Using Concept Mapping within a Community–Academic Partnership to Examine Obesity among Mexican Immigrants

Abstract

Background

Weight gain is common following migration to a new country. Mexican immigrants have a disparate prevalence of overweight/obesity and food insecurity. Social stressors, such as unemployment, discrimination, and the threat of deportation, may fuel both food insecurity and weight gain in this population.

Objectives

We sought to (1) examine community-defined causes and correlates of obesity among Mexican-Americans; (2) determine how current social stressors, policies, and programs impact food insecurity and obesity; and (3) identify community-defined priorities for preventive interventions.

Methods

Group concept mapping (GCM) was used in a community–academic partnership (CAP) to describe the factors contributing to weight gain and obesity among Mexican immigrant families. Activities included community brainstorming, sorting and rating, multivariate statistical analysis, and community interpretation of results.

Results

Eighty statements were generated in the brainstorming sessions. These statements were sorted into nine clusters, which were organized into three regions: (1) intrapersonal factors; (2) community-level factors; and (3) social policy–related barriers. Statements reflecting the impact of immigration-related stressors were found in all three regions, addressing participants’ fears of deportation, and the prioritization of resources away from healthy eating, resulting in food insecurity. Community members identified five priority areas for intervention planning: (1) lack of exercise; (2) lack of knowledge of a healthy diet; (3) expense of healthy foods; (4) “junk” food; and (5) stress management.

Conclusions

Results suggest high levels of social stress are contributing to food insecurity and obesity among Mexican immigrant families. Areas identified for intervention planning reflect the need for a multifaceted approach toward obesity prevention.

Obesity is a major global public health problem, given that one-third of the world’s population is classified as overweight or obese and two-thirds of these persons live in developing countries.¹ Weight gain is common following immigration,^2–4 and Mexican-Americans (the largest subgroup of Hispanic immigrants) have a disparate prevalence of overweight and obesity (78.8%), as compared with non-Hispanic Whites (66.7%).⁵ In addition, the risk of excess weight increases exponentially with time living in the United States.^6,7

Paradoxically, obese Mexican immigrants are also at higher risk for food insecurity, defined as a lack of consistent access to enough food for an active, healthy life.⁸ Food insecurity affects up to 46% of Mexican origin women and 34% of their US-born children.⁹ Although interventions to improve diet and related health outcomes have largely targeted individual knowledge, attitudes, and behaviors,^6–9 there are multiple social and community-level factors that impact and limit the choices low-income individuals make.¹⁰

Food insecurity has been linked to obesity and high levels of social stress among low-income populations.^11–14 Social stressors, such as unemployment and discrimination, may fuel food insecurity and drive the consumption of cheap, readily available, high-sugar foods, resulting in the metabolic syndrome.^15,16 For Mexican immigrant families, the threat of deportation is an omnipresent social stressor.¹⁷ In families with mixed-immigration status, the parent who remains behind in the United States frequently struggles to provide food for the children.^18–20 Deportation raids in a community cause increased fear and mistrust among immigrants, who may shun the very social programs that safeguard against food insecurity.^21–23 In response to these threats, immigrants may sequester themselves at home, increasing their social isolation and emotional distress, all of which have been linked with food insecurity.²³

CAPs show great promise in helping to identify the population-specific risk factors which contribute to health disparities such as obesity and in developing effective community-level prevention and intervention approaches to ameliorate these disparities. More specifically, they may provide insight into how to keep a community healthy in the midst of stress.

With funding from an R13 grant from National Institutes of Health /National Institute of Child Health Development, the purpose of this community-based participatory research (CBPR) project was to develop a comprehensive CAP between New Brunswick Tomorrow (NBT) and Rutgers, The State University of New Jersey, designed to address health issues among Mexican immigrant families. The focus on obesity is reflected in the CAP’s first joint initiative, Project PESO (People Engaged in Stopping Obesity). This work was informed by a syndemic perspective to the management of obesity, acknowledging the interaction of various biopsychosocial and social factors to impact community health concerns.^24,25 The results will be used to design a syndemic community-level intervention to address obesity among Mexican immigrant families.

GCM was used to identify the community-perceived factors felt to contribute to weight gain and obesity among Mexican immigrants. GCM is a mixed-methods approach that lends itself well to the conduct of CBPR.^26–30 The goal of GCM is to capture not only participants’ shared views, but also the unique views of individual participants.³⁰

The objectives of Project PESO were:

1. To examine community-defined causes and correlates of obesity and barriers to accessing preventive/health protective resources;

2. To determine how current social stressors, policies, and programs impact food insecurity and obesity; and

3. To identify community-defined priorities for preventive interventions that can address obesogenic factors unique to the target population and incorporate “community wisdom” into intervention designs.

METHODS

History of the CAP

The city of New Brunswick, New Jersey, is home to the flagship campus of Rutgers, The State University of New Jersey (RU), and since the early 2000s, a sizeable Mexican immigrant population. Roughly 56% of the city’s 56,427 residents are Hispanic, a significant percentage of which are Mexican immigrant families.³¹ The prevalence of childhood obesity among Hispanics in New Brunswick is 31%, a figure higher than any other urban area in New Jersey and considerably higher than the national average of 17%.³² Data from the Behavioral Risk Factor Surveillance System indicate a similar percentage (31.9%) of Hispanic adults in New Jersey are obese.³³ As many as 52% of New Brunswick residents report being food insecure.³⁴

Established in 1978, NBT is a non-profit organization that works with more than 30 community-based organizations (CBOs) to ensure that health, human service, and social issues are addressed that complement the physical and cultural revival of the city (www.nbtomorrow.org/about). NBT has spearheaded community programs that engage the Latino population, such as Esperanza Neighborhood, Ciclovia, and Live Well-Vivir Bien New Brunswick. Since 2000, RU researchers have partnered with the Health Task Force of NBT to develop community health initiatives. One such program is a partnership between RU School of Nursing and Lazos America Unida (LAU), a Mexican-American organization headquartered in the city, to train community health workers/promotoras de salud. Using the definition of CAPs developed by Drahota et al,³⁵ the academic members and their community partners at LAU proposed the development of a formal CAP to enhance the capacity of researchers and CBOs to address health disparities in New Brunswick using CBPR methods. With funding from the Robert Wood Johnson Foundation, RU investigators formed the Greater New Brunswick Community Health Collaborative, a multidisciplinary group of scholars interested in community-partnered research collaborations and approached NBT to form a partnership. An organizational chart of the resulting partnership is presented in Figure 1.

Figure 1.

Organizational chart for the community–academic partnership (CAP) between New Brunswick Tomorrow and Rutgers University.

One of the first activities of the CAP was to establish a community advisory board (CAB) to (1) provide support for the ongoing CAP; (2) provide a mechanism for building capacity in the community and RU; and (3) guide collaborative research efforts. The CAB is a mix of service providers, consumers, academicians, community leaders, and representatives from CBOs. With the help of the CAB, we developed an agreement of participation among the community and academic partners. The agreement outlined the purpose of the partnership, guidance principles of the partnership, frequency and location of meetings, designation of leadership roles, and governing and voting policies. A set of “operating norms” for the CAP were developed to encourage all partners to become involved in the governance and day-to-day operations of the partnership.³⁶ Minutes were kept for each meeting and are available for viewing by community members and academic partners. Consistent with principles of CBPR, community members actively participated in every aspect of Project PESO.

GCM

An introductory workshop was held for community members and academicians to learn about GCM and provided hands-on opportunities to practice. The workshop was conducted by a former RU faculty member and her community partner who are experienced with GCM. Community members liked the visual depiction of the results in the concept maps and felt GCM would work well with in both English and Spanish. Based upon their suggestions, some modifications were made to the format of the Project PESO brainstorming sessions. For example, it was decided each group would determine the language used in the session (English, Spanish, or both) and participants would write their responses on slips of paper rather than enter them online. As a GCM activity, Project PESO was then divided into five steps: (a) community preparation; (b) community brainstorming; (c) community sorting and rating; (d) multivariate statistical analysis, and (e) community interpretation and utilization of results.³⁷

Promotora Training

In the weeks before the study began, the principal investigator (KD) and the Promotora Coordinator (MV) conducted training sessions with the promotoras at the offices of LAU. These sessioheritagens included general topics such as ethics training/CITI certification, recruitment strategies, and public speaking skills (including peer-reviewed mock presentations). Additional skill building sessions were held before the sorting and rating events and the community interpretive session. Other CBOs affiliated with NBT volunteered time and space for training of the promotoras in study-related activities.

IRB approval

Approval for Project PESO was granted by the Electronic Institutional Review Board at RU. Written consent was not required for the adult participants, since the data did not contain identifiers and was grouped together for analysis. Consent forms (in English and Spanish) were provided to interested adolescents, who returned the signed forms and provided assent on the day of the session.

Community Preparation

Three conference-style meetings were held to introduce the research topic to the community and academic partners. The first meeting targeted RU faculty and students, while the second meeting focused on CBOs and individual members of the Mexican immigrant community. The third meeting was designed to bring together community members and academicians through (1) a discussion of the CBPR process; (2) elicitation of community members’ concerns about obesity and how these concerns fit with their priorities; and (3) ascertainment if concerns of academicians, CBOs, and community members were in alignment. After the third meeting, there was strong agreement among the three groups that obesity and the escalating risk of obesogenic diseases were the biggest public health threats to Mexican immigrant families. The focus prompt, “What do you think are the most important factors that contribute to weight gain among Mexican immigrants?” was developed and piloted with members of CBOs to affirm that the participants would understand the intent and generate ideas about the issue.

Brainstorming and Generation of Items

Two purposive sampling methods (homogenous sampling followed by snowball sampling) were used to recruit participants in the brainstorming sessions. Homogeneous sampling is a purposive sampling technique that aims to recruit participants who are similar in characteristics such as age, gender and ethnicity.³⁸ In Project PESO, bilingual promotoras invited immigrant Mexican adults who were members of LAU or one of several other CBOs affiliated with NBT to attend a community brainstorming session, scheduled at various locations and times throughout the city. Adults were initially recruited at face to face meetings of the CBOs and through the CBOs’ Facebook websites. Teenagers were recruited through flyers placed in their high schools and by face to face contacts by younger promotoras who had graduated from the same schools. Adults and teenagers were then encouraged to recruit other potential participants from among their acquaintances (snowball sampling). This type of purposive sampling has been shown to be particularly effective in recruiting populations that are difficult to reach because it takes advantage of established social networks of persons with characteristics of interest.³⁹ This combination of purposive approaches offers a wider range of sampling opportunities from which a researcher can draw.⁴⁰

Each structured brainstorming session was composed of five to fifteen residents. Four of the sessions targeted immigrant adults; three of these adult brainstorming sessions took place in area Catholic churches, while the fourth group was held at a participant’s home. The fifth brainstorming event targeted F and took place in the RU School of Nursing. Refreshments were served, and childcare was offered at all events. At each location, a pair of promotoras introduced the Project PESO brainstorming process and facilitated an open discussion around the focus prompt: “What do you think are the most important factors that contribute to weight gain among Mexican immigrants?” At the request of the participants, the adult sessions were conducted in Spanish and the teen session was conducted in English. Verbal statements were recorded on a large poster board for all participants to view. To promote participation by all, participants were encouraged to write their ideas on slips of paper, which were then added to the list of generated items. Once all five brainstorming sessions were finished, the full set of statements was examined by four members of the community who had been trained as coders. The coders reviewed the statements for common content, redundancies, spelling, and grammatical errors. This process, referred to as idea synthesis, ensures that items are clear, relevant, and not redundant.⁴¹ The coders then collectively agreed to condense the initial statements into a final set of 80 brainstorming statements.

Sorting and Rating of Items

During this step, the list of 80 brainstorming statements was individually sorted and rated by participants, who were either members of the community or RU faculty members/students. The decision to include representatives from both the community and RU was based on comments from CAB members, who felt this approach (1) was consistent with the philosophy of the CAP and (2) would provide data on similari ties and differences between the perceptions of community members and academics. The sorters and raters were recruited by the principal investigator and the promotoras using purposive sampling and none had participated in the brainstorming activities. Each of the final brainstorming statements was printed on a separate card, and each of the sorters/raters received a complete set of cards. The sorters/raters were given instructions to sort the statements into piles in a manner that made sense to them and to provide a label describing the contents of each pile.⁴² The sorters/raters then rated the relative importance of each statement in contributing to weight gain among Mexican immigrants using a four-point Likert scale, where 1 = relatively unimportant; 2 = slightly important; 3 = moderately important, and 4 = very/extremely important.

Multivariate Statistical Analysis

In this step, the data from Project PESO was translated into English and entered into the Concept Systems software program (Concept Systems, Inc, Ithaca, NY).³⁷ Data entry was conducted by a PhD student (LG). In the GCM process, the data from the sorting activity is first aggregated into a total similarity matrix. Subsequently, multidimensional scaling is applied to the total similarity matrix to create a point map, which represents the spatial arrangement of the numbered brainstorming statements. The distance between statement points indicates the perceived relationship between items. Statements that are located near each other on the point map are more likely to be viewed as conceptually related. In concept mapping, a stress value between 0–1 is calculated to assess the degree to which the point map represents the distribution of relationships found in the total similarity matrix. High stress values suggest considerable variability in the way participants group the statements. In general, stress values will be lower (the map is a better fit) when there are more statements and more sorters.⁴² In a pooled study analysis of 69 studies, Rosas and Kane⁴³ reported the average stress value across the studies was 0.285, with a range from 0.17 to 0.34. Hierarchical cluster analysis was then applied, using the location of the points on the map as input so that clusters of points presumably represent broader concepts based on similarly located points. This partitioning of the points on the map produces a cluster map.

Community Interpretation Session

In the final GCM interpretive session for Project PESO, eight community members were recruited by the promotoras to validate the clusters and to assist with interpretation of the final concept map. Large copies of the maps were projected on the wall and paper copies of the final statement list, statements by cluster, cluster map, and cluster map with regions were distributed to community members.⁴⁴ Following a facilitator-guided review of the maps in Spanish, participants reviewed each of the cluster labels as a group to establish consensus that the labels captured the content of each cluster. Participants were then asked to select two to three items within each cluster that would be important to consider in the design of an intervention.

RESULTS

Brainstorming Data

A total of 70 community residents of Mexican descent participated in one of five brainstorming sessions. Sixty-two of the participants were females (88%) and eight (12%) were males. Ages ranged from 16 to 50 years. In response to the focus prompt, a total of 120 initial statements were produced in the five brainstorming sessions. The four coders collectively agreed to condense the 120 statements into 80 final brainstorming statements. Formal brainstorming sessions lasted 60 to 90 minutes each, but informal discussions frequently continued long after the session ended. Many of the comments specifically pertaining to immigration-related stressors were made in these informal discussions. A list of the final brainstorming statements organized by cluster appears in Table 1.

Table 1.

Brainstorming Statements and Average Ratings Organized by Cluster

Cluster	Statements	Average Rating
Barriers to physical activity (M = 3.24, SD = 0.19)
	1 Lack of exercise	3.56
	52 The lack of exercise due to being tired from work	3.28
	62 All day sitting and standing at work, there is a need to be more active	3.19
	75 Unsafe neighborhoods	3.16
	64 People are intimidated to go out to the parks alone	3.00
Financial restraints (M= 2.99, SD = 0.60)
	11 The financial difficulty to get help	3.44
	37 Resources and money	3.38
	41 Healthy food such as fresh vegetables are too expensive	3.38
	67 Fast food is cheaper	3.38
	58 Can’t afford healthy food because they are expensive	3.28
	54 The nearest gym downtown is too expensive	2.97
	7 Resources are not free	2.91
	34 No set schedule to eat during work	2.78
	79 No roller-skating rink	1.44
Time/scheduling (M= 3.04, SD = 0.19)
	61 There is lack of information of health awareness in minorities’ community	3.34
	65 People have misinformation on how to have a personalized diet	3.28
	73 Have to work after school	3.16
	60 We eat what is available to us when we need to eat something whether it is fast food or not	3.13
	44 People have less time to prepare food on the fly	3.00
	36 Lack of being well-rested	2.97
	32 Working too much to prepare food	2.94
	49 We don’t have a consistent schedule as to what time to eat	2.78
	74 Take care of younger siblings after school	2.78
Intrinsic motivational (M = 2.91, SD = 0.22)
	27 Lack of awareness on the importance of having a healthy weight	3.22
	10 The need to take the initiative and to go out and exercise in the local parks	3.03
	19 As women we lack personal time	3.00
	16 Difficulties faced in achieving and keeping high personal desire to do things	2.94
	15 People always look for excuses not to exercise	2.75
	17 Sometimes we just go to the doctor and think it is over	2.53
Mental health and eating (M= 3.05, SD = 0.23)
	28 Stress	3.47
	59 Stress and insecurity	3.28
	26 Suffering more from anxiety and eating more	3.25
	8 Lack of personal desire and self-discipline	3.22
	14 People lack the motivation	3.19
	21 No self-discipline	3.13
	13 Anxiety impedes desire and self-discipline	2.97
	63 Many people feel shame about weight	2.94
	56 Drugs and alcohol	2.94
	45 Addictions	2.91
	20 Women don’t care for themselves	2.72
	46 Comparison problem leading people to frustration and depression	2.69
Culture and gender- based norms (M = 2.54, SD= 0.50)
	5 There is a need for parents to get involved	3.34
	55 Lack of support from parents	3.28
	2 We eat what we want	3.03
	4 There is help for children, but adults don’t use it	2.69
	24 Husband does not like to eat healthy	2.50
	30 Women don’t challenge the excuses that they and their children make up	2.45
	6 Adults do not integrate	2.38
	78 Eating to socialize	2.34
	12 It’s a child’s lack of interest	2.32
	29 Children use excuses like “Es que el papa es gordito”	2.03
	25 Latinos are hard-headed	1.59
Eating habits (M = 2.81, SD = 0.39)
	35 Children eating junk food	3.47
	51 Lack of vegetables in their diets	3.38
	23 Not enough greens	3.25
	22 Large food portions	2.91
	42 Children are obligated to eat what their parents say	2.91
	39 Inconsistency from too much or too little food	2.88
	66 Parents make the food	2.81
	43 Children are no longer eating healthy, they refuse to eat vegetables	2.75
	57 Excess of food	2.72
	38 Extremes in diets	2.69
	33 Malnutrition due to Mexican diet	2.56
	18 Some children do not want to eat home cooked meals	2.31
	40 Eating too fast	1.97
Fast food (M = 2.89, SD = 0.31)
	3 Diet	3.41
	53 Fast food and canned food	3.06
	48 Children are eating junk food after school so they do not eat the food at home	2.94
	76 Convenience stores tempting with chips and soda	2.88
	31 Food staples that are processed in US	2.53
	80 Genetics	2.50
School (M = 2.75, SD = 0.27)
	9 Kids do not have a lot of physical activity in school although projects and resources exist	3.19
		2.94
	72 Students only have gym class two times a week	2.91
	69 Some school lunches don’t offer many vegetables	2.91
	70 Schools do offer fruits, but students don’t choose healthy	2.78
	50 Most meals are at school where they have less control	2.72
	68 High school lunches include mostly pizza	2.59
	71 There is healthy food at home, so children pick less healthy foods at school	2.50
	77 School breakfast has cinnamon rolls	2.22

Generated by Mexican-Americans (N = 70) as part of a group concept mapping project in New Brunswick, New Jersey.

Sorting and Rating Data

An additional 32 persons participated in the sorting and ranking activities; 17 were community members and 15 were academics (a mix of faculty members and students). The ages of sorters and raters ranged from 16 to 63 years; 19 were females (59%) and 13 were males (41%). During the sorting activity, community members and academics sorted the statements in a way that reflected an understanding of the factors related to weight gain among Mexican immigrants. These clusters and their content included (1) barriers to physical activity—environmental and work-related factors that impede exercise; (2) financial restraints—meager household income coupled with the expense of eating healthy; (3) time and scheduling—work-related barriers to preparation of food and regular mealtimes; (4) intrinsic motivational factors—lack of personal desire and discipline for healthy behaviors; (5) mental health and eating—influence of stress, anxiety, and depression on eating habits; (6) culture- and gender-based norms—how traditional Mexican beliefs about women and family can interfere with healthy eating; (7) access to healthy foods—unhealthy eating habits that promote weight gain; (8) fast foods—preference for and easy availability of fast foods; and (9) school-related barriers—difficulty in maintaining a healthy diet during the school day. The nine clusters were then organized into three regions, or larger groupings of constructs: (1) intrapersonal factors; (2) community-level factors; and (3) social policy-related factors. Figure 2 shows the final cluster map with regions. The numbered points on each cluster represent the numbers of the brainstorming statements identified in Table 1. The multidimensional scaling analysis of the total similarity matrix converged after 22 iterations, yielding a stress value of 0.31, indicating a reasonable fit between the original sort data and the point map.⁴³ On the map, smaller clusters, such as barriers to physical activity, represent a collection of more conceptually coherent items. The cluster closest to the middle of the map is analogous to a core cluster or anchor, representing a key theme that is connected to other statements. In this map, time management was perceived to be a key theme and a major barrier to healthier lifestyles.

Figure 2.

Community–academic cluster map with regions generated by Mexican-Americans (n = 17) and Rutgers faculty and students (n = 15) as part of a group concept mapping project in New Brunswick, New Jersey.

Although both community members and academics agreed on the cluster labels, the two groups differed somewhat on the placement of statements in each cluster and region and the importance rating of the cluster. The ratings from participants were first averaged for each item, then for each cluster based on the specific items belonging to that cluster. These results are shown in Table 1 by average cluster order, with the highest to lowest item average arranged within. To assess for differences in the importance of the clusters between community members and academics, an absolute pattern match was constructed using the Concept Systems Global Max software. The results are shown in Figure 3. T-tests revealed significant differences between the two groups for only two clusters: (1) school: community members (M = 2.94, SD = 0.29), and academics (M = 2.53, SD = 0.31), t(16) = −2.88, p < .02, and (2) eating habits: community members (M = 2.99, SD = .031) and academics (M = 2.61, SD = 0.53), t(24) = −2.27, p < .05. Although there seemed to be a difference between the two groups on Financial restraints: community members (M = 2.83, SD = 0.53) and academics (M = 3.18, SD = 0.73), the difference was not significant, t(15) = 1.16, p = .13.

Figure 3.

Relative pattern matching between Mexican-Americans (n = 17) and Rutgers faculty and students (n = 15) as part of a group concept mapping project in New Brunswick, New Jersey.

Community Interpretation Data

Five of the eight invited community members participated in the map interpretation sessions; three had been participants in the brainstorming sessions, while the other two had not. Three of the five were females (60%) and two were males (40%). Ages ranged from 18 to 56 years. The final nine cluster map with three regions was presented to this group for review and validation. Although there was 100% agreement with cluster names, participants illuminated some aspects of the clusters that were not originally observed in the brainstorming content. For example, in the barriers to physical activity cluster, one participant noted that few children in this community join sports teams, especially when significant parental involvement is required. In the financial barriers cluster, two women discussed how they budget their resources, including looking for sales, buying food items in bulk, and freezing food for future use. In the time management cluster, participants talked about managing their energy to make better use of their time, especially when working three or four jobs. Participants saw this pattern as necesario para sobrevivir—necessary for survival. There were many new comments supporting the mental health and eating cluster. Participants identified loneliness, isolation, and subsequent overeating as common responses to anti-immigrant sentiment. Several community members noted Mexican-born parents often feel guilty their US-born children are caught up in deportation concerns. As a result, they cope by indulging their children, permitting them to eat whatever they want. Other parents are so exhausted when they return from work, they have little energy to supervise their children’s eating habits or to help them learn to cook. One participant added many women in the community have a history of sexual abuse, resulting in shame and insecurity and that they cope by overeating. In the fast foods cluster, one member noted it is difficult to find healthy food choices in the small bodegas near her home. One needs to travel, usually by car, to a larger grocery store and this is time consuming. Other participants noted that when several families share an apartment, they may not have immediate access to cooking facilities, and fast food is the easiest option. Last, there were a few additional comments about the school cluster. One younger participant recalled high school students often buy lunch from vending machines or order from a nearby fast food restaurant, both of which are faster than cafeteria service. The young man recalled that among his peers, eating fast food created an image that soy americano—I am an American. These comments were valuable, as they put a finer point on the content in each of the clusters.

In the final discussion, participants revisited the original brainstorming statements to identify what they felt to be the most important elements in the design of an intervention to lower obesity rates among Mexican immigrants. The statements they chose were (1) lack of exercise; (2) lack of knowledge about a healthy diet; (3) healthy foods are expensive; (4) junk food; and (5) stress management.

DISCUSSION

We found GCM to be a useful, pragmatic device to facilitate the CBPR process in Project PESO and foster the growth of a comprehensive CAP that engaged both community members and academicians. GCM helped us to clarify the stakeholder-authored factors perceived to contribute to obesity in this community, assess community and academic priorities, and frame action and assessment. In Project PESO, the promotoras played a key role in facilitating various components of GCM, including community preparation, brainstorming, sorting and rating, and community interpretation activities. Participants reported a variety of factors they felt contribute to the rising prevalence of overweight and obesity in their community. These issues went well beyond the standard suggestions to “eat less and exercise more” and included biological, behavioral, and social determinants of health.

Results suggest that high levels of social stress are contributing to food insecurity and obesity among Mexican immigrant families. The statements in region 3 (social policy-related factors) specifically addressed participants’ fears of deportation and the subsequent prioritization of scarce temporal, financial, emotional, and social resources away from healthy eating, resulting in food insecurity. These findings support results suggesting that anti-immigrant sentiment and local immigration enforcement policies are associated with an increase in the food insecurity risk of Mexican noncitizen households.^45–47 Indeed, it has been suggested that in immigrant communities, local and federal immigration enforcement activity is analogous to a negative social determinant to health.⁴⁷ While much attention has been drawn to the public health threats posed by new immigrants,^48–50 we are only beginning to understand the dangers of anti-immigrant sentiment and policies on the physical and mental health of immigrants and their families.^51–53

As the core cluster in the map, time management emerged as a challenge to healthy eating. Because of the need to work multiple jobs, many immigrants reported working all day without eating, then making up for it by eating cheap, readily available fast food later at night. Surprisingly, the Fair Labor Standards Act does not require employers to provide meal breaks for their workers,⁵⁴ so an undocumented worker is not likely to request time to eat, for fear of losing his or her job. The irony of this situation is that a significant percentage of the estimated 11.3 million undocumented workers in the United States are employed in food related industries, for example, agriculture and food service.⁵⁵ Results suggest time management appears to impact teenagers as well as adults., current events have intensified this shift in the parent–child relationship.^56,57 Fears of detainment/deportation lead parents to send children to do grocery shopping. Teenagers are often expected to prepare meals for themselves and younger siblings and parents admitted they are too busy working multiple jobs to teach their children to cook. It becomes obvious that the federal government’s punitive immigration policies have impacted undocumented adults as well as their US-born children.

Interestingly, we observed differences in the sorting and rating patterns of community and academic participants. Community members were more likely to attribute the difficulties encountered in maintaining a healthy diet to intrapersonal failures, while academics ascribed many of the same difficulties to community restraints and pejorative social policies. The tendency for Hispanic women to blame themselves for the twin crises of food insecurity and obesity is likely deeply rooted in the concept of marianismo^58,59 and the sense that “foodwork”⁶⁰ is women’s work. Van Esterik⁶¹ has noted that for many women, the inability to provide food for their family is analogous to a form of torture. Conversely, academics who view food insecurity as solely a form of structural violence⁶² may miss key learning needs identified by community members, such as how to buy healthy foods on a limited budget or how to avoid stress related eating. The importance of community input was also evident in the rating differences between community members and academics. Community members identified the nutritional environment in schools and stress-related eating habits as priority areas, underscoring the value of community wisdom in health planning. The five areas earmarked for intervention planning coincided with the results of pattern matching in Figure 2 and were chosen from all three regions of the map, reflecting a syndemic approach⁶³ toward obesity prevention. Given the unique group perspectives, it may be worthwhile to systematically examine the similarities and differences in the way community and academic participants structured the content.⁶⁴ Evaluation of these differential patterns could yield greater insight regarding intervention design and communication.

There are a few limitations to the study. The study was conducted at a time when the federal government expanded “enforcement priorities” so that all undocumented immigrants have become targets for arrest and deportation. This includes individuals who have lived in the United States for many years, have US-born children, and those with no history of prior arrest. Because of these policies, the adult members of the Mexican immigrant community preferred that the brainstorming sessions were open only to community members, the promotoras and the principal investigator. For the same reasons, most of the brainstorming took place in area churches located within immigrant neighborhoods. In addition to purposive sampling, the majority of brainstorming participants were women (88%) which may limit the generalizability of the findings.

This GCM study was an initial step toward building a CAP to explore the relationships among social stressors, food insecurity, and obesity among Mexican immigrants. GCM proved to be a useful tool with which to build this partnership, while focusing on the larger goal of promoting community health in the midst of stress.