Introduction

Spatial epidemiology is a fast-growing field, a modern adaptation of the established methodology of geographic analysis, which dates to the 18^th century [1]. Traditionally, epidemiologic research focuses on person and time, with little attention to the implications of place and space due to project design, data limitations, etc. [2]. However, the built (e.g., proximity and density of alcohol and tobacco retail outlets in a neighborhood) and social (e.g., crime rates, racial composition of a neighborhood, etc.) environments are critical factors affecting health behaviors and health outcomes [3, 4]. Many studies have shown the importance of activity space and residential context on the incidence and prevalence of risk behaviors among populations vulnerable to health disparities in the United States [5–12], including among sexual minorities (those identifying as lesbian, gay, bisexual, queer, or other non-heterosexual identities) [13, 14].

Over recent decades, the application of technology has enriched the complexity, sophistication, and utility of spatial epidemiology [15–17]. The development of geographic information systems (GIS) has provided a powerful capability for researchers to rapidly examine spatial patterns and processes. GIS data and spatial methods have been used in many epidemiologic applications such as disease mapping, rate smoothing, and spatial modeling [18]. With the rapid increase in computing power, advancement in technologies have extended GIS capacities. For example, the availability of low-cost personal Global Positioning Systems (GPS) has made their use increasingly popular—a development that has been discussed in many recent health research studies [9, 19, 20]. GPS tracking generates geographic data, which allows for near continuous monitoring of mobility (i.e., activity paths) across space. GPS technology not only allows researchers to investigate the impact of spatial contexts when wedded with GIS data, but also to consider the timing and the duration of exposure within spatial contexts [21, 22].

GPS technology has been widely used among diverse populations [7, 9, 14, 23–25] to assess contextual determinants and correlates of different health behaviors such as health-related product purchasing behaviors, physical activity, dietary intake, and smoking behaviors [10, 12, 23, 26]. However, some studies indicate that participants have concerns regarding confidentiality and privacy when GPS data are collected that may affect the acceptability of these procedures [27, 28]. This may especially be the case for population subgroups that have been historically marginalized or mistreated. This indicates the need to assess the feasibility and acceptability of wearing GPS devices among sexual minority individuals. Additionally, there is a need to examine whether or not GPS technology can be readily paired with data on stress and negative emotion. Pairing GPS and affective data has the benefits of better understanding how the external context may affect internal experiences.

Historically, data measuring experiences of mood and stress have been collected via paper and pencil questionnaires that rely on the retrospective recall of thoughts and feelings experienced over a specified timeframe (e.g., previous week, month, year, etc.). Recognizing the limitations of memory recall and the potential for response bias, there has been a push for objective methodological innovations in naturalistic studies where participants navigate their environment while collecting biomarker samples (e.g., saliva)–an objective neuroendocrine assessment–that can be paired with the real time measurement of mood at several timepoints across the day [29]. Leveraging Ecological Momentary Assessment (EMA) methodology to collect mood data on a smartphone–or equivalent device–can better capture real time fluctuations in self-reported mood that can be paired with neuroendocrine reactivity (e.g., cortisol) while minimizing measurement error [30].

Many studies have shown that salivary cortisol represents a nonintrusive assessment of physiological stress reactivity that has been cross validated with serum cortisol in naturalistic settings, which allows for the simultaneous measurement of diurnal cortisol and the associated time-varying fluctuation in physiologic stress responses across a specified timeframe [31–44]. More specifically, higher levels of salivary cortisol were associated with increased levels of self-reported stress and negative emotions [31, 32]. Similar results were observed by other studies conducted with sexual minority participants [37–40]. Prior work suggested there were no differences in diurnal cortisol rhythms across sexual orientation groups [37], possibly due to the exhibition of physiologic resilience of the sexual minority participants, whereas differences were observed by race/ethnicity [38]. While there is a dearth in the amount of neuroendocrine studies that have focused on this population subgroup [37–44], a recent study (Growing Up Today Study, GUTS) by Austin and colleagues found a low response rate of young lesbian, gay and bisexual individuals providing saliva samples [37]. Thus, the literature would benefit from additional feasibility studies examining saliva sample collection among sexual minorities in naturalistic environments, particularly when such protocols include other procedures that may impact a person’s choice to participate.

Attitudes toward sexual minorities have changed remarkably in the United States during the past three decades; for example, the National Institute of Health now recognizes sexual minorities as a health disparities group and there is greater appreciation for understanding their collective experience that oftentimes diverge from heterosexual norms [45, 46]. For example, sexual minorities experience higher risks for many negative physical and mental health outcomes (e.g., chronic disease, human immunodeficiency virus (HIV) infection, and anxiety disorders [47–51]), and non-health promoting behaviors (e.g., smoking, heavy drinking, and illicit drug use [50]) relative to heterosexuals, with some variation between sexual minority sub-groups. Chronic exposure to stress represents a transdiagnostic vulnerability that is linked to health disparities experienced by marginalized communities [50–56]. Increased stress levels may be triggered by many factors including incessant experiences of discrimination and homophobia. Growing evidence suggests sexual minorities are exposed to more stress than their heterosexual counterparts [52–55]. The health disparities experienced among sexual minorities has been conceptualized through the minority stress theory that suggests that discrimination, stigma, and prejudice create a stressful social environment, which can in turn cause mental or physical health issues [56]. Specifically, Meyer’s minority stress model posits that sexual minorities are exposed to myriad and detrimental stressors related to social stigma and discrimination, which in turn contribute to the higher prevalence of health-risk behaviors and associated disorders and diseases [57–60]. A meta-analysis [56] found lesbian, gay, and bisexual individuals were more likely to have a mood, anxiety, or substance abuse disorder in their lifetime when compared to heterosexual individuals.

There are published studies that have collected salivary biomarkers among sexual minorities in a naturalistic setting [37–40, 61]; however, we were unaware of studies that have specifically combined the use of GPS and saliva collection among this group. To address this gap in the literature, we assess the feasibility of coupling GPS technology with objective salivary biomarker collection in this population. Tracking activity paths using GPS is a novel approach that provides researchers with an objective assessment of physical location where people interact with the built and social environmental contexts without the need for retrospective recall. GPS technology has the capacity to assess spatial mobility and polygamy in real time to understand the characteristics of participant activity spaces across contexts of interest (e.g., built or social environmental exposures) through the automated collection of latitude and longitude data [62–66]. Likewise, assessing stress physiology using inexpensive saliva collection protocols provides an objective measure of neuroendocrine reactivity to stress that is easy to collect and less invasive than to other modalities (i.e., serum via venipuncture). Moreover, stress-related steroid hormones like cortisol could then be paired with real time objective (e.g., GPS) and self-report data (e.g., perceived stress, mood, etc.) to provide a more comprehensive insight into the experience of sexual minority groups grappling with chronic exposure to stress in their day-to-day lives.

It is imperative that sexual minorities–a health disparity population–be included in cutting-edge research that can identify novel mechanisms for understanding how activity space couples with mood to have a deleterious effect on stress dysregulation–a known risk factor for chronic diseases. Integrating inexpensive GPS technology with a salivary biomarker collection protocol provides an innovative opportunity for quantifying stress physiology in real time as sexual minorities navigate the real-world environment. Despite increasing acknowledgements of the importance of the built and social environments for health, the feasibility of pairing the use of a GPS device with salivary biomarker collection in these communities remains poorly understood.

The purpose of this study was to assess the feasibility of recruiting adults identifying as sexual minorities to participate in a study involving: (1) location and wear-time data using a GPS monitor in their natural setting; (2) collection of saliva-based biomarker samples at specified times throughout a single day; and (3) assessment of participant characteristics that may inform within-group variation in the acceptability and feasibility of such methodological protocols.

Materials and methods

Results

Overall

Of the 124 participants, 70.97% (n = 88) participants had complete GPS and saliva data; 8.06% (n = 10) participants had complete GPS data but not saliva data; 8.06% (n = 10) participants had complete saliva data but not GPS data; and 12.90% (n = 16) participants did not have either GPS or saliva data. Of the 124 participants, the average age was 36.59 (±14.86) years and 50% were cisgender male; 52.42% were monosexual; 39.84% reported being Black/African American; 36.59% reported being non-Latino White; and 65.32% reported a high school education or less education (Table 1). No significant differences were found in participants’ characteristics across the aforementioned 4 GPS/saliva compliance groupings.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of participants by different types of data completion on participants providing entire 6 saliva samples including those who violated directions.

https://doi.org/10.1371/journal.pone.0250333.t001

More than three-fourths of the participants (n = 98, 79.03%) provided all 6 saliva samples. There were 4 (3.23%), 2 (1.61%) and 1 (0.81%) participants who provided only the initial one, the initial two, or three saliva samples, respectively. There were 19 participants (15.32%) who did not provide any saliva samples. Overall, 24 participants potentially violated the directions for saliva collection and were further excluded. Of the 100 participants who did not violate directions, there were no significant differences in socio-demographic variables across the 4 GPS/saliva compliance groups (Table 2). Additionally, there were no socio-demographic differences between violators (24 = 124–100) and non-violators (74 = 5+69) with the exception of age. The violators were significantly younger than non-violators (28.25 vs. 38.18, p = 0.0057) (Table 3).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of participants by different types of data completion on participants providing entire 6 saliva samples without directions violated.

https://doi.org/10.1371/journal.pone.0250333.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics of participants by providing entire 6 saliva samples without directions violated versus those with directions violated.

https://doi.org/10.1371/journal.pone.0250333.t003

Of 124 participants, 16 participants (12.90%) did not provide any useable GPS data. There were 108 (87.10%) participants who provided at least one day of GPS data and 98 (79.03%) with at least 4 days of data. GPS wear day values varied among the 108 participants with GPS data (M = 7.33; SD = 2.51; Range: 1–16). Fewer than half (n = 50, 40.32%) wore the GPS device one week or less. Slightly over a third of participants wore a GPS device for 8 days (n = 41, 33.06%). Of those with GPS data, there were 17 participants (13.71%) who had GPS data for a period of 9 or more days. Except 2 participants who wore the GPS at least 15 days, participants who wore the GPS 7 days had the highest average daily valid GPS points (i.e., latitude and longitude values) of 453.74 (SD = 1893.28), followed by those who wore the GPS 8 days who had mean of 140.05 (SD = 450.39) valid GPS points per day (Table 4).

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. GPS wear time and average daily valid GPS points by different types of data completion.

https://doi.org/10.1371/journal.pone.0250333.t004

Discussion

To the best of our knowledge, no study has investigated the feasibility of combined use of GPS information and saliva collection among sexual minorities. The incidence of health disparities among individuals identifying as sexual minorities has gained increasing scholarly and theoretical attention [34, 35]. Sexual minorities comprise diverse sexual identities, orientations and practices but as a group tend to be exposed to higher degree of stress from social stigma and discrimination than heterosexuals, resulting in a greater potential for adverse health behaviors and health outcomes [48–52]. Results of the current pilot study indicate that research protocols pairing GPS data collection with participant-collected salivary cortisol data are feasible among this population subgroup as evidenced by high percent of GPS valid days (79.03% had greater than 4 days) and low proportion of completed saliva samples involving direction violations (0.17% - 2.89% of completed saliva samples provided by study team, see Materials and Methods section for more details). Sexual minority individuals provided rich contextual data on their activity paths by wearing a GPS device and on their stress physiology patterns by providing the saliva samples. The GPS device provided automatic collection of data that could be easily synchronized with secondary GIS data on the built environment, and with biological markers of stress as assessed using saliva samples collected with minimal burden and disruption of typical daily routines on participants. The combined used of GPS and saliva samples collection would allow for analyses in examining how changes in the social environment may impact individual physiological stress processes. Although we did not also collect EMA data regarding daily or moment-to-moment mood fluctuations in this protocol, future work could extend into this area to further advance understanding of the interplay between stress-related steroid hormones like cortisol, built and social environmental exposures, real-time self-reported mood data to provide a more comprehensive insight into the experience of sexual minority groups grappling with chronic exposure to stress in their day-to-day lives. GPS data could be collected by the smartphone or via smartphone paired with a dedicated GPS device in order to assess data reliability and possibly remedy any gaps in data that may emerge.

The use of GPS devices has been widely accepted in the general population (typically these are embedded in smartphones). However, relatively few studies employing a GPS device have been conducted in community samples targeting sexual minorities. Despite the apparent convenience of the GPS device wearing and saliva sample collection over the course of the day, it is also important to know about the acceptability of these methods of data collection within a research protocol. Prior work suggests that acceptability can be gleaned by testing the study procedure/protocols, as a dimension of feasibility [77, 78]. Given this, our data could possibly suggest that methods were acceptable to study participants as indicated by high compliance with the protocol/procedures; however, future studies might incorporate feedback and thematically-analyzed qualitative pre- and/or post-interviews to further assess attitudes toward the saliva sample collections and GPS device data collection that may affect use. This would help provide insights that address the complexity of feasibility and potential acceptability of simultaneous use of GPS device and saliva collection in sexual minority communities.

We found that the majority of the participants in this study wore the GPS device, a potential indicator of protocol acceptability, with no significant differences on socio-demographic variables by GPS/saliva compliance groupings, which is consistent with previous research [14, 79–81]. While our study showed 79.03% of participants had at least 4 days GPS data, it is challenging to compare our results with those of other studies due to the longer GPS sampling epoch set up in this study. For example, Duncan and colleagues (2016) showed the feasibility and acceptability of using GPS methods among a sample of young men who have sex with men (YMSM) in New York City. They reported having at least one hour of GPS data (GPS device was set to record information in 30-second intervals) for 1 day, and 84% of their sample had at least 1 hour on 7 or more days for those YMSM (n = 75, 92% 22 or 23 years old) with any GPS data. Another study applied a similar GPS protocol, allowing 10-second epochs with a two-week monitoring period [82]. The high monitoring frequency enhanced the overall quality of the GPS data, and the comparatively longer collection period captured more variations of travel behaviors of the participants [82]. Nevertheless, an important conclusion from these studies—including present study—is that this form of data collection is feasible for research participants from diverse sexual minority groups.

In the current study, there was wide variability in the amount of GPS data that were collected per participant. One obvious contributor was the research design that did not require a specified duration of days, which is a potential limitation of the present study. However, other factors may also be relevant here, including satellite signal loss, atmospheric conditions, and interference from terrestrial objects (multipath error) contributing to missing coordinates [83]. Additionally, it is possible that the device was not worn regularly or daily due to competing interests and activities, including that some participants may have regarded the device as noticeable or cumbersome in some situations (e.g., while exercising, in their place of employment, etc.). Finally, it is worth mentioning that all participants who were screened eligible for participation in this study chose to participate and were not deterred by study procedures, including the impact of GPS data collection on their privacy. However, generally speaking, individuals in historically marginalized groups like sexual minority persons, may possess a heightened sense of intrusion into their privacy from the tracking of their locations [84, 85]. Therefore, future studies with different samples may see greater variability in the acceptability of these procedures. Moreover, results may not generalize to studies that prescribe a “minimum wear time” for the GPS monitoring device inasmuch as that may present additional burden to the participant. This is an important limitation as a recent study recommended 14 wear days for activity path studies, though fewer days may be required depending on the way activity space is measured (e.g., kernel density), environmental attributes of interest to the study, and activity-space representation of the data for analyses [85].

This study also successfully employed saliva collection among a diverse group of adult sexual minorities. Compared to other types of biological samples, saliva is relatively easy to gather. Of 124 participants, more than three-quarters of participants successfully collected 6 entire saliva samples. When accounting for the return of full 6 samples without any violations (e.g., eating food within 30 minutes prior to the collection), compliance dropped from 79.03% of the sample to 58.06% of the sample. Nevertheless, prior studies have required <6 daily saliva samples for analysis (e.g., Cook et al. [38]), suggesting that remaining data may still be valid for analysis after eliminating potential violations without necessitating the listwise deletion of the subject’s remaining saliva data output (e.g., cortisol values).

Other studies among subgroups of the sexual minority community have yielded similarly promising results. For example, in a recent saliva and hair collection feasibility study exploring HIV infection and chronic stress among a sample of 10 transwomen in Puerto Rico, participants were all willing to provide a one-time saliva sample with the assistance of the research team [41–44]. Saliva sample self- or researcher-collection has also been used among young sexual minority men [84, 86], and no problems or concerns were documented in obtaining multiple saliva samples over a longer period of time (e.g., 4 saliva samples each day over the course of 5 consecutive days in a study by Cook et al. [38]). However, it is noteworthy that our study excluded pregnant and lactating women as well as individuals taking medications that could not be postponed during the saliva collection protocol. Thus, feasibility data are limited to individuals who were eligible and/or who agreed to these procedures. Overall, we had 18 ineligibilities (164/182 = 90.1%) with 4 eligible individuals not interested in participation after they heard more about the study. Additionally, the accuracy of cortisol values assayed from saliva samples should be used with caution especially for those participants who reported certain behaviors (e.g., tobacco use and drinking) or non-compliance of saliva collection process (e.g., freeze the sample immediately). However, this represented a relatively low proportion of the completed saliva samples provided to the study team by participants (0.17% - 2.89%).

This feasibility study has limitations. First, the sample was relatively small overall—though large for a feasibility study on a historically marginalized group—and from a single city. The relatively small sample and unique participant characteristics contributed to the possibility that findings cannot be generalized other or all sexual minority adult groups residing in other areas. However, as the principal goal of the study was to determine feasibility, we do not make claims to generalizability. Second, it was not possible to disaggregate the participants’ GPS data by the number of hours participants wore the GPS. Third, no qualitative data regarding the acceptability and perception of GPS and saliva collection were collected from the participants after the assessment period. Such data could have allowed us to assess the burden on sexual minority participants of wearing a GPS or collecting saliva samples. Present work examined compliance with procedures relevant to a convenience sample that consented and enrolled in the study; results cannot determine if the protocol was fully acceptable to participants or that they had no concerns about compliance with it. Fourth, the non-compliance of wearing a GPS device throughout the day could skew the findings in terms of constructing the activity space. Finally, the sexual minority community comprises a diversity of gender and sexual identities; our inclusion criteria was not restrictive in this sense and although we report here on “sexual minorities” generally, it is important to recognize that there may be within-group identification differences (not assessed here) that affect the protocol acceptability or feasibility. Despite these limitations, this study provides important preliminary evidence supporting the combined use of GPS devices and self-collection of saliva samples among individuals identifying as sexual minorities.

The study’s strengths include the unique sample, diverse racial/ethnic composition, inclusion of cisgender men and women and transgender individuals, and the distinctive methodology of the combined use of a GPS device and saliva collection. The simultaneous use of GPS monitors and saliva collection to assess sexual minority individuals’ location and stress biomarkers will allow scientists to investigate the complex effects of the built and social environments and stress. Our findings may reflect characteristics of the sample, which reflects a limited range of population and geographic representation. Further research should investigate the extent to which similar findings might be obtained for sexual minorities in other geographical areas.

Conclusion

This study demonstrated the feasibility of simultaneous use of a GPS device and collection of saliva samples. GPS methods may be used to measure mobility and environmental exposure, while collected saliva samples may be used for examining stress physiology from diurnal patterns of salivary cortisol. Given the health disparities faced by some sexual minority groups, this feasibility study shed light on the potential use of GPS and saliva to enhance the understanding of the mechanisms underlying the complex interaction between environmental contexts and stress among sexual minority adults. Future research that are powered to investigate causal mechanisms underlying stress dysregulation and the social determinants of health in sexual minority communities is warranted.

Supporting information

Acknowledgments

The authors wish to thank Lauren Lee and Surya Chinamuthevi for their assistance on this manuscript. The first-author obtained permission from Ms. Lee and Mr. Chinamuthevi to name them in the acknowledgements.