Abstract
Purpose: This study investigates how the density of neighborhood Food Outlets, specifically fast-food restaurants, influences obesity rates among older adults, considering their lifestyle, psychological factors, and demographics.
Design: A cross-sectional, multilevel study design was employed.
Setting: The research focused on census block groups within the Urban Growth Boundary of metropolitan Portland, Oregon, representing diverse neighborhoods.
Subjects: A sample of 1,221 residents (average age 65) was randomly selected from 120 neighborhoods, achieving a 48% response rate.
Measures: Fast-food restaurant density was measured using Geographic Information Systems (GIS). Participants’ Body Mass Index (BMI), frequency of fast-food visits, fried food intake, physical activity levels, self-efficacy in eating fruits and vegetables, income, and race/ethnicity were assessed.
Analyses: Multilevel logistic regression analyses were conducted to examine the relationships.
Results: Significant links were found between individual characteristics and obesity in high food outlet density neighborhoods compared to low-density areas. The odds ratios (OR) for obesity and 95% Confidence Intervals (CI) were: 1.878 (CI=1.006-3.496) for weekly fast-food visits; 1.792 (CI=1.006-3.190) for not meeting physical activity guidelines; 1.212 (CI=1.057-1.391) for low confidence in healthy eating; and 8.057 (CI=1.705-38.086) for non-Hispanic Black residents.
Conclusion: Higher density of neighborhood food outlets is associated with less healthy lifestyles, poorer psychosocial profiles, and a greater risk of obesity in older adults.
Keywords: Fast Food, Food Outlets, Environment Design, Community Health, Obesity, Public Health Policy, Prevention Research, Research, Modeling, Relationship Testing, Non-Experimental Study, Biometric Outcome Measure, Local Community Setting, Fitness, Physical Activity, Built Environment Strategy, Adults, Seniors, Geographic Location Target Population.
Introduction
Obesity is a growing public health crisis in the U.S., with prevalence rates increasing significantly among adults [1-5]. This rise positions overweight and obesity as major health concerns [6-9]. Adult obesity is linked to a higher risk of various health problems, including heart disease, type 2 diabetes, hypertension, certain cancers, and osteoarthritis [8, 10, 11]. Beyond physical ailments, obesity can also lead to psychological distress, such as depression, body image issues, and low self-esteem. It reduces life expectancy [12] and escalates healthcare and societal costs [13].
While the causes of obesity are multifaceted, encompassing lifestyle choices and genetics, mounting evidence points to the role of unhealthy environments, particularly the accessibility of certain types of food outlets, in contributing to overweight and obesity [14-18]. The increased availability of fast-food restaurants in local areas is a significant environmental factor under scrutiny. Data reveals a substantial surge in fast-food food outlets across the U.S. in recent decades [19, 20], especially in lower-income and predominantly Black urban areas [20, 21]. This trend aligns with increased spending on fast food [19, 22] and higher energy intake from food consumed away from home, notably from fast-food food outlets [19, 23-25]. Studies have linked fast food consumption to elevated BMI and weight gain [26-30]. For instance, a 15-year study indicated that adults eating fast food more than twice weekly gained significantly more weight and developed insulin resistance at a faster rate than those eating fast food less than once a week [30]. Furthermore, a correlation has been observed between the density of fast-food food outlets in neighborhoods and state obesity rates [16], and between residents in areas with more fast-food food outlets relative to full-service restaurants and higher weight status [31].
While existing research emphasizes the connection between obesogenic environments [32] and obesity, much of it focuses on the direct impacts of either neighborhood or individual factors independently. Given the growing understanding that weight gain and obesity are complex outcomes of interactions between biological, behavioral, and environmental elements [32], it is logical to consider that individual traits are linked to their surrounding social and neighborhood contexts. This suggests a potential interplay between neighborhood and individual characteristics in influencing obesity [33].
Therefore, this study aimed to determine if the associations between specific individual behavioral, psychosocial, and sociodemographic characteristics and obesity are influenced by the availability of local food outlets, specifically fast-food restaurants. Using a multilevel analysis, we tested the hypothesis that the strength of the relationship between individual measures—such as weekly fast-food visits, self-efficacy in fruit and vegetable consumption, fried food intake, physical activity levels—and obesity varies based on the density of fast-food food outlets in older adults’ neighborhoods. Building on prior research [20, 21, 34], household income and race/ethnicity were also included as sociodemographic factors in our analysis.
Methods
Design
A cross-sectional study with a multistage, stratified sampling design was used to obtain a representative sample of adults aged 50-75 from U.S. census block groups within the Portland, Oregon, metropolitan area’s Urban Growth Boundary (www.metro-region.org). Census block groups served as proxies for neighborhoods and the primary sampling unit (PSU). The sampling process involved three stages: (1) random selection of block groups representing diverse urban forms (land use mix), socioeconomic status (median income), and ethnic diversity; (2) random selection of households within these block groups; and (3) recruitment of eligible residents from selected households. The study was conducted between 2006 and 2007, with ethical approval from the Oregon Research Institute Institutional Review Board.
Sample
In 2006, 120 neighborhoods were chosen, stratified by urban form, median household income, and race/ethnicity. Households (single and multi-family) were identified using a commercial database (www.surveysampling.com) compiled from telephone directories, voter registration, and driver’s license data from 2006. This database provided age and contact information for recruitment. Sample sizes varied from 5-8 residents in smaller block groups to 9-21 in larger ones, proportionally allocated. Eligibility criteria included adults aged 50-75, English speaking, independently mobile (including cane use), and without major cognitive impairments.
Initial contact was made via mail, followed by phone calls. Non-responsive households were replaced through continued random sampling until target numbers were reached in each neighborhood. The final sample consisted of 1,221 participants (48% response rate). Data collection involved in-person interviews covering sociodemographics, diet, physical activity, body measurements, and perceptions of the neighborhood environment. All participants provided informed consent and received compensation for their participation.
Measures
Individual-Level Measures
Body Mass Index (BMI)
Weight and height were measured objectively and used to calculate BMI (kg/m²). For analysis, BMI was categorized: 1 = obese (BMI ≥ 30); 0 = non-obese (BMI < 30).
Eating-out behavior
Two questions assessed weekly fast-food visits: “How often do you eat at fast-food places like McDonald’s, Burger King, KFC, or Pizza Hut?” and “How often do you go to buffet-style restaurants?”. Responses ranged from (1) never to (6) daily. For analysis, a binary variable was created: 1 = eats out at fast food/buffets 1-2+ times weekly; 0 = less than once per week. The 12-month test-retest reliability was .75.
Eating self-efficacy
A 10-item scale adapted from Resincow et al. [35] measured confidence in eating more fruits and vegetables. Responses ranged from 1 (Not at all confident) to 10 (Completely confident). Scale validity and reliability were previously established [35], with a Cronbach’s alpha of .85. Scores were reversed for interpretation, with higher scores indicating lower confidence in healthy eating.
Fried food consumption
Participants were asked, “How many servings of fried food do you eat in a typical week?”. A binary score was created: 1 = 1+ servings; 0 = no servings.
Fruits and vegetables intake
The “All-Day” Fruit and Vegetable Screener [36] assessed intake frequency and portion sizes. Validity and reliability were established by developers [36]. Scores were calculated per National Cancer Institute guidelines [37], with higher scores indicating greater intake.
Physical activity
The Behavioral Risk Factor Surveillance System Survey (BRFSS) [38] was used to assess physical activity levels. Questions covered days per week and time per day spent in moderate and vigorous activity. Levels were categorized based on American College of Sports Medicine and CDC guidelines [39-41]: (a) meeting guidelines, (b) insufficiently active, (c) inactive. For analysis, categories (b) and (c) were combined into “not meeting recommended physical activity levels” (1 = not meeting; 0 = meeting).
Sociodemographic characteristics
Measures included: (a) age (continuous, 50-75 years); (b) gender (1 = male; 0 = female); (c) education (1 = high school or less; 0 = some college or more); (d) household income (1 = ≤$29,999; 0 = ≥$30,000); (e) race/ethnicity (1 = non-Hispanic Black; 0 = other); (f) employment (1 = employed; 0 = not employed); (g) home ownership (1 = yes; 0 = no); (h) alcohol use (1 = current; 0 = never/no use); (i) tobacco use (1 = current; 0 = never/no use); and (j) health status (5 = excellent to 1 = poor).
Neighborhood Measures
Fast-food restaurants
Commercial data from infoUSA (www.infousa.com), updated in 2006, provided information on fast-food food outlets in sampled block groups. Data, based on SIC codes, included fast-food chains like McDonald’s and Burger King. These data were geocoded and analyzed using GIS software (ArcView) [42]. Density was calculated as the number of fast-food food outlets per square mile per neighborhood and standardized for analysis. Neighborhoods were then classified into high and low density based on quartile ranges.
Land use mix
Data from Portland’s Regional Land Information System (RLIS; www.metro-region.org) were used to generate a land use mix index [43]. Values near 0 indicated single-use environments (e.g., residential suburbs), and values near 1 indicated mixed-use environments.
Residential density
Calculated as the number of persons per residential acre in each block group.
Other sociodemographics
2000 census data provided neighborhood-level socioeconomic measures: (a) median household income, (b) percentage of non-Hispanic Black residents, and (c) percentage of Hispanic residents. These, along with land use mix and residential density, were used as covariates.
Analysis
Multilevel random effects logistic regression models were used, with neighborhoods as level-2 units and residents as level-1 units, using Hierarchical Linear and Nonlinear Modeling software [44]. The models predicted obesity (BMI ≥30) based on neighborhood-level variables (fast-food food outlet density, land use mix, sociodemographics) and individual-level variables (eating-out behavior, self-efficacy, diet, physical activity, sociodemographics). A priori interaction terms were included: fast-food food outlet density by (a) fast-food visits, (b) physical activity, (c) self-efficacy, (d) fried food intake, (e) race/ethnicity, and (f) household income.
Results
The majority of participants were male (57%) and non-Hispanic White (92%), with an average age of 62 (±7 years), some college education or higher (77%), and a household income of $30,000 or more (73%). Overall, 38.2% (n=466) were obese (BMI ≥ 30), and approximately 36% did not meet physical activity recommendations. 72% reported fried food consumption (1-23 servings/week), and 24% visited fast-food food outlets 1-2+ times weekly. The mean self-efficacy score was low (2.57 ± 2).
Table 1 shows descriptive statistics of individual characteristics across low and high fast-food food outlet density neighborhoods. Higher density was associated with increased obesity among frequent fast-food visitors, those not meeting physical activity guidelines, and those with lower self-efficacy. Obesity prevalence also increased among non-Hispanic Black residents in high-density neighborhoods.
Table 1. Aggregated Descriptive Statistics on Obesity by Individual Characteristics Across High- and Low-Density of Neighborhood Fast-Food Outlets
| Density of Neighborhood Fast-Food Restaurants | |
|---|---|
| Low-Density | |
| Aggregated Individual-Level Measures at the Neighborhood level | Prevalence of Obesity (number of neighborhoods) |
| Visits to fast-food restaurants | |
| Yes (n = 113) | 36.08 (n = 58) |
| No (n = 7) | 23.75 (n = 2) |
| Meeting recommended physical activity | |
| Not meeting (n = 115) | 35.80 (n = 56) |
| Meeting (n = 5) | 33.75 (n = 4) |
| Low self-efficacy in eating fruits and vegetables | |
| Above or at median (n = 62) | 35.51 (n = 26) |
| Below median (n = 58) | 35.78 (n = 34) |
| Fried food consumption | |
| Yes (N = 120) | 35.67 (n = 60) |
| No | 0 |
| Race/ethnicity | |
| Non-Hispanic black (n = 29) | 33.34 (n = 11) |
| Others (n = 91) | 36.19 (n = 49) |
Of the initial six interaction terms, density of fast-food food outlets by fried food consumption and household income were not significant (p = .43, p = .21). The model was rerun without these. Table 2 presents odds ratios from the revised model. Significant cross-level interactions were found, adjusted for covariates. Residents in high-density fast-food food outlet neighborhoods visiting fast-food or buffet restaurants 1-2+ times weekly were 1.878 times (95% CI: 1.006, 3.496, p < .05) more likely to be obese than those in low-density areas. Similar results were found for those not meeting physical activity guidelines (OR=1.792, 95% CI: 1.006, 3.190, p < .05), those with low self-efficacy (OR =1.212, 95% CI: 1.057, 1.391, p < .005), and non-Hispanic Black residents (OR =8.057, 95% CI: 1.705, 38.086, p < .005).
Table 2. Multilevel Logistic Regression Analyses of Associations between Obesity and Resident-Level Measures of Characteristics and Neighborhood-Level Measures of Density of Fast-Food Outlets
| Obesity (BMI ≥ 30) | |||||
|---|---|---|---|---|---|
| Cross-level interaction effect | b coefficient | Standard Error | t-value | Odds Ratio | 95% CI |
| Density of fast-food outlets by visits to fast-food restaurants | 0.629 | 0.318 | 1.980 | 1.878 | 1.006, 3.496 |
| Density of fast-food outlets by not meeting recommended physical activity | 0.583 | 0.294 | 1.982 | 1.792 | 1.006, 3.190 |
| Density of fast-food outlets by self-efficacy of eating fruits and vegetables | 0.193 | 0.070 | 2.752 | 1.212 | 1.057, 1.391 |
| Density of fast-food outlets by race/ethnicity (non-Hispanic Black) | 2.087 | 0.792 | 2.633 | 8.057 | 1.705, 38.086 |
*Odds ratios adjusted for resident-level variables: age, gender, education, income, employment, home ownership, alcohol and tobacco use, vegetable intake, fried food consumption; and neighborhood-level variables: land use mix, residential density, median income, percentage of non-Hispanic Black and Hispanic residents.
Discussion
This study explored the interaction between individual factors (eating habits, physical activity, self-efficacy, ethnicity) and the built environment, specifically the density of local fast-food food outlets, in relation to obesity in older adults. Findings support the hypothesis that in neighborhoods with more fast-food food outlets, the link between these individual factors and obesity is stronger. Specifically, older adults who frequently visit fast-food food outlets, have lower self-efficacy in healthy eating, do not meet physical activity guidelines, and are non-Hispanic Black are more likely to be obese in high food outlet density neighborhoods.
These results suggest that while individual lifestyles and psychological factors are significant in obesity risk, the environmental context, such as the availability of fast-food food outlets, can amplify this risk. The multilevel design of this study allows for a simultaneous examination of neighborhood and individual factors, addressing a methodological gap in prior research [45].
Consistent with ecological models [33, 46, 47] and prior empirical findings [16, 26, 28, 30, 31], this study extends research by specifically examining cross-level interactions between individual behaviors and the neighborhood food outlet environment in contributing to obesity. The study also reinforces findings on race/ethnicity and obesity [20, 21, 34], showing that African Americans in high fast-food food outlet density neighborhoods are at higher obesity risk.
Limitations
The cross-sectional design limits causal conclusions. We cannot definitively say that food outlet distribution causes fast-food visits and weight gain. Future longitudinal studies should examine temporal relationships between environmental factors, individual behaviors, and obesity over time. Observing changes in built environment characteristics like food outlet density requires longer timeframes than changes in individual lifestyles, posing a challenge for interaction effect studies.
The focus on fast-food food outlets is another limitation. The study did not consider workplace environments or factors like lack of sidewalks, which might influence sedentary behavior and reliance on food outlets. Neighborhood characteristics like car-dependency could also affect eating habits. Future research should incorporate these broader built environment factors.
Finally, self-reported fast-food visits are a limitation. The exact location of visits was not recorded, so we could not confirm if they were within the study area. Future studies should use GIS to track visited food outlet locations relative to residence and neighborhood density for greater accuracy.
Significance and Implications
This study highlights the public health significance of obesogenic environments, such as high fast-food food outlet density, in exacerbating unhealthy lifestyles and obesity. It supports the need to consider built environment influences in addressing the obesity epidemic. From a public health perspective, these findings suggest the need for policies that regulate the growth of fast-food food outlets in neighborhoods and promote healthier food environments [31]. Without addressing environmental factors, the impact of individual-focused health education programs may be limited.
Conclusion
This research contributes to the limited body of multilevel studies on neighborhood built environment factors and obesity in older adults. It emphasizes the importance of considering fast-food food outlet availability as a moderator in the relationship between individual lifestyles, self-efficacy, demographics, and obesity. High concentrations of fast-food food outlets, combined with unhealthy lifestyles and poor psychosocial profiles, increase obesity risk. Public health strategies to reduce obesity should include policy changes to regulate fast-food food outlet proliferation in communities.
So What?
This study demonstrates how neighborhood fast-food food outlet density influences the relationship between individual characteristics and obesity in older adults. It underscores the importance of obesogenic environments and their impact on health behaviors. Understanding the interaction between food outlet environments, lifestyle, and demographics is crucial for developing effective land use and public health policies to mitigate the negative effects of detrimental food environments on public health.
Contributor Information
Fuzhong Li, Oregon Research Institute, Eugene, Oregon
Peter Harmer, Department of Exercise Science, Willamette University.
Bradley J. Cardinal, Department of Exercise Science, Oregon State University
Mark Bosworth, Metro Regional Services, Portland, Oregon 97232
Deb Johnson-Shelton, Oregon Research Institute, Eugene, Oregon
