Physical Activity in Older Mexican Americans Living in Two Cities on the U.S.-Mexico Border

Background: There is limited information on physical activity in marginalized older populations like that on the U.S.-Mexico border. This study aims to understand physical activity engagement among older Hispanics residing in two U.S.-Mexico Border counties. Methods: The International Physical Activity Questionnaire (IPAQ) was used to measure physical activity in El Paso and Cameron County, Texas. Physical activity levels were reported for vigorous, moderate, and walking met/mins. Adjusted and unadjusted modeling was conducted to determine county differences and sociodemographic covariates. Results: There were 784 participants and 92.9% were less than 65 years of age. El Paso participants reported a significantly greater natural log met/mins of vigorous (β = 1.34, p = 0.000) and walking (β = 0.331, p = 0.006). Significant sociodemographic covariates in El Paso for vigorous met/mins were gender (females β = −1.20, p = 0.003), having a regular doctor (β = −0.779, p = 0.029), and acculturation (β = 0.513, p = 0.019). Significant associations in Cameron County were having a regular doctor (β = −1.03, p = 0.000) and fair/poor health status (β = −0.475, p = 0.001). Conclusion: Level of physical activity may differ in older Hispanics by urban context on the U.S.-Mexico border. Future physical activity programs to promote physical activity should take context into consideration.


Introduction
There is compelling evidence that an insufficient amount of physical activity is associated with an increased risk of chronic metabolic, cardiovascular disease, and several forms of cancer [1][2][3][4]. The scientific evidence of the benefits of regular physical activity is indisputable, and the benefits far outweighs the risks for most adults. Current Centers for Disease Control and Prevention (CDC) guidelines recommend that older adults engage in either moderately intensive cardiorespiratory physical activity training (brisk walking) for at least 30 min five or more days per week (150 min per week) or vigorous cardiorespiratory physical activity training (jogging, running, aerobic dance) for 20 min or more per day for three days or more per week [5]. Older adults with medical comorbidities can safely participate in physical activity with the advice of their health care professional, and regular physical activity provides substantial benefits in reducing chronic disease symptoms and comorbidities. [5]. Furthermore, engaging in these activities can help achieve weight loss and maintain a lower body mass index (BMI) [6]. address, due for colorectal cancer screening at time of recruitment, and no blood in the stool for the previous three months.

Dependent Variable
International Physical Activity Questionnaire (IPAQ). We made use of the IPAQ to measure physical activity across an inclusive set of activities, such as leisure time, domestic, work-related, and transport-related movements [35]. The IPAQ (short version) was used to measure vigorous activity (hard physical effort that makes you breathe much harder than normal), moderate activity (moderate physical effort that makes you breathe somewhat harder than normal), and walking (including at work, home, solely for recreation, sport, physical activity, or leisure) [36,37]. Participants were asked to report the number of days per week and the number of hours/minutes per day these activities were performed. The IPAQ has a reliability of 0.80 and validity of 0.30 internationally [37].
Using the energy requirement criteria determined for metabolic equivalents (METs) in physical activity testing, all responses for activity measures were converted from hours into minutes to produce a continuous score. MET is the ratio of a person's working metabolic rate relative to the resting metabolic rate. One MET is defined as the energy cost of sitting quietly and is equivalent to a caloric consumption of 1 kcal/kg/h [36]. For each specific type of activity, the frequency (measured in days per week) and duration (time per day) were recorded; vigorous activity was defined as 8 METs, moderate as 4 METs, and walking as 3.3 METs. Each form of activity was assigned a formula in order to determine a score for the participant calculated by multiplying its METs by frequency (days per week) and by duration (time per minutes). The subject's overall physical activity was calculated by using the total physical activity MET-minutes/week summing (walking + moderate + vigorous) MET-min/week scores. Listed below are the designated formulas for each activity [37]: Walking MET-min/week = 3.3 × walking minutes × walking days. Moderate MET-min/week = 4.0 × moderate-intensity activity minutes × moderate days Vigorous MET-min/week = 8.0 × vigorous-intensity activity minutes × vigorous intensity days. We made use of the three categories-vigorous, moderate, and walking met/mins-for our analysis.

Analysis
Statistical analysis was conducted using Stata MP 14 [38]. Continuous variables were described using mean, standard deviation, and frequency, and percentages were used to describe categorical variables. We used Spearman's correlation coefficient test to evaluate the association between IPAQ scores and continuous demographic variables. Pearson's chi-squared test was used to calculate scores between categorical demographic characteristics and the association with physical activity by county.
p-values less than 5% were considered significant results. Adjusted multivariable ordinary least squares (OLS) regression analysis was conducted to assess differences by site for each IPAQ outcome. Because of significant skewness in all three IPAQ outcomes, natural log transformations were used in this analysis.

Results
The majority of the participants were over 65 years of age (92.9%), were female (78.6%), and born in Mexico (90.3%) (see Table 1). Additionally, most had less than a high school level of education (77.8%), made less than $10,000 per year, and spoke mostly Spanish (average score 1.31 out of 5). Most were married (63.2%), rated their health as excellent, very good, or good (55.2%), never smoked (59.4%), and did not have a regular doctor (61.8%). Significant differences between El Paso and Cameron Counties were gender, educational attainment, and health status. When compared to Cameron County, a greater proportion in El Paso were male (25.3% vs. 15.6%, p-value 0.001), a lower proportion with high school or greater education (74.3% vs. 84.5%, p-value 0.000), and rated their health as excellent/very good or good (58.0% vs. 50.9%, p-value 0.051). Finally, near significant differences were observed for having a regular doctor (El Paso 37.3% vs. Cameron 44.4%, p = 0.081). Table 2 presents the unadjusted associations between demographic variables and the IPAQ outcomes by county. Only significant covariates from the univariate analysis in Table 1 were used. Cameron County participants reported on average less total vigorous activity than in El Paso (283.6 vs. 1641.1 met/mins). Significant differences were observed in El Paso by gender, whereas men had on average higher met/mins of vigorous (3325.8 vs. 1071.5, p < 0.001), moderate (1022.3 vs. 482.1, p < 0.01), and walking (1039.6 vs. 528.6, p < 0.001) compared to women. In El Paso, participants reported higher average met/mins of walking in the unadjusted analysis if they had a greater than high school educational attainment (888.3 vs. 573.4, p < 0.05), speak more English than Spanish (1338.5 vs. 576.2, p < 0.05), are current smokers (1040.9 vs. 470.1, p < 0.05), and did not have a regular doctor (805.9 vs. 424.4, p < 0.01). In Cameron County, participants who reported not having a regular doctor reported higher average met/mins of vigorous activity (386.7 vs. 148.6, p < 0.05) but lower average walking met/mins (185.3 vs. 301.8, p < 0.01). Also, in Cameron County, participants who reported their health as excellent, very good, or good reported higher average met/mins of walking (284.5 vs. 189.9, p < 0.05). Table 3 presents the results stratified by county and the between-county comparison OLS regression results for the natural log transformation IPAQ outcomes. We ran two models. In the first model, we compared El Paso to Cameron County, adjusting for gender, education, health status, regular doctor, smoking history, and acculturation. In this analysis, participants in El Paso reported 3.8 times higher average natural log IPAQ met/mins of vigorous activity (β = 1.34, p = 0.000) and 1.39 times higher for walking (β = 0.331, p = 0.006). We then ran separate models for El Paso and Cameron Counties to determine significant within-county covariates for our IPAQ outcomes. Females in El Paso on average engaged in 26.4% (β = 1.33, p = 0.001) less vigorous physical activity than men. Participants who had a regular doctor had a 45.8% (β = −0.780, p = 0.027) lower average vigorous met/mins in El Paso and 64% lower in Cameron County (β = −1.03, p = 0.000). Additionally, El Paso participants had on average a 167% increase in vigorous activity for each point increase in acculturation (β = 0.513, p = 0.019). In Cameron County, health status was associated with 38% less met/mins of walking (β = −0.475, p = 0.001) for participants rating their health as fair or poor.

Discussion
This study's findings revealed that on the U.S.-Mexico border, there were significant differences by county. Participants in El Paso were nearly three times more active when compared to Cameron County participants. Our findings were consistent with the leisure time physical inactivity prevalence documented by national surveillance programs like the Behavioral Risk Factor Surveillance Survey (BRFSS) (El Paso 27.0% vs. Cameron County 36.8%) [39], adding to the narrative that along the U.S.-Mexico border, there are contextual differences in physical activity engagement. Socioeconomic predictors did vary by context, although they did not fully explain the differences between El Paso and Cameron Counties for vigorous or walking activity.
Even adjusting for several individual factors, differences continued to exist for both vigorous and walking activity by county. These findings may suggest that there are unmeasured attributes of each location that may be responsible for these differences. Despite both being on the U.S.-Mexico border, El Paso is a much larger county than Cameron County and, as such, may have more amenities for physical activity, such as better access to parks, trails, and bike paths, which may in turn be responsible for the higher levels of physical activity. In the few studies that have compared metropolitan areas, size and socioeconomic context have been identified as important factors in physical activity amenities [40,41]. Although the exact mechanisms are not well known, individual and community resources may limit or enable real or perceived options for activity. More research is needed to better understand contextual variations along the border, such as rural/urban or built environment policies that may contribute to differentials in physical activity engagement by county in this region.
Many studies have established socioeconomic correlations to physical activity resources at a neighborhood level [42][43][44], but few have done so by metropolitan area. In this study, participants in Cameron County that had a regular doctor or fair/poor health engaged in less physical activity. Socioeconomic characteristics that were associated with physical activity in El Paso were gender, having a regular doctor, and acculturation. Other studies have shown similar levels of physical activity and socioeconomic predictors in Cameron County [34], but no study has compared the same socioeconomic characteristics to different contexts in the United States or on the border [45,46]. Participants who identified as having a regular medical provider may be less active because of existing health conditions (e.g., fair/poor health status) in Cameron County. Healthcare access is limited in Cameron County, and it is recognized as medically underserved [47]. Less access, in this context, may translate into poorer overall health and less capacity to engage in physical activity.
In the context of El Paso, acculturation and male gender were predictive of more physical activity engagement, which is consistent with the literature from other contexts [48]. Men have been well documented to engage in physical activity more often than women [49,50], although few studies have compared gender differences in Hispanic groups [51]. Acculturation has been well studied in the Hispanic health literature [52,53]. In general, greater acculturation to the United States is associated with more physical activity but a less healthy diet [54][55][56]. El Paso, due to its size and fewer socioeconomic burdens, may be more in line with other large U.S. cities, despite being on the border, and less similar to more rural, less developed border areas like Cameron County, resulting in similar predictors to other large metropolitan areas. More research on border context comparisons are needed to disentangle what is truly border from rural/urban or other factors that may influence physical activity that may distinguish one are from the other.
There are a number of limitations to this study. Most notably, the data was collected from only El Paso and Cameron Counties, and inferences cannot be made to other areas of the U.S.-Mexico border or other Hispanic subpopulations. Additionally, since this data is cross sectional, we were unable to establish any causal relationships. Furthermore, while the IPAQ is widely used, it is a subjective measure of reporting physical activity which could over-or underestimate actual levels of activity. Finally, participants in the ACCION program were recruited because they were overdue for colon cancer screening and uninsured, so the sample represents a subpopulation of those living on the U.S.-Mexico border.
The strengths of this study include the large sample of older adults, the comparison of two metropolitan areas on the border, and the use of a widely used, validated measure of physical activity. Future studies should investigate the age-specific benefits of physical activity in older Hispanic populations living in different contexts in the U.S.-Mexico border region and other areas across the U.S. In addition, clinic-based interventions using the primary care physician (PCP) to counsel patients about age-appropriate physical activity levels may be an effective approach to increasing physical activity levels in this population. There is limited physical activity education currently being conducted in clinical settings, particularly among Mexican American populations [57,58]. Other areas where further investigation is needed is a comparison of different Hispanic subgroups and variations in levels of physical activity that may be impacted by context.

Conclusions
Overall, physical activity levels are much lower on U.S.-Mexico border than what has been documented nationally. In addition, Cameron County, a more socioeconomically burdened county on the U.S.-Mexico border, may be at increased risk for obesity and associated chronic diseases due to high levels of physical inactivity established in this study. Physical activity is a low-cost solution to a growing epidemic of chronic disease in this region, and finding ways to increase activity will in the long run reverse current trends and improve the wellbeing of the older population of Mexican Americans.
Author Contributions: G.V. lead author, primary writer of paper content; J.S. assisted with writing, editing and formatting paper; J.M. assisted with writing and editing paper; G.S. assisted with writing and proofreading paper; S.F.-L. conducted data analysis; A.A. conducted data analysis; N.K.S. study PI, assisted with writing and editing paper.

Conflicts of Interest:
The authors declare no conflict of interest.