In Los Angeles County (~9.8 million residents), health disparities are striking among economically disadvantaged communities [1]. Obesity prevalence is highest among cities with the greatest indices of economic hardship (Table 1). East Compton, for example, has one of the highest rates in the county (39.9%, the city is economically ranked last out of 127 communities), while the city of San Marino has one of the lowest (8.4%, economically ranked first). These marked disparities are observed by race and ethnicity as well, with obesity being more pronounced among Latinos (29.4%) and African Americans (29.2%) [1, 2]. Collectively, this community snapshot paints a picture of significant health disparities in the region [3, 4].

In the literature, factors such as demographics, geography, culture, community resiliency, and access to affordable, healthy foods have been found to be important mediators of obesity risk [5, 6]. Emerging evidence suggests that, to reduce this risk, interdisciplinary interventions—especially in nutrition—should be implemented across multiple sectors (e.g., healthcare, public health, education, transportation, and food environments) [5–7]. In applying this evidence, federal and local health authorities have begun to take notable actions; that is, many recent federally funded obesity prevention efforts have employed an array of practice-based system and environmental (SE) change strategies to improve food environments across the United States [5, 7]. Between 2010 and 2012, for example, the Centers for Disease Control and Prevention (CDC) Communities Putting Prevention to Work (CPPW) program targeted health inequalities in several underserved communities in Los Angeles County. Through this funding, the Los Angeles County Department of Public Health implemented a number of nutrition interventions in the region. These interventions included (a) modifying food services and vending practices at food venues operated by county and city governments (e.g., incorporating healthy nutrition standards through the contracting process with food vendors or suppliers) [8]; (b) converting corner stores or other stores in low-income neighborhoods to food outlets which offer more fresh fruits and vegetables; and (c) utilizing outreach and health marketing to educate the public about the adverse effects of excess sugary drink consumption (Table 2) [9].

As in other communities, assuring community acceptance of SE modifications to the food environment requires in-depth knowledge and understanding of the key health behaviors and characteristics of targeted subpopulations [10]. To date, few ongoing public health strategies have tailored intervention programs to address these groups' unique needs. Access to more granular, community-level health data could change this practice by helping to better inform and guide planning and program improvements in these communities.

Capitalizing on the results from a local health and nutrition examination survey, we contribute to this gap in public health practice by studying a population that was exposed to and may have been affected by these and other nutrition interventions implemented in urban Los Angeles County during 2010–2012. The study examined local health data including predictors of healthy eating among a subset of low-income adults who receive free/low-cost services from multipurpose, public health centers in the jurisdiction. Policy and practice implications are discussed within the context of program improvement and future obesity prevention planning for the region.

Data from the first round of a local health and nutrition examination survey in Los Angeles County was collected during the first 15 months of the CPPW obesity prevention program. The survey included a subset of adults residing in low-income neighborhoods (verified using residential zip codes). Information collected by the survey included (a) objectively measured height and weight; (b) objectively measured waist circumference and blood pressure; (c) self-reported smoking status; (d) self-reported dietary behaviors; (e) ratings of self-efficacy in healthy eating and exercise; and (f) sociodemographics.

Of the 1,393 prospective survey participants approached, 983 met eligibility criteria and were scheduled appointments. Of these, a total of 720 were low-income adults and completed the survey for a response rate of 74% for the subset. A large proportion of participants were black (40%) or Latino (34%), between the ages of 25 and 44 years (48%), and women (57%). More than one-third had a high school education or less (39%), nearly one-quarter were college graduates (22%), and over one-half were unemployed or underemployed (58%). Approximately two-thirds (68%) of the participants were overweight and/or obese; 30% were in the prehypertension range based on objectively measured blood pressure readings (Table 3). Although only 28% were reported to be current smokers, approximately 63% indicated exposure to second-hand smoke. In general, fruit and vegetable consumption was relatively low in the group, with only about one-fourth consuming four or more servings of fruits and/or vegetables per day (26%).

In the comparison analysis (see Table 4), Latinos were more likely than whites to be overweight and obese (adjusted odds ratio [AOR] = 3.9, 95% confidence interval [CI] = 2.2, 6.9). Similarly, blacks were more likely than whites to be overweight and obese (AOR = 2.1, 95% CI = 1.2, 3.5). Latinos were generally less likely to smoke, as compared to whites (AOR = 0.4, 95% CI = 0.2, 0.7). Based on objectively measured blood pressures, Latinos and blacks experienced a greater burden of elevated blood pressure readings than whites: 49% of Latinos (AOR = 1.2, 95% CI = 0.7, 2.0) and 55% of blacks (AOR = 1.4, 95% CI = 0.8, 2.4) had readings in the prehypertension and hypertension ranges.

In multivariable regression analyses (see Table 5), being a woman and having a high self-efficacy for regularly reading Nutrition Facts labels were strong predictors of high fruit and vegetable consumption. In the final model, women were 1.5 times (95% CI = 1.0, 2.1) more likely than men to consume 4+ servings of fruits and vegetables per day. Participants with high self-efficacy in reading Nutrition Facts labels were 2.4 times (95% CI = 1.7, 3.5) more likely than their counterparts (with low self-efficacy) to do the same. The Hosmer and Lemeshow Goodness-of-Fit test indicated that these models were compatible with the data presented (χ² = 5.57, P = .70).

Guided by a socioecological framework [5], the present study conducted a series of analyses to examine key characteristics of a subpopulation disproportionately affected by overweight and obesity in Los Angeles County [1]. This priority group is one of several vulnerable groups targeted by a number of nutrition-focused obesity prevention interventions in the region (Table 1) [7, 8]. Although prior efforts have relied on national and/or county surveillance databases to aid program planning [1, 2, 4, 10, 20], this study is among the first to collect more granular, community-level health data that are representative of the groups targeted by program interventions that sought to make changes to the food environment. These data have implications for quality improvement, especially for local health authorities and community-based organizations seeking to improve or better tailor program delivery to their intended audiences [21].

While emerging evidence supports the use of system and environmental change strategies [5, 7, 22], there remains a paucity of research that has fully elucidated the interactions between these structural modifications and individual health behavior change. To achieve meaningful outcomes in community and individual health, interventions often require substantive tailoring to match the needs and the unique social, epidemiological, and ecological characteristics of the target subpopulations [21, 23]. Oka and colleagues (2013), for instance, analyzed a community-based epidemiologic survey using multilevel modeling to better understand area-based variations in obesity [24]. They demonstrated differences in obesity prevalence by gender and race/ethnicity at the neighborhood level and concluded that, to be effective, future interventions/programs should address these and other neighborhood-specific characteristics.

In the present study, the sampled population had high prevalence of overweight and obesity; this was accentuated for Latinos and blacks. This high prevalence, however, is not uniquely different from the documented evidence in the literature for US minorities [6, 25]. In the literature, disparities in obesity burden, including associated conditions such as hypertension, generally clustered in vulnerable groups, frequently confounded by multiple social and environmental factors that are not solely explained by socioeconomic status [26]. These factors have included but are not limited to racism [27, 28], residential segregation [29], and the built environment [30].

The most striking finding in the study was that gender and self-efficacy were strong predictors of healthy eating (e.g., fruit and vegetable consumption), even after controlling for a number of confounding variables including other demographics in the sampled group. This was somewhat unexpected given that the subset of adults included in the analysis represented a source population believed to be ready for and would benefit from structural changes made to the food environment (e.g., healthy food procurement, corner store conversion, 100% healthy vending machine policy, and competitive pricing of healthy foods in food venues). However, experiential information suggests that due to perceived lower educational attainment and poor nutrition in this population, differential patterns of receptivity or readiness to capitalize on these changes were likely common. As such, after adjusting for covariates such as age, race, and education, survey participants were significantly more likely to consume fruits and vegetables than other participants when they were women and had higher self-efficacy in reading Nutrition Facts labels.

From a practical standpoint, intervening with women who typically make food selection decisions for their entire household (nutrition gatekeepers) may be advantageous to the overall effort to reduce obesity in Los Angeles County, as it can concurrently model positive changes in the diet of the entire family unit and can be applied across generations [31, 32]. Intervening with women can also indirectly target other members of the household, especially men who often eat poorly [33, 34]. The potentially additive effect that may result from this action could augment the structural changes (i.e., through system and environmental interventions) made to the food environment by the recent federal and local obesity prevention initiatives. Similarly, improving the level of confidence in practicing healthy eating behaviors may also help accelerate the community acceptance of changes made to the food environments by these efforts. To achieve optimal interventional effects, priming at-risk groups to accept and take advantage of structural improvements may be as important as improving the food environments themselves [10, 35].

The present study was subject to a number of limitations. First, generalizability of the findings to the general population in Los Angeles County was not feasible, as the sampled group principally represented the region's low-income adult population. This, however, should not be considered a study weakness, as the demographics of the group aligned closely with the intended audiences of the various federal and local obesity prevention efforts in the region. Second, large confidence intervals were observed for some of the point estimates in the analyses. For most of these, the smaller sample size(s) of the referent group (i.e., whites) likely contributed to the imprecision. Third, self-selection and self-reporting bias likely led survey participants to over- or underestimate their food frequency and reports of self-confidence (self-efficacy in healthy eating). Fourth, the study design was cross-sectional in nature and thus was only able to describe the regional health profiles of the subpopulation at one given point in time. Finally, measurement errors, including misclassifications, likely introduced additional bias to the descriptive and comparison analyses. These potential errors, however, were minimized through iterative use of well-defined, standardized measurement protocols and rigorous training of field staff responsible for data collection.

Despite these limitations, community-level health data on particular subpopulations in Los Angeles County highlight the feasibility and utility of collecting these kinds of data to address socioecological factors that drive healthy eating in urban settings. To increase desired eating behaviors, both structural (system or environmental change) and individual-level approaches (e.g., changes in knowledge, attitudes, beliefs, intentions, and self-efficacy) should be employed. Developing nutrition interventions tailored to the unique characteristics of targeted subpopulations can help prepare individuals to take advantage of the structural improvements or resources that are made available to them by obesity prevention initiatives. Such tailoring of structural change interventions can be synergistically augmented by culturally sensitive public education and/or community engagement that seeks to address the underlying gender norms and self-confidence mediators known to shape eating behaviors.

Although local health data can help facilitate community planning and acceptance of system-level and environmental changes to the food environment, area-specific health profiles and behavioral determinants of healthy eating in the targeted subpopulations are often not well-characterized prior to nutrition program implementation or for use in subsequent program improvement efforts. The present study addresses this gap in public health practice by providing actionable data that the LACDPH can use to further address health disparities in the region. Many of the lessons learned in Los Angeles County may have similar applications in other US communities. Local or community health assessments represent a set of tools that is often underutilized by public health authorities. Ultimately, these chronic disease surveillance and assessment tools that document more granular information about subpopulations' health status provide the opportunity for communities to tailor multisector public health programs to intended audiences.