1. Introduction
A range of factors shape the well-being of Latinx farmworkers in the United States (US). Their work is often contingent and low-paid [
1,
2,
3]. They labor in, and live in, rural places with fewer resources [
4]. They have limited access to insurance and physical and mental health care [
5], and face well-documented occupational health and safety risks [
2,
6]. Family separation, social hierarchies, insufficient social networks and other forms of social suffering also influence Latinx farmworker well-being [
1,
7,
8,
9]. Further, immigration politics have been central to shaping farmworker livelihoods in the US, and immigration status affects access to federally-funded safety nets which can improve health and welfare [
1,
10].
Collectively, these factors create an environment in which Latinx farmworkers are likely to suffer—physically, socially and psychologically. However, another form of marginalization often goes unrecognized in consideration of the experience of Latinx farmworkers: gender. In recent decades, there has been a marked increase in the proportion of women in the US agricultural workforce [
11], growing from 25% in 1989 to 32% in 2016 [
12,
13]. However, there is a limited body of research that examines the experiences of women farmworkers, particularly Latina farmworkers [
14]. These workers deserve attention not only because their numbers are increasing across the agricultural industry, but also because the additional marginalization associated with gender may lead to further disadvantage.
Women farmworkers are often more economically vulnerable, relative to men, in part because of gender-segregated work [
15]. Further, women farmworkers remain primarily responsible for childcare and household labor in the private sphere [
14]. Labor in the public and private spheres may conflict, decreasing well-being [
16], and Latina farmworkers are more likely to report work–family conflict and less support from supervisors [
15]. Women farmworkers are also more likely to experience elevated depressive symptoms, relative to men [
17], and the prevalence of depression is high in general among Latina women in farmworker families [
18,
19,
20].
Latina farmworkers also may experience increased barriers to occupational health and safety, such as ill-fitting protective equipment or reduced access to safety training [
21]. A study of 220 Latina women in farmworker families, a third of whom themselves worked in agriculture, found that just 50 participants (38%) reported always having access to proper safety equipment [
22]. These barriers to occupational health and safety may result in increased exposure to pesticides; in the US, two studies employing the Sentinel Event Notification System for Occupational Risks (SENSOR)-Pesticides program found that acute pesticide poisonings were nearly twice as prevalent in women working in agriculture compared to men [
23,
24]. Other studies have found higher levels of pesticide exposure among Latina nursery workers compared to controls [
25], and Latina farmworkers were found to have urinary pesticide metabolite levels above those measured in nationally representative samples, although these levels were not higher than Latina non-farmworkers in the same study [
26].
While Latina farmworkers may, therefore, face multiple, varied challenges to health and well-being, there remains more to be learned about their perspectives on these challenges. In this manuscript, we employ the concept of structural vulnerability to frame our understanding of the well-being of Latina farmworkers. Structural vulnerability asserts that existing social hierarchies shape the degree to which an individual is vulnerable within a community (or society), and calls attention to the political and social contexts within which people labor [
27]. The factors noted above, which may influence well-being among Latina farmworkers, are an outcome of structural vulnerability, and stem from inequalities based on gender, race and ethnicity, socioeconomic status, and space and place.
The connection between structural inequality and farmworker well-being has been made elsewhere, such as in the work of Holmes and Horton, who note the ways in which economic and social forces shape physical, social and psychological well-being among Latinx farmworkers [
28,
29]. However, less attention been given to the ways in which gender shapes well-being among Latina farmworkers. Yet as noted above, gender inequalities may be a key driver in understanding the vulnerabilities Latina farmworkers face, including those related to well-being [
30].
In the research presented here, we aimed to characterize factors contributing to and challenging well-being among a cohort of Latina farmworkers in Idaho. We developed a mixed-method approach to gather information on multiple dimensions of well-being, and further aimed to understand aspects of well-being that matter most to Latina farmworkers.
2. Materials and Methods
This study aimed to identify challenges to well-being, and to assess related social, cultural and workplace-related risk factors, among Latina farmworkers in Idaho. While we recognize the complexities of well-being, many scholars have argued for its usefulness as a tool for measuring quality of life [
31,
32,
33]. In assessing well-being and related risk factors, we employed an interdisciplinary approach including four components: surveys, focus groups, interviews, and biological monitoring. All aspects of this study were reviewed and approved by our university’s Institutional Review Board (IRB).
2.1. Recruitment, Enrollment and Consent
Our study sample included women aged 18 years or older who identified as Latina or Hispanic farmworkers during the screening process. Data collection occurred in either English or Spanish based on participant preference. All participants were recruited from southwest Idaho between October 2018 and June 2019. In-person recruitment occurred at local community-based organizations including Migrant and Farmworker Head Start programs, community events such as local festivals and health fairs, and churches and grocery stores. We committed in advance to sharing the findings of our study with community-based organizations, which we ultimately did in the form of a research report and through in-person presentations. Conducting research with a rural, vulnerable population has many challenges. In order to successfully recruit participants in this study, we worked in partnership with multiple trusted community organizations; included bilingual research staff in all study components; scheduled interviews and focus groups primarily on evenings and weekends; welcomed children during interviews, focus groups and survey completion; and, in some cases, provided childcare during study activities.
Consent was obtained separately for each form of data collection (surveys, focus groups, interviews, and biological monitoring, each discussed below). While all participants completed the survey component of the study, participants were allowed to opt in or out of the additional components. Grocery store gift cards ranging in value from $10 to $25 compensated participation in each component. The survey coversheet stated “If you agree to participate in this survey, please turn the page to begin,” and survey completion thus indicated consent to participate in this aspect of the research. No identifiers were collected as part of the survey. Following survey completion, women were asked whether they were willing to be contacted for participation in additional study components. Those who provided informed consent for additional components also provided their name and contact information (which was then linked to their survey IDs) to schedule future study procedures. All identifying information was coded and stored in a locked-file cabinet per IRB protocol.
2.2. Surveys
The survey included six domains of inquiry: sociodemographics; food security and food access; housing conditions; social isolation; access to medical care; and occupational hazards. Survey items were based on several previously validated instruments, including the National Agricultural Workplace Survey [
34], surveys of housing conditions and social isolation among farmworkers in the southeastern US [
35,
36], as well as a survey previously developed to assess food quality and availability among Latina farmworkers in Idaho [
14]. This survey was in paper format (English and Spanish), was completed individually by the respondent, and required between 15 and 35 min to complete.
2.3. Focus Groups
Study participants who indicated a willingness to be contacted for participation in focus groups were called within a week of survey completion. Focus group participants were also recruited via snowball sampling and flyers, which were posted in grocery stores and housing complexes proximate to farmworker community housing and shared among community partners. Five focus groups were held at community centers and restaurants in southwestern Idaho between January and May 2019. Focus group participants who had not previously completed the study survey were asked to do so during the focus group itself.
Focus groups were primarily conducted in Spanish and lasted between 60 and 120 min. A Spanish-speaking member of the research team attended each focus group to facilitate discussion with non-English speakers. Researchers also took detailed notes and/or audio recorded the sessions, and recordings were transcribed and translated to English. Discussion occurred through guided facilitation around participants’ own definitions of well-being and the dimensions of well-being of greatest concern and importance. Facilitated discussion also centered around how farm work influences well-being, and the strategies that farmworkers use to improve well-being. Participants were also asked to identify the relative importance of multiple aspects of well-being and how well-being varies seasonally in both facilitated discussion, and through activities. We provided meals for participants and any family members in attendance.
2.4. Interviews
Study participants willing to take part in the interview component of the study were contacted by phone to schedule the interviews. These semi-structured interviews occurred between March and June of 2019, lasted between 45 and 90 min, and occurred at a location of the participants’ choosing. Interview participants were asked a range of questions about their experiences with farm work, including benefits and challenges. Interviews were conducted in English or Spanish, based on participant preference. All interviews were audio-recorded, transcribed and translated as needed.
2.5. Biological Monitoring
Study participants who indicated a willingness to participate in the biological monitoring component of the study were called within a week of survey completion; others were approached directly during focus groups. Each participant was asked to provide two urine samples to reflect exposures both during the time of year when pesticides are not actively applied in Idaho agriculture (the “non-spray season”, defined as January 1 through April 14) and the active agricultural season during which pesticides are commonly sprayed (the “spray season”, defined as April 15 through June 30). Spray season designation was based on local farming practices in 2019.
Samples were transported to our laboratory on ice, analyzed for specific gravity, and stored at −80 °C until overnight shipment on dry ice to an external laboratory. There, they were analyzed for eleven metabolites of common insecticides and herbicides [
37]. These included the metabolites of four organophosphate insecticides, five pyrethroid insecticides, and two herbicides. Included with the sample shipment were six duplicate samples, which were each analyzed for all eleven biomarkers; laboratory analyses were blinded to the identity of these quality assurance (QA) samples.
2.6. Advocate Interviews
We interviewed five farmworker advocates, identified as professionals who work in support of farmworkers and their families in southwestern Idaho. During these interviews, we asked about organizational mission, job description, and perceptions of the rewards and challenges among Latina farmworkers. The results of these interviews provided additional perspectives and context on the experiences of Latina farmworkers.
2.7. Data Analysis
Survey data were entered into Microsoft Excel and duplicate data checks were made by two researchers to ensure the accuracy of the data entry. Survey data were analyzed using STATA 13 (StataCorp; College Station, TX, USA). Frequencies of missingness, counts of categorical responses, and summary statistics were generated for all items.
Focus group and interview data were analyzed using NVivo (QSR International, V12), a computer-assisted qualitative data analysis program. The qualitative data were coded using both inductive and deductive coding techniques. Multiple members of the research team worked together to develop a coding scheme. They then conducted line-by-line coding of the transcripts to assign codes to the text. They did so independently, and results were compared to ensure inter-coder reliability.
For biomonitoring results, we first calculated the relative percent difference (RPD) between duplicate pairs of QA samples and assessed the frequency of detection for all biomarkers. We decided a priori to focus our analysis on biomarkers with an RPD <40% and a detection frequency >50% based on the treatment of censored data recommended by Antweiler and Taylor [
38]. For biomarkers detected in at least 50% of samples, concentrations below the detection limit were replaced with a value equal to the limit of detection divided by the square root of 2 [
39]. Prior to all analyses, we employed specific gravity measurements to adjust for urinary dilution according to the method described by Chiu et al. [
40]. We calculated summary statistics for all biomarkers and compared measured concentrations across seasons and across job tasks using the nonparametric Mann–Whitney U test for non-normally distributed data. We also identified outliers in the urinary biomonitoring dataset, defined as concentrations greater than or equal to 1.5 times the interquartile range (IQR) above the third quartile or 1.5 times the IQR below the first quartile. Extreme outliers were defined as three times the IQR above and below the third and first quartiles, respectively.
4. Discussion
This study aimed to describe challenges to well-being among a growing, yet understudied, population of Latina farmworkers. While significant literature exists describing the health-related risks faced by men who labor in agriculture [
2,
5,
6,
7], far less attention has been paid to the experiences of women in this same workforce. Building from our conceptual framework, our findings suggest that gender may play a critical role in the unique challenges facing farmworkers, as several of the themes that emerged in our work had very specific gendered aspects.
For instance, Latina farmworkers in this study frequently described the challenges of balancing the demands of agricultural work with their household responsibilities. In particular, childcare responsibilities—both the act of caring for children and the work involved in acquiring other forms of childcare during their work shifts—played a central role in the data. Women in this study also consistently expressed concerns around food security and food provisioning, for which they felt primary responsibility. These challenges and concerns may stem from gendered responsibilities which are an outcome of social hierarchies.
Consistent with these findings, previous research has found Latina farmworkers to experience greater stress and anxiety than Latinas who were employed in other occupations [
30]. Further, employed Latinas (both farmworkers and non-farmworkers) were found to experience greater stress and anxiety than unemployed Latinas [
30]. The authors of this work suggest several potential factors as possible causes of this increased stress and anxiety, including work–family balance and the “significant domestic responsibilities” these women shoulder.
Even concerns around pesticide exposure may have a gendered component, as women may be less likely to receive pesticide safety training than men and may be less likely to be issued personal protective equipment [
21]. Female farmworkers may also be more susceptible to adverse health effects associated with pesticide exposures than their male counterparts [
42], particularly during hormonal-based processes such as pregnancy, lactation and menopause [
43]. In previous studies, pregnant farmworkers have expressed concern that pesticide exposure could be hazardous to pregnancy health [
44]. While the sample size for the biological monitoring component of this study was relatively small, we did measure the highest exposures in samples collected during the agricultural spray season from women who reported mixing and applying pesticides. These outlying measurements were more than an order of magnitude higher than the 95th percentile of exposures measured in women and Mexican-Americans in the National Health and Nutrition Examination Survey [
45,
46,
47], and notably higher for 3-PBA and
trans-DCCA compared to the highest levels measured among migrant male farmworkers in Sonora, Mexico [
48] (see
Table 3). Further, the single highest exposure was measured in one such woman who reported that she never received any pesticide safety training.
Strengths of this study include a mixed-methods approach that allowed us to investigate various aspects of well-being using multiple measures, including surveys, interviews, focus groups, and pesticide biomonitoring. We also benefited from strong relationships with community-based organizations, which allowed us to recruit a reasonably sized cohort from a difficult to reach population. However, we acknowledge that this study was limited by the fact that we recruited a convenience sample, which is not necessarily representative of the Latina farmworker population in the US or even in the region of southwestern Idaho from which we recruited our participants. Furthermore, although we never intended for all study participants to take part in all of the study components, we recognize that each of the additional components was only completed by a subset of the overall cohort.