovar M, Coser P Jr, & Leary M. (2014, July 3). More than meets the IRB–American Indian and Alaska Native (AIAN) research ethics: Considering the Havasupai Tribe case [Episode 7], HRPO Podcasts. Podcast retrieved from: https://digitalcommons.wustl.edu/hrpopods/7/ [Google Scholar]
  50. Turner RJ (2013). Understanding health disparities: the relavence of the stress process model. Social and Mental Health, 3(3), 170–186. DOI: 10.1177/2156869313488121 [DOI] [Google Scholar]
  51. Walls ML, Sittner KJ, Aronson BD, Forsberg AK, Whitbeck LB, & al’Absi M (2017). Stress exposure and physical, mental, and behavioral health among American Indian adults with type 2 diabetes. International Journal of Environmental Research and Public Health, 16; 14(9), 1074 DOI: 10.3390/ijerph14091074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wheaton B (1994). Sampling the stress universe In Avison WR, & Gotlib IH, (Eds.), Stress and mental health (pp. 77–114). Boston, MA: Springer. [Google Scholar]
  53. Whitbeck LB, Adams GW, Hoyt DR, & Chen X (2004). Conceptualizing and measuring historical trauma among American Indian people. American Journal of Community Psychology, 33(3–4), 119–130. DOI: 10.1023/B:AJCP.0000027000.77357.31 [DOI] [PubMed] [Google Scholar]
  54. Williams DR, Priest N, & Anderson NB (2016). Understanding associations among race, socioeconomic status, and health: Patterns and prospects. Journal of Health Psychology, 35(4), 407–411. DOI: 10.1037/hea0000242 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES