Factors associated with sugar-sweetened beverage and water consumption among adults in rural Baja California

Abstract

Introduction:

This study explored factors associated with sugar-sweetened beverage (SSB) and water consumption among adults in a small community in rural Mexico.

Methods:

We conducted a cross-sectional survey using door-to-door sampling in an unincorporated community. Measures included self-reported consumption of beverages, personal preference, behavioral, and social environmental factors. Hierarchical multivariate linear regression models tested associations between social environmental factors and beverage consumption.

Results:

Participants (n = 89) reported consuming high amounts of sugar-sweetened beverages, averaging 29.9 oz (884.2 mL) per day. Daily water intake was approximately 33.9 oz (1002.5 mL) per day. Personal preference (β = 1.58; 95% CI 1.20, 2.04), behavioral (β = 1.38; 95% CI 1.10, 1.74), and social environmental factors (β = 1.38; 95% CI 1.02, 1.86) were associated with greater daily consumption of SSBs.

Conclusions:

Adults reported consuming high amounts of SSBs and low amounts of water. These behaviors correlated with personal preference, behavioral, and social environmental factors. A multidimensional approach that considers personal, behavioral, and social environmental factors could reduce SSB consumption and improve health outcomes in rural Mexican communities.

Introduction

Excess body weight and obesity are global public health concerns associated with chronic diseases including cardiovascular disease, cancer, and most diabetes cases (1,2). Mexico has among the highest obesity (BMI ⩾ 30 kg/m²) rates globally, ranking second among adults and fourth among children (2). More than one-third (36.9%) of Mexican adults have obesity and abdominal obesity is estimated to be 81.0%, according to the 2022 National Health and Nutrition Survey (ENSANUT) (3). Addressing the underlying causes of excess body weight and obesity is crucial to reducing adverse health outcomes in Mexico.

Consumption of sugar-sweetened beverages (SSBs) is associated with higher rates of overweight and obesity (4–6). SSBs represent a significant source of added sugars in the Mexican diet (7,8), accounting for 68% of added sugars consumed, compared with 37.1% in the United States (U.S.) (5,9). In 2016, 85.3% of the Mexican population self-reported regularly consuming SSBs (10). The average daily consumption of 8 oz (236 mL) servings of SSBs in Mexico is higher (2.6) when compared with other upper middle-income countries like Colombia (1.8), Peru (0.6) and Ecuador (0.5), and twice that of the U.S. (1.3) (11).

A 2020 study of Mexican adults found that 6.9% of all-cause mortality (40,842 excess deaths/year) and 19% of diabetes, cardiovascular disease, and obesity-related cancer mortality (37,000 excess deaths/year) could be attributed to SSB consumption (4). Deaths attributed to SSB consumption are much higher in Mexico (12.1%) compared with Colombia (3.8%), Ecuador (2.4%), Peru (0.4%), and the U.S. (2.3%) (11). Reducing SSB intake is a public health priority in Mexico and necessitates research to understand the factors contributing to consumption (4,6).

Rural Mexican populations face unique challenges related to healthcare access, nutrition education, and healthy food options, which can contribute to higher rates of SSB intake and related health issues (12). One study in a rural and primarily indigenous community of Baja California showed an obesity prevalence of 40.1% among adult women and 25.7% among adult men (13), notably higher than the national averages of 34.5% and 24.1%, respectively (2). The same study reported type 2 diabetes prevalence at 21.8%, 6% higher than the national average of 15.6% (13). There is limited research exploring SSB consumption and its motivators in rural Mexican adult populations.

SSB consumption is influenced by social factors and norms. Mexico experienced a nutrition transition from 1999 to 2012 in which consumption of SSBs increased substantially (14). A national tax introduced in 2014 reduced SSB sales; however, long-term results appeared to wane over time (15,16). Advertising strategies in Mexico widely promote the consumption of SSBs, limiting the impact of taxation (17). Festive occasions and lunchtime menus in Mexico typically include offerings of SSBs, establishing SSB consumption as a social norm (18,19).

Investigating water consumption alongside SSB consumption allows for a fuller understanding of behavioral factors associated with beverage choice. A study of plain water and SSB consumption among Mexican school-age children found that when combined, high consumption of plain water and low consumption of SSB was associated with less total energy intake (20). The physical environment can be a factor and promoting water consumption may be less effective in areas with limited access to water. A recent study found that individuals with intermittent water supply consumed more SSBs (21).

Multiple interventions have aimed to reduce SSB consumption in Mexico, including taxation (22), media campaigns (23), and school-based programs (20). Despite these efforts, SSB consumption remains high, suggesting a need to better understand the drivers of beverage choice.

This study characterized SSB and water consumption behaviors among adults in a small rural community in Baja California, Mexico, and explored potential factors influencing those behaviors.

Methods

Study setting

This study was conducted with a community-engaged program led by one Mexican and two U.S. universities that provides health services and education in a small unincorporated neighborhood in San Quintin, Baja California, Mexico (24). San Quintin is an agricultural region 300 km south of the U.S.–Mexico border that attracts migrant workers from southern Mexico. The study community, Colonia Lomas de San Ramón, covers approximately 0.57 km² and has an estimated population of 4,432 inhabitants (25), 60% of whom identify as indigenous (13). Food and beverage access include small local markets and larger grocery stores located on the main highway.

Participants and recruitment

Data were collected via a household survey in October 2022. Participants were recruited door-to-door and were eligible if they resided in the community, were 18 years or older, and spoke Spanish. Maps of the community were used to identify streets for which teams of trained, bilingual, graduate student data collectors were responsible. Data collectors sampled every other house and used a die to determine which adult member of multi-person households to survey. Data collectors explained the study’s purpose and obtained verbal consent. Participants received reusable water bottles and educational handouts as incentives.

Conceptual model

This study is guided by Triadic Reciprocal Determinism (TRD) from Social Cognitive Theory (26). While SSB consumption is an individual behavior, multilevel factors and external influences drive that behavior (27,28). The TRD posits that health behavior is influenced by multiple factors and focuses on the dynamic interaction between personal, behavioral, and environmental factors (26). Personal factors encompass beliefs, attitudes, expectations, and self-perceptions; behavioral factors include actions, choices, and efforts; environmental factors refer to an individual’s surroundings, both physical and social, and can include other people’s actions and influences (26).

TRD suggests that interconnected personal, environmental, and behavioral factors reciprocally influence each other (26). To better understand the multiple influences on SSB and plain water consumption, we applied TRD to measure constructs within each of the three categories.

Measures

The 110-item survey was adapted from instruments validated in contexts that varied from the study setting, including university students in the U.S. and high school students in Kenya (29–31). Survey items only available in English (30,32) were translated to Spanish using a single-translation approach and pilot-tested with Spanish-speaking Mexican American adults living in the U.S. Adjustments to the questions were made based on feedback from the pilot test and were intended to minimize participant burden and improve clarity. Questions were simplified and shortened for clarity and to be compatible with literacy level.

Beverage consumption

The survey assessed self-reported weekly consumption of 15 beverages, including plain water, based on ENSANUT and supplemented with visual flashcards (29). Store-bought beverages included soda, fruit juice, coffee, tea, flavored milk, aguas frescas, energy drinks, and sports drinks. Homemade beverages were considered SSBs when participants reported adding sugar. To quantify consumption, standard Mexican beverage portion sizes were used, such as one 8 oz (236 mL) cup of coffee and one 12 oz (350 mL) bottle of soda. This approach was guided by the Beverage Intake Questionnaire, which demonstrated that self-reported beverage intake was significantly correlated with actual intake (r = 0.49, P < .01) (30).

Triadic reciprocal determinism

Questions were derived from the validated Spanish-Language Questionnaire on Adolescent Eating Behavior Based on Social Cognitive Theory (Cronbach α = 0.82) (31) and the Nutrition Knowledge, Attitudes, and Consumption of Sugar-Sweetened Beverages Questionnaire (Cronbach α = 0.72) (32) to ensure relevance and appropriateness for the study population.

Personal factors measured included personal knowledge of SSBs and personal preferences for beverages. Behavioral factors included adding sugar to homemade beverages. Environmental factors included both physical access to water and SSBs and social influences including household member behaviors. For personal knowledge items, open-ended knowledge questions were recoded into dichotomous responses for 'correcť and 'incorrecť responses (0 for correct and 1 for incorrect or ‘don’t know’ responses). Composite index scores were created for each TRD construct, encompassing personal knowledge (PK) and personal preference (PP), behavior of adding sugar to beverages (BS), and physical environmental (EP) and social environmental (ES) factors. Survey questions and index scores are detailed in Supplemental Table 1.

Statistical analysis

Data were analyzed using SPSS (SPSS version 28.0.1.1, SPSS Inc., Armonk, NY, 2021). Descriptive statistics including means, standard deviations, and ranges were calculated for continuous variables. Bivariate analysis using Spearman’s correlation coefficient assessed the strength and direction of relationships between TRD constructs (independent variables), sociodemographic characteristics (covariates), and beverage consumption (outcome variables) owing to its suitability for non-parametric data (33).

Several hierarchical multiple linear regression models were constructed with TRD index scores as independent variables, and SSB and water consumption as dependent variables, to determine the variance explained and association with self-reported SSB and water consumption. Multicollinearity was assessed using the Variance Inflation Factor (VIF), and all variables with VIF of less than 2 (VIF < 2) were retained in the models (34). Additionally, bivariate correlations among all independent variables were assessed for collinearity (R ⩾ 0.5).

Sociodemographic factors were placed in block 1 and treated as control variables for this analysis to establish a baseline model, control for confounding, and to assess the incremental variance explained (e.g., R²) by subsequent independent variables of beverage consumption (35). Personal, behavioral, and environmental factors were placed in blocks 2, 3, and 4, respectively, in accordance with the varying groupings of TRD. Log₁₀ transformation was applied to normalize the distribution of SSB consumption. The results of the Log₁₀-transformed data were subsequently back-transformed to the original units to facilitate interpretation of the results. A significance level of P < .05 was used to determine statistical significance. For the back-transformed results, statistical significance was determined when the 95% confidence intervals excluded 1.

Ethical approvals were obtained from the San Diego State University Institutional Review Board (HS-2022–0190) and the Universidad Autónoma de Baja California Bioethics Committee of the School of Medicine and Psychology (1344/2022–2).

Results

Descriptive statistics

Of the 200 households approached, 96 were surveyed, representing roughly 10% of households in the community. Reasons for non-participation were: not at home (n = 90), refusal (n = 11), and ineligibility (n = 3). Of the 96 surveys collected, 89 were included for analysis after excluding seven surveys for incomplete data. Data from two participants with extremely implausible water consumption were excluded.

Descriptive statistics are presented in Table 1. Average daily SSB consumption equated to 29.9 (SD 17.1) oz (884.2 [SD 505.7] mL). Participants consumed an average of 33.9 (SD 18.1) oz (1002.5 [SD 535.3] mL) of water per day and most obtained drinking water outside of the home. Coffee with added sugar and soda were the SSBs most consumed. Men reported higher SSB and water consumption than women.

Table 1.

Socio-demographic, beverage consumption, triadic reciprocal determinism index scores, and perceptions among adults living in rural Baja California, Mexico (n=89).

Demographics	Mean (SD)	Beverage Consumption	Mean (SD)
Age (years)	36 (14)	SSB daily (oz)	29.9 (17.1)
	n (%)	(mL)	884.2 (505.7)
Sex (female)	68 (76)	Water daily (oz)	33.9 (18.1)
Educational attainment		(mL)	1002.5 (535.3)
None	10 (11)
Some Primary	9 (10)	TRD Index Scores	Mean (SD) [range possible]
Primary	14 (16)	Personal Knowledge (PK)	5.26 (1.2) [0–8]
Secondary	31 (35)	Personal Preference (PP)	0.66 (0.7) [0.2]
High School	15 (17)	Physical Environmental (EP)	2.12 (0.56) [0–4]
College	10 (11)	Social Environmental (ES)	2.34 (0.72) [0–3]
Children living in the home	71 (80)	Behavior of Adding Sugar (BS)	1.62 (0.87) [0–3]
Language other than Spanish  Mixteco	12 (14)	Perceptions	n (%)
Triqui	8 (9)	SSBs are easy to find (agree)	79 (88)
Zapotec	1 (1)	SSBs are affordable (agree)	23 (26)
English  Mixteco & English	2 (2) 1 (1)	Children in the community drink a lot of SSBs (agree)	75 (84)
Drinking water source  Outside the home	86 (97)	The community drinks a lot of SSBs (agree)	79 (89)
At home	2 (2)
Missing	1 (1)
Children living in the home drink SSBs	54 (76)

TRD: Triadic Reciprocal Determinism; SSB: sugar-sweetened beverage; PK: personal knowledge; PP: personal preference; BS: behavior of adding sugar to beverages; EP: physical environmental factors; ES: social environmental factors

Bivariate analyses using Spearman’s correlation

Personal preference (PP) and the behavior of adding sugar to beverages (BS) were significantly and positively correlated with daily SSB consumption (r = .35, P < .01 and r = .23, P = .03, respectively). Personal knowledge (PK) and educational attainment were positively correlated with average daily water consumption (r = .21, P = .04; and r = .26, P = .02, respectively), while PP was negatively correlated with average daily water consumption (r = −.36, P < .01). Average daily SSB and water consumption were not statistically significantly correlated (r = −.16, P = .14). The results of the bivariate analyses using Spearman’s correlation are presented in Supplemental Table 2.

Multivariate hierarchical linear regression

The results of the hierarchical regressions are presented in Table 2. The first model, with SSB consumption in ounces as the outcome variable, showed that block 2 (F = 2.90; P < .01), block 3 (F = 4.01; P < .01), and block 4 (F = 3.84; P < .01) significantly contributed to the regression model (all results for SSBs are back transformed). Sociodemographic factors accounted for 6% of the variance, while incorporating personal factors, behavioral, and environmental factors subsequently added an additional 12%, 8%, and 5% of variance explained, respectively, for daily SSB consumption. Altogether, the four blocks accounted for 30% of the total variance in the model.

Table 2.

Hierarchical models of correlates for sugar-sweetened beverage and water consumption.

Sugar-sweetened beverage consumption — daily ounces (n=89)a
	Block 1				Block 2				Block 3				Block 4
Independent variables	R2 = 0.06				R2 = 0.18,R2 ∆ 0.12				R2 = 0.26, R2 ∆ 0.08				R2 = 0.30, R2 ∆ 0.05
Demographics	10β	10SE	10Std. β	1095%CI	10β	10SE	10Std. β	1095%CI	10β	10SE	10Std. β	1095% CI	10β	10SE	10Std. β	1095%CI
Age (years)	0.98	1.01	0.60	0.95,1.002	0.99	1.01	0.78	0.98,1.02	1.01	1.01	0.72	0.76,1.01	0.98	1.01	0.71	0.98,1.01
Sex (female)	1.32	1.29	1.41	0.85, 2.40	1.51	1.29	1.51	0.93,2.51	1.62	1.26	1.58	1.02,2.63	1.66	1.26	1.62	1.02,2.63
Education	0.83	1.23	0.78	0.55, 1.26	0.95	1.23	0.93	0.63,1.44	0.91	1.02	0.89	0.62,1.38	0.85	1.23	0.79	0.56,1.29
Children (y/n)	0.87	1.32	0.89	0.50, 1.55	0.81	1.32	0.81	0.47,1.38	0.76	1.29	0.78	0.46,1.29	0.62	1.32	0.64	0.36,1.07
Personal
Knowledge (PK)					1.02	1.10	1.05	0.85,1.23	0.99	1.10	0.98	0.83,1.20	0.98	1.10	0.93	0.81,1.17
Preference (PP)					1.62	1.15	2.29	1.23,2.14	1.58	1.15	2.19	1.20,2.09	1.58	1.15	2.19	1.20,2.04
Behavioral
Adding sugar (BS)									1.41	1.12	1.95	1.12,1.79	1.38	1.12	1.90	1.10,1.74
Environmental
Physical (EP)													1.12	1.23	1.15	0.76,1.66
Social (ES)													1.38	1.17	1.66	1.02,1.86
Water consumption - Daily ounces (n=87)
	Block 1				Block 2				Block 3				Block 4
Independent variables	R2 = 0.12				R2 = 0.26, R2 ∆ 0.15				R2 = 0.27, R2 ∆ 0.01				R2 = 0.32, R2 ∆ 0.04
Demographics	β	SE	Std. β	95% CI	β	SE	Std. β	95% CI	β	SE	Std. β	95% CI	β	SE	Std. β	95% CI
Age (years)	0.03	0.16	0.03	−0.29,0.36	−0.11	0.15	−0.08	−0.41,0.20	−0.10	0.15	−0.08	−0.40,0.21	−0.11	0.15	−0.09	−0.41,0.20
Sex (female)	9.76	4.62	0.22	0.57,18.95	8.17	4.30	0.19	−0.38,16.72	7.91	4.30	0.18	−0.66,16.47	8.73	4.28	0.20	0.22,17.25
Education	7.54	3.50	0.27	0.59,14.49	3.31	3.48	0.12	−3.62,10.23	3.60	3.49	0.13	−3.36,10.55	6.14	3.65	0.22	−1.13,13.40
Children (y/n)	5.47	4.94	0.12	−4.36,15.30	6.16	4.58	0.14	−2.96,15.28	6.41	4.59	0.14	−2.73,15.54	8.35	4.79	0.19	−1.18,17.88
Personal
Knowledge (PK)					2.00	1.60	0.13	−1.18,5.18	2.18	1.61	0.14	−1.02,5.38	1.87	1.61	0.12	−1.33,5.08
Preference (PP)					−8.94	2.42	−0.38	−13.75,−4.14	−8.77	2.42	−0.37	−13.59;−3.95	−9.11	2.39	−0.38	−13.86,−4.35
Behavioral
Adding sugar (BS)									−2.02	2.02	−0.10	−6.05;2.01	−1.96	2.00	−0.10	−5.94,2.02
Environmental
Physical (EP)													−6.75	3.51	−0.21	−13.74,0.25
Social (ES)													−1.78	2.60	−0.07	−6.96,3.41

^aSSB was log 10-transformed for regression analysis, then back-transformed as 10^x for all parameters.

Education categories: elementary or less (reference), middle school and high school, college or more.

Children living in the home (yes vs. no).

PK: personal knowledge; PP: personal preference; BS: behavior-of adding sugar; EP: physical environmental factors; ES: social-environmental factors.

∆: difference in R² from-previous block.

Boldface values represent significant associations.

In the final block, when controlling for age, sex, educational attainment, children living in the house, and all other covariates, higher personal preference (PP) scores (β = 1.58; 95% CI 1.20, 2.04), behavior of adding sugar (BS) scores (β = 1.38; 95% CI 1.10, 1.74), and social environmental (ES) scores (β = 1.38; 95% CI 1.02, 1.86) were significantly and positively associated with greater daily consumption of SSBs. One-point increases in PP, BS, and ES scores were associated with 1.58 oz (46.7 mL), 1.38 oz (40.8 mL), and 1.38 oz (40.8 mL) increases in SSBs consumed per day, respectively.

Daily water consumption in ounces was the dependent variable in the second regression model. In this model, block 2 (F = 4.8; P < .01), block 3 (F = 4.3; P < .01), and block 4 (F = 3.9; P < .01) were significantly associated with water consumption. Sociodemographic factors accounted for 12% of the variance, while personal, behavioral, and environmental factors added an additional 15%, 1%, and 4% variance, respectively, in daily water consumption. In total, all four blocks accounted for 32% of the variance (respectively) in the model for daily water consumption.

When controlling for age, sex, educational attainment, children living in the home, and all other covariates, personal preference (PP) was negatively associated with daily water consumption. Each one-point increase in PP score for juice or soda was associated with a 9.1 oz (269 mL) decrease in water consumed per day. Other factors were not significantly associated with daily water consumption.

Discussion

This study aimed to characterize SSB consumption and factors influencing water and SSB consumption behaviors among adults in a small rural community in Baja California, Mexico. We found that adults in the study consumed high amounts of SSBs and low amounts of water and that personal preference was the strongest correlate for both SSB and water consumption.

Study participants consumed high amounts of SSBs with daily intake averaging 29.9 oz (884.2 mL)—nearly triple the national average of 11 oz/day (325.3 mL) (36), and well over the American Heart Association’s recommendation not to exceed 36 oz/week (1064.6 mL) (37). Daily water intake was approximately 33.9 oz (1002.5 mL), aligning with another study’s finding of 32 oz/day (946.3 mL) among Mexican adults with low socioeconomic status (SES) (38). The primary sources of SSBs were soda, coffee with added sugar, and store-bought fruit juice. High SSB and low water consumption may be contributing to the community’s high rates of obesity, diabetes, and metabolic syndrome (13), emphasizing the importance of reducing SSB consumption.

The results of the bivariate analysis showed that individuals with low educational attainment consumed less water; however, this association was not significant in the multivariate hierarchical linear regression model. Higher educational attainment was associated with greater personal knowledge of SSBs, consistent with studies showing a positive association between educational attainment and nutrition knowledge (28). In our study personal knowledge was not associated with SSB consumption. It is unclear how educational attainment and personal knowledge influence SSB and water consumption in this population and how these findings should be interpreted to inform intervention development.

Personal preference was the strongest correlate for both SSBs and water consumption compared with all TRD constructs. A large scoping review of factors influencing SSB consumption among global indigenous populations found that personal taste preference outweighed all other factors influencing SSB consumption, which may help explain why interventions targeting factors such as water quality, policy, and education have had limited effects (27).

Our study found that the behavior of adding sugar (BS) and social environmental factors (ES) were positively associated with SSB consumption, suggesting that behavioral and social environmental factors play an important role in individual beverage choices. Interpersonal factors have long been recognized as significant in shaping health behaviors (39). Among Mexican Indigenous rural communities like the one in this study, food and beverage choice hold important social and cultural significance, something that is often ignored by standard nutrition guidelines (40).

Our findings provide insight into the effectiveness of policy efforts to reduce SSB consumption in rural Mexico. The 2014 tax on SSBs reduced sales, but the effect diminished after three years (16). Our study found that the second most consumed SSB was homemade coffee with added sugar, which is not affected by a tax on store-bought SSBs. Similarly, a warning label requirement for ultra processed foods implemented in 2020 may change some behaviors but not added sugars in homemade beverages (41).

Access to clean water is a challenge in many unincorporated communities like the one in this study, where water is brought in on trucks. Intermittent water supply can lead to higher SSB consumption (21). In a large review of factors influencing SSB consumption among Indigenous populations worldwide, availability of SSBs was the second most common factor influencing consumption of SSBs, however efforts to reduce consumption through increasing cost of SSBs or reducing the costs of alternatives were not effective (27). While participants in our study reported that SSBs were easy to access in the community, the association between the physical environment and SSB consumption was not significant.

Our study has certain limitations. The sample size for this study was small and thus potentially underpowered for some of the analyses. An a priori sample size and power calculation was not conducted. A post-hoc power calculation suggested that the achieved power ranged from 20% to 95% for a sample size of 89, depending on the associations tested. There is risk of a type 1 error due to the multiple statistical tests conducted. The study was limited to one community, limiting generalizability to other areas within the region. Nearly half of approached homes were unoccupied during data collection, potentially introducing systematic bias between respondents and non-respondents. The survey methodology may have introduced both self-report and recall bias. For survey items not available in Spanish and requiring translation, we used a single-translation approach, which may have led to loss of the intended original meaning. We did not measure exact amounts (e.g., grams) of sugars in SSBs, limiting our ability to estimate the energy intake from added sugars. Another limitation concerns the measurement of TRD constructs. For several TRD questions, the answer choices were changed from the original format to dichotomous (e.g., yes vs. no) to simplify answer choice selection, therefore the summary variables resulting from these data more accurately represent an index score rather than the original construct. Because of this modification, assessment of internal consistency would not provide a meaningful estimate of construct reliability.

Conclusions

This study identifies the need for contextual approaches to reducing SSB consumption. Our study highlights the limits of taxation on reducing SSB consumption. While taxation can lower purchases of store-bought SSBs, results appear to be short term, and taxation does not influence consumption of homemade beverages with added sugar. More research is needed to understand how personal preference could be targeted as an intervention strategy to support increased water consumption and decreased consumption of SSBs among rural Mexican populations consuming high amounts of SSBs. Given the role of taste preference and the behavior of adding sugar in SSB consumption, interventions might test the acceptability of alternatives to SSBs based on taste preferences. Future interventions could more effectively engage rural Indigenous communities in Mexico by considering the social environmental and cultural roles that food and beverages play and developing tailored strategies for behavior change. A multidimensional approach that considers personal, behavioral, and social environmental factors could effectively reduce SSB consumption and improve health outcomes in rural communities in Mexico.

Supplementary Material

Supplemental material for this article is available online.