Background

The burgeoning of non-communicable diseases (NCDs) worldwide brings a growing threat to public health and to economies: costs are projected to exceed $30 trillion over the next 20 years [1, 2]. NCDs now account for more than two-thirds of deaths globally [3], and in the United States (US), NCDs including cardiovascular disease, cancer, chronic respiratory disease and diabetes, contribute to 7 out of 10 deaths [1, 4, 5]. NCDs also disproportionately affect low-income and minority populations in the US [6]. They are pronounced among Mexican Americans, especially those living along the US-Mexico border, the geographical region extending north from Mexico into the US for 100 km [7]. Recent research shows that Hispanics living along the US-Mexico border experience levels of NCDs resembling the morbidity and mortality rates found in low-and middle-income countries [1, 8].

NCDs are among the most preventable of diseases, primarily because they are a result of individual health behaviors [4]. Preventive health behaviors, including smoke-free living, physical activity, fruit and vegetable consumption, and avoidance of heavy alcohol use, help to prevent or alleviate the suffering and early mortality related to NCDs [4] (see Table 1). It is estimated that if behavioral and environmental risks associated with NCDs were removed it is estimated that at least 80 % of all heart disease, stroke, and Type II diabetes and more than 40 % of cancer cases could be prevented [9]. However, little research examined behaviors preventive of NCDs among Hispanics, particularly along the US-Mexico border. This study therefore examined the extent to which preventive health behaviors were self-reported among Hispanics, primarily of Mexican descent, in two weighted samples, one from the US as a whole and the other from the US-Mexico border. Given the documented chronic disease prevalence along the border, we expected to see fewer preventive health behaviors, including smoke-free living, physical activity, fruit and vegetable consumption, and avoidance of heavy alcohol use in Hispanics living along the border than in Hispanics nationally.

Table 1 Guidelines and data on preventive health behaviors associated with non-communicable diseases (NCDS)
Full size table

Methods

The study used two datasets containing detailed information on Hispanic populations. One sample included 29,942 Hispanic respondents from the national Behavioral Risk Factor Surveillance System (BRFSS) in 2009 [10]. The other sample included a total of 1,408 participants from the Cameron County Hispanic Cohort (CCHC) who were enrolled between the years 2008 and 2011 [11].

BRFSS is a national on-going telephone survey of self-reported health conducted yearly by the Centers for Disease Control and Prevention [10]. Data were collected over a 12 month period in 2009 for all states and territories of the United States [10]. For this study, the sample was restricted to respondents self-identified as Hispanic. Identification of Hispanic subgroups by country of ancestral descent is unavailable via BRFSS, but, 63 % of Hispanics nationwide identify as of Mexican descent [10].

The CCHC, a prospective cohort study, began in 2004 and includes adults aged 18 years and older living in Brownsville, TX, a low-economic midsize city along the US-Mexico border with a population that is 93.2 % Hispanic, with 86.2 % of Mexican descent [12]. Cohort participants were recruited from randomly selected households using a two-stage cluster systematic sampling method. Participants received clinical examinations, answered a behavioral health questionnaire, and had anthropometric measures taken; a description of these has been previously published elsewhere [11]. Surveys were conducted in the participant’s preferred language (English or Spanish) by trained bilingual staff. Data on tobacco use, physical activity, fruit and vegetable consumption, and alcohol use were drawn from participants’ behavioral measures. Participants of Mexican descent were identified by ancestry variables when a participant, parent, or grandparent indicated nativity in Mexico. The Committee for the Protection of Human Subjects at the University of Texas Health Science Center at Houston approved all protocols and consent forms.

Smoke-free living

To assess the prevalence of smoke-free living, CCHC participants were asked, “Do you now smoke cigarettes?” with a response of “yes” labeled as smoker and “no” labeled as nonsmoker. BRFSS respondents were asked, “Do you now smoke cigarettes every day, some days, or not at all?” with “every day” and “some days” responses collapsed to indicate a smoker, and a response of “not as all” labeled as nonsmoker [15].

Physical activity

To assess the extent of physical activity, CCHC participants were assessed using validated scales to examine moderate and vigorous physical activity of at least 10-minute bouts during the past 7 days [16, 17]. On the BRFSS, respondents reported the number of days per week and the total time per day of moderate and vigorous physical activity of at least 10 minute bouts [15]. CCHC and BRFSS survey items were used to estimate metabolic equivalents (METs) minutes of energy expenditure in the past 7 days, where participants with > 600 MET adjusted minutes were considered to meet physical activity guidelines.

Fruit and vegetable consumption

Using the Two-item Food Frequency Questionnaire [18], fruit and vegetable consumption was assessed in the CCHC by asking participants how many portions of fruit and vegetables they ate daily. A portion size was described as a ½ cup of fresh, frozen, or canned produce or a medium-sized piece of produce. Consumption of five or more fruit and vegetable portions was considered meeting guidelines [19, 20]. On the BRFSS, participants reported the number of servings of fruits and vegetables daily, and those who reported five or more servings per day were considered to meet guidelines [15].

Heavy alcohol use

CCHC participants were asked the number of alcoholic drinks (12 oz. beer, 8 oz. wine cooler, 4 oz. wine, or 1 oz. liquor) consumed each week. The total was divided by 7 to obtain an estimated daily number of alcoholic drinks consumed. In the BRFSS, one alcoholic drink was measured as a 12 oz. beer, a 5 oz. glass of wine, or one shot of liquor [15]. For both samples, heavy alcohol use for males was identified as having more than two drinks per day and for females having more than one drink per day [15].

Sociodemographic variables

Data on age, gender, educational attainment (high school diploma or less than high school diploma) employment (employed or unemployed), preferred language (English or Spanish), and health insurance (insured or uninsured) were calculated on both samples.

Statistical analysis

To derive representative population-based estimates, all analyses were performed using sampling weights for the CCHC and the BRFSS samples. The two samples were compared based on sociodemographic characteristics and preventive health behaviors using generalized estimating equations (GEE) correlated outcomes to account for potential clustering among participants within CCHC households and census tracts and blocks. Specifically, to compare proportions by dataset (CCHC or BRFSS) and obtain crude and adjusted p-values, we used univariate and multivariate GEE regression models, respectively, with logit link function for binomial responses. Similarly, to compare means by dataset and obtain crude and adjusted p-values, we used univariate and multivariate GEE regression models with identity link function for continuous responses. All statistical tests were two-sided with a significance level of 0.05. Analyses were performed using SAS version 9.1 [21].

Results

On average, sampled participants from the CCHC were significantly older than BRFSS participants (CCHC: 47.22 ± 1.0 BRFSS: 40.47 ± 0.2, p-value < 0.0001) (Table 2). The CCHC sample also had a higher proportion of females than the BRFSS sample (CCHC: 58.3 %, BRFSS: 49.7 %, p-value < 0.0004), and a higher percentage of participants who did not graduate from high school (CCHC: 47.9 %, BRFSS: 31.2 %, p-value < 0.0001). Among CCHC participants, the percentage employed was significantly lower than among BRFSS participants (CCHC: 47.6 %, BRFSS: 55 %, p-value = 0.0041). When stratified by gender, there were fewer employed males (CCHC: 63.8 %, BRFSS: 65.6 %, p-value 0.6493) and fewer employed females (CCHC: 36.0 %, BRFSS: 44.3 %, p-value = 0.0017) in the CCHC sample.

Table 2 Sociodemographic characteristics of CCHC (2008–2011) and BRFSS (2009) Hispanic respondents
Full size table

Table 3 summarizes the preventive health behaviors of smoke-free living, meeting physical activity guidelines, fruit and vegetable consumption, and avoiding heavy alcohol use. After adjusting for age, educational attainment, employment, preferred language, and insurance, BRFSS respondents were significantly more likely than CCHC participants to meet recommended physical activity guidelines (44.14 % vs. 33.3 %, adjusted p-value = 0.0186). BRFSS respondents were also significantly more likely than CCHC participants to meet fruit and vegetable consumption guidelines (21.93 % vs. 14.8 %, adjusted p-value = < 0.0001).

Table 3 Preventive Health Behaviors of ncds among CCHC (2008–2011) and BRFSS (2009) Hispanics respondents
Full size table

Table 4 summarizes preventive health behaviors stratified by gender for both samples. When adjusted for age, educational attainment, employment, preferred language, and insurance, BRFSS male respondents were significantly more likely than CCHC male participants to meet fruit and vegetable consumption guidelines (46.61 % vs. 39.8 %, adjusted p-value = 0.0034). BRFSS male respondents were also significantly more likely to avoid heavy alcohol use (93.86 % vs. 91.0 %, adjusted p-value = 0.0236). Among females, BRFSS respondents were significantly more likely to meet physical activity guidelines (41.68 % vs. 39.8 %, adjusted p-value = 0.0015) and to consume adequate fruits and vegetables (25.16 % vs. 11.4 %, adjusted p-value = 0.0417).

Table 4 Preventive Health Behaviors of NCDs among CCHC (2008–2011) and BRFSS Hispanics respondents (2009) by gender
Full size table

Discussion

This study found Hispanics nationally and Hispanics living on the US-Mexico border reported low engagement in preventive health behaviors, and Hispanics living on the border engaged the least. These findings are important since Hispanics are the fastest growing population in the US and thus will influence future wellness projections and health care costs. Previous reports on disease rates from the US-Mexico border have shown significantly higher rates of overweight, obesity, and diabetes than in national samples [9, 10, 22, 23]. The results of this study suggest that the population along the US-Mexico border will continue to have higher levels of NCDs. The study found that Hispanics living on the US-Mexico border were significantly less likely to engage in the preventive health behaviors of physical activity and fruit and vegetable consumption than Hispanics nationally. Also Hispanic males nationally were significantly more likely than Hispanic males on the US-Mexico border to avoid heavy alcohol use.

The limitations of this study include differences in data collection and sampling between the two datasets. In-person interviews were used to collect behavioral information from participants in the CCHC, while BRFSS data were collected through telephone interviews. This difference introduced potentially different respondent – interviewer biases since they did not sit face-to-face during the BRFSS interviews as they did in the CCHC study. However, wording of questions was similar for both samples and coding for preventive behaviors of interest was standardized in the analysis.

It should also be noted that the majority of CCHC participants were of Mexican descent, while BRFSS respondents were probably more diverse in their representation of Latin-descent populations. However, while past research on preventive health behaviors has shown differences between Latin descent populations in alcohol use [24–26], tobacco use [27–30], and physical activity [31–33], there are larger differences between this group and other racial/ethnic groups [32–37]. A final limitation to this study is the possible under-sampling of those with severe disparities among BRFSS participants resulting from the intent of BRFSS to produce results generalizable to the entire US. When considering national surveys, it is thus important to pay attention to under-sampled, understudied, but highly disparate areas such as the US-Mexico border.

Conclusions

Despite its limitations the study provides clear evidence of disparities in preventive health behaviors for NCDs among Hispanics along the US-Mexico border. The lack of preventive health behaviors among Hispanics along the border is particularly disturbing because it mirrors features found among low and middle income populations worldwide, where NCDs are escalating [3]. High rates of poverty among the rapidly growing population along the US-Mexico border, low levels of educational attainment, and lack of health insurance increase the complexity of addressing NCDs. Common obstacles faced in addressing disparities in preventative health behaviors include food insecurity, abundant access to unhealthy foods, and insufficient environmental infrastructure to promote preventive behaviors [38]. Our findings suggest that health resources should be used to disseminate and implement evidence-based practices, programs, and policies to promote preventive behaviors. A multipronged approach would span the ecological model [39] such that preventive health behaviors could be aligned with system and policy changes to support individuals in adopting and maintaining such behaviors long-term. Our findings also suggest a need for synergistic approaches to nurturing multiple preventive behaviors, rather than addressing each one as an isolated behavior. Given the disparities found in NCDs faced by Hispanics nationally and on the US Mexico border, particularly, multipronged approaches to foster preventive health behaviors are overdue.