An Innovative Approach to Promote Weight Loss Among Mexican Immigrants: A Pilot Study

Abstract

Background:

Mexican Americans are among the highest risk groups for obesity and its associated health consequences, including diabetes, heart disease, and cancer.

Methods:

154 overweight/obese Mexican Americans recruited from the Mexican Consulate in New York City were enrolled in COMIDA (Consumo de Opciones Más Ideales De Alimentos) (Eating More Ideal Food Options), a 12-week Spanish-language lifestyle intervention that included a dietary counseling session, weight-loss resources, and thrice-weekly text messages. Participants’ weight (primary outcome); dietary intake, physical activity, and nutrition knowledge (secondary outcomes) were assessed pre- and post-intervention.

Results:

Of the 109 who completed follow-up, 28% lost ≥5% of their baseline body weight. Post-intervention, participants consumed more fruit and less soda, sweet pastries, fried foods and red meat; increased physical activity; and evidenced greater nutrition knowledge.

Discussion:

A community-based lifestyle intervention with automated components such as text messaging may be a scalable, cost-effective approach to address overweight/obesity among underserved populations.

An Innovative Approach to Promote Weight Loss Among Mexican Immigrants: A Pilot Study

Obesity and its associated health consequences disproportionately affect underrepresented populations. Mexican Americans (MAs) are among the highest risk groups. Over 1 in 2 MA men and women are obese and trends in obesity prevalence from 1999–2000 to 2017–2018 show more rapid increases in obesity and severe obesity in MAs than in non-Hispanic Whites (NHWs) [1]. The high incidence of overweight/obesity and of abdominal obesity confer greater risk of cardiovascular disease (CVD) and metabolic syndrome [2]. In a study of U.S. Latinos from diverse backgrounds, DM prevalence ranged from 10.2% among South Americans and was highest, at 18.3%, among MAs [3]. Obesity is also a risk factor for some cancer types, including breast (postmenopausal), endometrial, prostate, and colon [2].

Social, systemic, and lifestyle factors, such as energy-dense, nutrient-poor diets [4], physical inactivity [5], lack of health education [6], and barriers to healthcare access [6] increase obesity risk among MAs. Acculturation to the U.S. has had deleterious effects on the MA diet and studies have suggested that acculturation and chronic stress can contribute to visceral adiposity accumulation and insulin resistance [7].

Five percent weight loss has been associated with clinically significant health benefits, including improvements in blood pressure, blood cholesterol, and blood sugar [8]. Numerous clinical trials have demonstrated that lifestyle interventions such as the Diabetes Prevention Program (DPP) [9] and Look AHEAD [10] can lead to meaningful weight loss and a reduction in CVD, DM, and cancer risk [8]. However, these studies were developed for the general population and were not designed to be translatable to community settings and thus have not been widely disseminated due to their high cost and participant burden [11], especially for groups with long work hours. In the few studies that examine the effectiveness of weight loss interventions among U.S. Latinos, Latinos lose less weight compared to NHWs and are more likely to regain weight once the intervention is over [12]. This may be due to the underutilization of preventive services by minority populations [13]. In the DPP study, Latinos were half as likely to attend DPP sessions as NHWs but there were no weight loss differences by ethnicity after controlling for attendance [14], suggesting that strategies to engage and retain Latinos may lead to more successful outcomes for this group.

Most research on obesity treatment has been conducted in NHW populations, and interventions that have been developed for Latinos have generally aggregated Latinos of various ethnicities into a single category [12]. In a systematic review of evidence-based obesity treatment interventions for Latino adults in the U.S., only 8 interventions were found that were intended for MAs [15]. Yet recommended lifestyle modifications (e.g. suggested recipes, physical activity recommendations) are more likely to be successful when tailored to participants’ culturally and regionally-based food and activity preferences [16].

Lifestyle interventions that use text messaging (i.e., short message service [SMS]) show promise for weight loss [17, 18], but very few of these have been developed or culturally adapted for Latinos. In a randomized controlled trial comparing SMS alone or in combination with brief, monthly health-coaching calls to reduce weight among overweight and obese English- and Spanish-speaking adults, a combination of SMS and health-coaching calls was most effective, suggesting that SMS content needs to be supported with other methods to affect weight-loss related behavior change [19].

There is a need for more data on innovative lifestyle interventions in community settings, adapted for high-risk underrepresented populations, that leverage technology such as text messaging to reduce participant burden and costs but still lead to dietary change and weight loss [20, 21]. This paper presents results from COMIDA (Consumo de Opciones Más Ideales De Alimentos) (Eating More Ideal Food Options), a 12-week, community-based Spanish-language pilot dietary counseling and education program that incorporates evidence- and theory-based cognitive-behavioral change strategies (including features from the DPP and Look AHEAD), includes text messages to encourage behavior change, and addresses specific cultural characteristics and needs of MAs [22]. It details the impact of the pilot intervention on participant weight (primary outcome), and dietary intake, physical activity, and nutrition knowledge (secondary outcomes).

Methods

Design

This was a nonexperimental single-arm, pre-post dietary counseling and education pilot intervention conducted between January 2016 and August 2018 in a community setting with a duration of 12 weeks. Study approval was obtained from the Memorial Sloan Kettering Cancer Center (MSK) Institutional Review Board.

Participants and Procedures

Recruitment was conducted through the Ventanillas de Salud (VDS; Health Windows) program, a partnership between the Mexican government and over 400 U.S. non-profit and private agencies designed to promote MA health by providing services such as mobile health visits, health fairs, case management, assistance with health insurance enrollment, and referrals to primary care providers. The VDS has 50 locations nationwide and reaches hundreds of thousands of MAs each year, including many who are uninsured and/or undocumented [23]. The New York Metropolitan area (NY) VDS is an academic-community partnership of the MSK Immigrant Health and Cancer Disparities Service and the Mexican Consulate in New York City (NYC). Individuals attending the VDS at the Mexican Consulate in NYC were approached for interest and eligibility on study recruitment days (2 rotating days per week) in Spanish by trained outreach staff.

Study inclusion criteria were: 1) self-identifies as Mexican or Mexican American, 2) seeking services at the NY VDS, 3) Spanish-speaking; 2) overweight (BMI of 25 – 29.9 kg/m²) or obese (BMI ≥ 30 kg/m²); 3) at least 18 years of age. As part of the screening process, trained research study staff measured participants’ height (using a measuring tape affixed to the wall) and weight (using a Taylor 7009 Lithium Electronic Scale), which were used to calculate BMI (kg/m²). Exclusion criteria were: 1) not being in the NYC area for the 3 month study duration; 2) being pregnant and/or breastfeeding; 3) the presence of a serious chronic disease such as cancer, kidney disease, or liver disease (individuals with diabetes, lactose intolerance, and/or high blood pressure were allowed to participate in the study); 4) having dietary restrictions; 5) not having a phone that accepts text messages; 6) presence of a serious psychiatric or cognitive impairment, and/or; 7) having another family member already enrolled in COMIDA. To address concerns that can deter research participation in this community (e.g., medical mistrust, fears of deportation), participants provided verbal informed consent prior to study participation. Participants received a $30 incentive.

Description of the Intervention

COMIDA is a Spanish-language, culturally and linguistically adapted twelve-week dietary counseling and education community-based pilot program designed to help MAs make healthy eating and physical activity choices and lose weight. After completion of baseline measurements and surveys (including a 24-hour dietary recall), individuals received an in-person, 45–60 minute individual dietary counseling session delivered by trained bilingual outreach staff. The session is based on the USDA MyPlate program (available in Spanish) [24], tailored to the Mexican population’s commonly eaten foods, grounded in Social Cognitive Theory (SCT) (i.e., focuses on self-efficacy and social norms) [25], and incorporates Motivational Interviewing [26] approaches (i.e., assessing motivation and stage of change, engaging in participant-led goal-setting, reflective listening). The counseling session is tailored to individuals’ responses to the 24-hour dietary recall and consists of ten modules: 1) Beverages: Avoid Sodas, Artificial and/or Sugary drinks; 2) Vegetables, Fruit, and Fiber; 3) Grains, Tortillas, and Fiber; 4) Proteins: Diversify your Proteins and Choose Lean Ones; 5) Fried Food; 6) Junk Food and Sweets; 7) Fast Food Restaurants and Eating Out; 8) Be More Active: Exercise; 9) Final Activities, Food Selection (Weekly Ad) and Building Your Meal (‘Build Your Tortilla’) (experiential or ‘modeling’ activities, e.g. using a supermarket ad to choose foods and build a meal using food models); 10) Conclusion and Goal-Setting. Modules, which last 3 to 5 minutes each, include an introduction, the module topics, questions and activities for the participant about diet/physical activity behaviors, and money-saving tips. The presentation was developed to present topics in order of importance, in the event that participants do not have sufficient time to complete the counseling session. A random sample of 20% of recorded sessions were reviewed and rated for treatment integrity and fidelity to SCT and MI techniques utilizing a COMIDA program implementation log.

Participants also received the following weight-loss resources at intake: low literacy Spanish-language written materials on diet and nutrition, written information on neighborhood food resources, measuring cups and spoons, a water bottle, and a MyPlate portion plate.

Over the ensuing twelve-weeks, participants received one-way thrice weekly diet/physical activity Spanish-language text messages (e.g., “Instead of snacking on chicharrones or other fried snacks, choose fruits and veggies with lime and chili”) to encourage behavior change.

Measures

All data were collected at baseline and at 3 months, except for demographic information, which was collected at baseline only. Bilingual research study staff administered all measures to participants in Spanish-language, face-to-face interviews.

Sociodemographic Variables

Demographic information collected included state of origin, years of residency in the U.S., education level, monthly household income, language proficiency, occupation, and history of injuries in the workplace.

Primary Outcome Measure

Weight.

Weight was assessed using a Taylor 7009 Lithium Electronic Scale by trained outreach staff at the VDS.

Secondary Outcome Measures

Dietary Intake.

Dietary intake was assessed using a modified version of the Dietary Screener of the 2011–12 California Health Interview Survey (CHIS). The CHIS is a standardized survey developed to capture a large sample of Hispanics of Mexican origin [27] and has been culturally and linguistically adapted in Spanish [28]. It includes self-reported items on dietary intake per week during the month before the interview: fruit (excluding juices), fried potatoes (including French fries, home fries, and hash browns), other vegetables (excluding fried potatoes), regular soda or pop that contains sugar (excluding diet soft drinks), and fast food. The scale was modified to be more inclusive of sweetened drinks, desserts, fried foods and red meats, by adding 7 items: energy drinks, sweetened fruit drinks, coffee and tea with sugar or honey added, sweet pastries (including cookies, cake, pie, or brownies), frozen desserts (including ice cream), fried foods, and red meats. The information was collected through twelve questions in the following format: “During the past month [“or in the past 7 days” for fast-food consumption], how often did you eat [food item name]?” Responses were standardized to reflect mean per-week intake frequency.

Physical Activity.

Physical activity was assessed using the Spanish version of standardized items from the National Health Interview Survey (NHIS) [29]. Participants reported how often (times per day/week/month/year) and for how long (minutes, hours) they engaged in vigorous, light/moderate, and/or strengthening physical activities. These items provided the average time per week participants engaged in physical activity.

Nutrition Knowledge.

Participants were administered a survey based on a tool used in previous research to assess nutrition knowledge among low-income Hispanics and African Americans [30]. Additional items are related to specific content on nutrition and diet included in the individual counseling and written materials. Ten items on nutrition knowledge comprised a nutrition knowledge score used in the statistical analyses; items that were answered correctly (i.e., accurately identifying foods high in saturated fat and foods that are a good source of calcium, fiber, etc.) were given a score of one and those answered incorrectly were scored as zero. A higher score indicated greater knowledge of nutrition.

Data Analysis

For participant demographics, categorical variables were summarized with frequencies and percentages and compared using Fisher’s exact test; continuous variables were summarized with mean ± standard deviation and compared using t-tests. For the primary outcome of weight, 5% weight loss (based on pre-intervention weight) was calculated post-intervention, dichotomized as yes/no, and summarized with frequencies and percentages. For dietary behavior measured by the modified CHIS, mean servings per week were reported across all CHIS domains. Mean servings per week pre- and post-intervention were compared using paired t-tests. To adjust for multiple comparisons, the Benjamini-Hochberg procedure was used with a false discovery rate of 5%. Cohen’s d (with 95% confidence interval) was calculated for each domain. Average time spent per week for vigorous, moderate and muscle strengthening physical activity, pre- and post-intervention were compared using paired t-tests. Two participants who reported extreme amounts of time spent exercising (more than 70 hours a week) were considered outliers and therefore excluded from the analysis. Because of nonnormal data, the distribution of participants’ nutrition knowledge score, pre- and post-intervention, were compared using Wilcoxon signed ranks test [31]. Missing data were determined to be missing at random and were excluded from the analysis.

Results

Participants

A total of 154 Latino/Hispanic individuals consented to participate in the COMIDA intervention. The average age was 40 years (SD=8.075) and 81% were female. Participants’ average duration of years living in the U.S. was 15 (SD=6.889). The most common occupations were in the restaurant and cleaning services industries. The median monthly household income was $1200 (interquartile range [IQR]=$1000).

Of 154 initial participants, 109 completed the 3-month exit survey and weight measurement (Figure 1). Staff made at least 10 attempts to reach participants for follow up, including during evenings and weekends. There were no significant differences in baseline characteristics between those who completed the exit survey (n=109) and the 45 participants who did not (Table 1).

Figure 1.

Modified CONSORT Flow Diagram for COMIDA Pilot Study

CONSORT flow diagram template, modified from http://www.consort-statement.org/consort-statement/ flow-diagram

Table 1.

Baseline characteristics between participants (N=109) and those lost to follow-up (N=45)

Variable		n (%)	n (%)	Sig.
		N=109	N=45
Demographics
Gender	Male	20 (18)	9 (21)	0.73
	Female	89 (82)	35 (79)
Education	0–6 years	35 (32)	15 (34)	0.90
	7–12 years	60 (56)	23 (52)
	> Some college	13 (12)	6 (14)
English	Very well	2(2)	0 (0)	0.82
proficiency	Well	18 (17)	10 (23)
	Not well	63 (59)	21 (49)
	Not at all	24 (22)	12 (28)
Spanish	Very well	66 (61)	30 (70)	0.81
proficiency	Well	41(38)	12 (28)
	Not well	1 (1)	1 (2)
	Not at all	0 (0)	0 (0)
Employment	Not currently working	41 (38)	13 (29)	0.66
	Restaurant worker	17 (16)	11 (24)
	Cleaning	20 (18)	7 (16)
	Construction/factory worker	8 (7)	2 (4)
	Retail/finance/business	4 (4)	3 (7)
	Nanny/babysitting	4 (4)	3 (7)
	Driver	3 (3)	2 (4)
	Hospitality/leisure	4 (4)	1 (2)
	Home attendant	4 (4)	1 (2)
	Other	4 (4)	2 (4)
Weight		M (SD)	M (SD)
Mean weight (lbs)		173.6 (39.14)	171.4 (34.257)	0.77
Self-reported food servings/wk		M (SD)	M (SD)
Fruit		5.6 (4.92)	5.9 (6.39)	0.66
Fried potatoes		0.9 (1.59)	1.0 (2.03)	0.76
Other vegetables		5.1 (6.48)	4.0 (4.09)	0.30
Soda		2.6 (4.65)	4.1 (5.75)	0.10
Fast food		3.9 (5.71)	3.3 (7.47)	0.60
Energy drinks		2.3 (1.75)	2.4 (2.71)	0.96
Sweetened fruit drinks		0.5 (1.34)	1.2 (3.50)	0.07
Coffee/tea with sugar		1.7 (2.35)	3.6 (8.04)	0.07
Sweet pastries		6.0 (5.67)	7.6 (10.89)	0.23
Frozen desserts		1.4 (4.98)	1.0 (2.31)	0.58
Fried foods		1.6 (2.42)	1.1 (1.66)	0.23
Red meats		2.0 (2.06)	2.7 (3.55)	0.13
Physical Activity (minutes/wk)		M (SD)	M (SD)
Moderate exercise		170 (213.88)	146 (139.77)	0.55
Muscle strengthening exercise		27 (115.47)	24 (69.44)	0.87
Nutrition Knowledge		M (SD)	M (SD)
Number of questions answered correctly		6 (1.42)	6 (1.21)	0.92

Primary Outcome (Weight)

When all individuals who enrolled in the study (n=154) were included, with the assumption that those who did not complete the exit weight measurement did not lose weight (n=45), 19% lost at least 5% of their baseline body weight. Among the n=109 who completed the follow up assessment, 28% lost at least 5% of their baseline body weight (mean weight loss was 5.3 pounds, greatest weight loss was 43 pounds).

Secondary Outcomes

Table 2 reports pre- and post-intervention means, standard deviations, and effect sizes in self-reported dietary intake, physical activity, and nutrition knowledge (N=109).

Table 2.

Secondary Outcomes (N=109)

	Pre-intervention	Post-intervention
	M (SD)	M (SD)	test statistic	Adj. p-valuea	d b	95% CIc
Self-Reported Food Servings/wk
Fruit	5.6 (4.92)	9.1 (7.67)	−4.24	<.001	0.71	0.50 – 2.01
Fried potatoes	0.9 (1.59)	0.6 (1.15)	2.08	0.06	−0.22	−0.41 – 1.92
Other vegetables	5.1 (6.48)	6.7 (5.82)	−1.91	0.08	0.24	0.05 – 1.96
Soda	2.6 (4.65)	1.2 (2.15)	3.29	<.01	−0.34	−0.53 – 1.91
Fast food	2.3 (1.75)	1.5 (1.00)	2.21	0.06	−0.47	−0.66 – 1.89
Energy drinks	0.5 (1.34)	0.4 (1.31)	0.47	0.64	−0.05	−0.24 – 1.94
Sweetened fruit drinks	1.7 (2.35)	1.3 (2.66)	1.23	0.24	−0.16	−0.34 – 1.93
Coffee/tea with sugar	6.0 (5.67)	4.6 (5.11)	2.22	0.06	−0.25	−0.44 – 1.92
Sweet pastries	3.9 (5.71)	1.5 (2.2)	4.36	<.001	−0.66	−0.87 – 1.87
Frozen desserts	1.4 (4.98)	0.6 (1.57)	1.56	0.15	−0.4	−0.60 – 1.90
Fried foods	1.6 (2.42)	0.6 (0.98)	4.27	<.001	−0.41	−0.60 – 1.90
Red meats	2 (2.06)	1.3 (1.67)	3.80	<.001	−0.34	−0.53 – 1.91
Physical Activity (minutes/wk)
Moderate exercise	170 (213.88)	328 (520.73)	1.66	0.67	-	-
Muscle strengthening exercise	27 (115.47)	33.92 (97.00)	0.38	0.60	-	-
Nutrition Knowledge
	Mdn (IQR) d	Mdn (IQR) d
Number correct	6 (2)	7 (2)	4.25	<.001	-	-

^aAdusted p-values were corrected for multiple testing using the Benjamini-Hochberg procedure with a 5% false discovery rate.

^bCohen’s d.

^c95% Confidence interval for the effect size.

^dMedians and interquartile ranges are reported.

Dietary Intake

Average fruit intake per week was significantly higher post-intervention (M=9.1, SD=7.67) compared to pre-intervention ((M=5.6, SD=4.92, p<0.001, t=−4.242, p<0.001,) constituting a medium effect (d=0.71). Average fried food consumption per week was significantly lower post intervention (M=0.6, SD=0.98) compared to pre-intervention (M=1.6, SD=2.42, p<0.001, t=4.274), with a medium effect (d=−0.41). Other significant differences in self-reported dietary intake per week included decreased consumption of soda (mean decrease=1.4, p=.002, t=3.291, d=−0.34), sweet pastries (mean decrease=2.4, p<.001, t=4.362, d=−0.66), and red meat (mean decrease=0.7, p<.001, t=3.804, d=−0.34).

Physical Activity

The average minutes per week spent on moderate physical activity was higher post intervention (M=328, SD= 520.73) compared to pre-intervention (M=170, SD=213.88, t=1.656), as was the average minutes per week spent on muscle strengthening physical activity (post-intervention: M=33.9, SD=97.00; pre-intervention: M=27.2, SD=115.47, t=0.377). These differences were not significant. Most participants (n=57) did not participate in any vigorous physical activity.

Nutrition Knowledge

Participants’ median post-intervention scores (7, IQR=2) were significantly higher than their median pre-intervention scores (6, IQR=2); Z =−4.246, p<.001.

Discussion

COMIDA had a modest impact on weight loss. Among those who completed the follow-up assessment, nearly a third lost at least 5% of their baseline body weight, potentially conferring clinically significant health benefits [8]. Notably, the mean weight loss of 5.3 pounds was within the range of that observed in other more intensive interventions with minority populations (a meta-analysis found a range of −16.3 to +1.3 pound mean weight change among 13 two-component interventions with an overall 7.7 pound mean weight change among the studies) [32]. Post-intervention, participants consumed more fruit and fewer fried foods, soda, sweet pastries, and red meat.

Effect sizes related to changes in dietary intake post-intervention varied, suggesting that the intervention was more effective in changing certain dietary behaviors than others. Although fast food consumption decreased post-intervention, this decrease was not statistically significant. It may be that for low-income immigrants—the average monthly household income of was well below the federal poverty line—fast food is a “stickier” habit than others. Fast food is an entire meal that requires no preparation and may be more difficult to omit when compared to smaller dietary changes such as adding fruit or opting out of red meat, soda, or pastries. Reasons for this may include parents working long hours for minimum wage or lower [33], dietary assimilation to the majority culture [34], and the food environment (i.e. accessibility of unhealthy food and scarcity of high-quality supermarkets in low-income and minority neighborhoods) [35].

A primary limitation of the study was the lack of a control group. A randomized controlled trial is needed to determine whether the changes observed can be attributed to the intervention. Another is the use of self-report data; participants may have over-estimated post-intervention changes to diet and physical activity due to a desire to report improvements following counseling. In addition, a number of participants did not complete the exit survey and BMI measurement (29%). However, loss to follow-up was less than in other community-based weight loss interventions for Hispanic populations, which in a systematic review ranged from 45–60% [12]. That is potentially because COMIDA was less burdensome for participants by requiring in-person participation only at study intake and exit. Participants may have been self-selected to be more amenable to a lifestyle intervention, as they were already attending a program with a health focus. Future work should consider longer term follow up (e.g. 6 months) to assess maintenance of change, an examination of which of the COMIDA components were most effective, and a cost analysis of intervention components.

Despite these limitations, the present study demonstrates feasibility and provides valuable information about the impact of a culturally-tailored and evidence-based pilot lifestyle intervention to reduce obesity among MAs. COMIDA resulted in modest changes in weight loss and dietary behaviors, comparable to those of more time and resource-intensive interventions with similar populations [12, 15], suggesting a lifestyle intervention with automated, time-saving components such as text messaging can be a scalable, cost-effective approach to address overweight/obesity among underserved and hard-to-reach populations [36, 37]. Taken together with previous studies [12, 15], our findings suggest that conducting a weight loss intervention with trusted community partners, addressing logistical barriers to participation such as time (e.g., by utilizing text messaging and modular delivery of intervention content), including a consideration of cultural influences on diet and physical activity preferences, and utilizing evidence- and SCT theory- based strategies can lead to more positive outcomes in MAs. This pilot study lays the groundwork for the development of future larger scale studies with longer follow-up duration to further assess whether community based, culturally tailored interventions such as COMIDA can lead to maintenance of weight loss and sustained dietary change.