Neighborhood Social Cohesion and Obesity in the United States

Abstract

Low neighborhood social cohesion (nSC) has been associated with obesity. Still, few studies have assessed the nSC-obesity relationship among a large, nationally representative, and racially/ethnically diverse sample of the United States population. To address this literature gap, we examined cross-sectional associations among 154,480 adult participants of the National Health Interview Survey (NHIS) from 2013–2018. We also determined if associations varied by race/ethnicity, sex/gender, age, annual household income, and food security status. Based on a 4-item scale from the Project on Human Development in Chicago Neighborhoods Community Survey, we categorized nSC as low, medium, and high. Based on body mass index (BMI) recommendations, we categorized obesity as ≥30 kg/m². We used Poisson regression with robust variance to directly estimate prevalence ratios (PRs) and 95% confidence intervals (CIs) while adjusting for sociodemographic characteristics, such as annual household income, educational attainment, and marital status, along with other confounders. Study participants’ mean age ± standard error was 47.1±0.1 years; most (69.2%) self-identified as Non-Hispanic (NH)-White, and 51.0% were women. NH-Black and Hispanic/Latinx adults comprised more of the population in neighborhoods with low nSC (14.0% NH-Black, 19.1% Hispanic/Latinx, and 61.8% NH-White) versus high nSC (7.7% NH-Black, 10.4% Hispanic/Latinx and 77.0% NH-White). Low vs. high nSC was associated with a 15% higher prevalence of obesity (PR=1.15 [95% CI: 1.12–1.18]), and the magnitude of the association was more substantial among NH-White (PR=1.21 [95% CI: 1.17–1.25]) compared to associations among Hispanic/Latinx (PR=1.04 [95% CI: 0.97–1.11]) and NH-Black (PR=1.01 [95% CI: 0.95–1.07]) adults. Low vs. high nSC was associated with a 20% higher prevalence of obesity in women (PR=1.20 [95% CI: 1.16–1.24]) compared to a 10% higher prevalence in men (PR=1.10 [95% CI: 1.06–1.14]). Low vs. high nSC was associated with a 19% higher prevalence of obesity among adults ≥50 years old (PR=1.19 [95% CI: 1.15–1.23]) compared to a 7% higher prevalence of obesity among adults <50 years old (PR=1.07 [95% CI: 1.03–1.11]). Efforts to address nSC may improve health and address health disparities.

INTRODUCTION

Obesity prevalence is alarmingly high in the United States (US). In 2018, 42.4% of adults in the US had obesity, and this estimate varied by race/ethnicity. For instance, the prevalence of obesity was estimated to be higher among Non-Hispanic (NH)-Black (49.6%), Hispanic/Latinx (44.8%), and NH-White (42.2%) compared to NH-Asian (17.4%) adults¹. However, these estimates did not consider the lower body mass index (BMI) thresholds for NH-Asians² or the likelihood that BMI thresholds by demographics (i.e. race/ethnicity; sex/gender)³. The prevalence of obesity further varies by sex/gender within racial/ethnic groups, where NH-Black women (56.9%) had the highest prevalence compared to both NH-Black men and men and women of other racial/ethnic groups¹. Obesity is also more common among older compared to younger adults and individuals with lower versus higher annual household incomes^4,5. Obesity contributes to myriad poor health outcomes, such as type 2 diabetes mellitus, cardiovascular diseases, and several types of cancer. This high health burden consequently poses a sizeable economic burden^6,7 with an estimated annual medical cost of $260.6 billion⁸. Previous efforts to address obesity in the US have included individual-level lifestyle interventions directed towards weight loss, which have largely been unsuccessful in the long term and generally have not considered the influence of upstream, societal factors such as social disadvantage, material deprivation, and other forms of exclusion⁹ that drive the production of disparities^6,10. Metabolism is threatened by the physical and social environment where one lives, works, and plays, as individual-level behaviors (e.g., eating, physical activity, sleep) do not occur in a vacuum and are instead influenced by socioenvironmental factors (e.g., neighborhood characteristics). Therefore, it is essential to understand how macro-level interventions focused on addressing, for example, poverty, food insecurity, and health care inequality, may address the root causes of obesity¹¹. Since the neighborhood environment can serve as a convenient, cost-efficient, and practical approach to reducing the burden of obesity, it is crucial to determine the potential role of attributes of the neighborhood environment¹². A promising feature of the understudied social neighborhood environment that can offer targets for intervention is neighborhood social cohesion (nSC).

nSC, defined as the degree of connectedness and solidarity among people in a community¹³, may provide more resources and support, healthy norms and values, and increased perceptions of a safe environment that may be useful for intervening in obesity⁴. Socially cohesive neighborhoods may protect against obesity by providing opportunities to socialize and offer social support, including mental (e.g., stimulation), emotional (e.g., well-being), and physical (e.g., walking) support¹⁴. Likewise, social support may also promote healthy norms among neighbors (e.g., nutritious food; cooking).⁴ Conversely, non-socially cohesive neighborhoods may contribute to obesity through, for instance, increased chronic stress that stimulates opioid release in the reward center of the brain¹⁵. This defense mechanism, which helps the body attenuate the detrimental effects of stress, leads to increased energy intake and fat accumulation over time¹⁵. Because women rely on social support for health and overall well-being more so than men, nSC may differentially impact women in a positive manner¹⁶. nSC may vary by race/ethnicity for various reasons, including governmental policies investing in predominantly NH-White neighborhoods in ways that may enhance social cohesion, such as parks/green space that encourage positive social interactions¹⁷ while disinvesting from predominantly NH-Black and Hispanic/Latinx neighborhoods¹⁸ where fewer parks/green space may be related to less social support¹⁹. A reliance on nSC may also be more critical to minoritized racial/ethnic groups since nSC may serve as an alternative avenue of support and resources in the context of historical and current exclusionary policies and practices²⁰. Likewise, people with low income and food insecurity may differentially experience and perceive nSC compared to high-income and food-secure individuals since high levels of cohesion may facilitate access to food supplies through, for example, neighbors or local food programs^21,22.

Furthermore, research shows that health is simultaneously shaped by multiple social factors, where compounding effects of racism and sexism and distinct forms of marginalization result in more detrimental impacts on health among NH-Black women, for example²³. Therefore, it is essential to incorporate an intersectionality framework to investigate joint health impacts of social inequalities across multiple social categories, such as race/ethnicity, sex/gender, and class (e.g., annual household income)²³. Given limited mobility and financial constraints, older adults, the most sedentary age group, heavily rely on their immediate surroundings and thus may benefit from high nSC²⁴. Although this assertion has been supported by previous studies²⁵, it has rarely been examined nationally among a racially/ethnically diverse sample. While a study by Yu and colleagues used nationally representative data from the National Health Interview Survey (NHIS), the study was limited to older adults. They examined racial/ethnic groups by income level only (i.e., older NH-White, NH-Black, Hispanic/Latinx, and other race older adults with low and high income)²⁵. The study was also limited to one year (i.e., 2013). Our study expands upon prior literature by using multiple survey years from the nationally representative NHIS to assess the relationship between nSC and obesity among NH-White, NH-Black, Hispanic/Latinx, and NH-Asian adults aged 18 years, and older by sex/gender, age, annual household income level, and food security status.

To address the aforementioned gaps in the literature, we used data from the NHIS to estimate cross-sectional associations between nSC and obesity. We also determined if associations varied by race/ethnicity, sex/gender, age, annual household income, and food security status. We hypothesized that the perception of living in a neighborhood with low and medium vs. high social cohesion would be associated with higher obesity among all participants and that the magnitude of the association would be more substantial for NH-Black, Hispanic/Latinx, NH-White participants vs. NH-Asian participants. We also hypothesized that the extent of the association would be more substantial among women vs. men, older vs. younger adults, individuals with low vs. high income, and individuals who are food insecure vs. secure among all racial/ethnic groups living in low vs. high nSC.

METHODS

Data Source: National Health Interview Survey

We obtained participant data from the NHIS using the Integrated Health Interview Series²⁶ for survey years 2013 to 2018. A detailed description of the NHIS procedures can be found elsewhere²⁷. The NHIS is an annual series of cross-sectional, household surveys conducted via computer-assisted, in-person interviews among the non-institutionalized US adult population (≥18 years). Using a three-stage stratified cluster probability sampling design, the NHIS obtains a nationally representative sample. A randomly selected adult from each household provides more specific health-related information to trained interviewers. The response rate for sample adults was 56.1% (range: 61.2% (2013) to 53.1% (2018)). We used sampling weights to account for the survey’s complex sampling design, non-response, and oversampling of certain groups (e.g., racial/ethnic minorities) that resulted in unequal probabilities of selection. The NHIS collected informed consent from each study participant. The National Institute of Environmental Health Sciences’ (NIEHS) Institutional Review Board (IRB) waived approval for publicly available secondary data with no identifiable information.

Study Population

Participants from all 50 states and the District of Columbia were included in the sample. Of the 190,113 NHIS participants, those with missing or implausible data on critical variables, including race/ethnicity (n=3,612), nSC (n=11,159), BMI (n=5,046), annual household income (n=14,540), and food security status (n=20) were excluded. We also excluded the Native American race (n=1,256) due to the small sample size resulting in a final analytical sample size of 154,480 participants (Supplemental Figure 1). Sociodemographic, health behaviors, and clinical characteristics were compared among those included to those excluded (Supplemental Table 1).

Exposure Assessment: Neighborhood Social Cohesion

nSC was measured using a modified four-item scale developed by the Project on Human Development in Chicago Neighborhoods Community Survey²⁸. Participants responded on a Likert scale (1=definitely agree; 2=somewhat agree; 3=somewhat disagree; and 4=definitely disagree) to the following four statements: 1) People in this neighborhood help each other out; 2) There are people I can count on in this neighborhood; 3) People in this neighborhood can be trusted; and 4) This is a close-knit neighborhood. Cronbach’s alpha (0.884) suggested high internal consistency. Responses were reverse coded, summed, and categorized as low (<11), medium (12–14), and high (≥15), as done in previous studies^29,30.

Outcome Assessment: Obesity

BMI was calculated by dividing self-reported weight in kilograms by self-reported height in meters squared and dichotomized as those without obesity (18.5-<29.9 kg/m²) vs. those with obesity (≥30 kg/m²). In a supplemental analysis, we considered the following categories: recommended/normal (18.5-<25 kg/m²), overweight (25–29.9 kg/m²) versus obesity (≥30 kg/m²). Considering the small sample of underweight (1.6%), we excluded them from the analysis.

Potential Confounders

Potential sociodemographic confounders considered were age (18–30, 31–49, and ≥50 years), sex/gender (women or men), race/ethnicity (NH-White, NH-Black, Hispanic/Latinx, and NH-Asian), annual household income (<$35,000, $35,000-$74,999, and ≥$75,000), educational attainment (<high school, high school graduate, some college, and ≥college), occupational class (professional/management, support services, or laborers), food security status (yes or no), region of residence (Northeast, Midwest, South, and West), and marital status (married/living with partner/cohabitating, divorced/widowed/separated, or single/no live-in partner). Behaviors, such as smoking, alcohol consumption, and physical activity, were not considered because they are likely mediators between nSC and obesity⁴. Likewise, health status and clinical characteristics (e.g., cardiovascular disease) were not considered as they are the most likely outcomes of obesity. Nonetheless, we observe the distribution of those variables by levels of nSC.

Potential Modifiers: Race/Ethnicity, Sex/Gender, Age, Annual Household Income, and Food Security Status

Participants self-identified their race/ethnicity, sex/gender, age, and annual household income. Race/ethnicity was categorized as NH-White alone, NH-Black alone, Hispanic/Latinx (of any race), and NH-Asian alone. Sex/gender was assessed in a binary manner and dichotomized as women or men. Age was categorized as <50 and ≥50 years. Annual household income was dichotomized between <$75,000 vs. ≥$75,000. Food security status was determined by responses to 10 food security questions, such as “In the last 30 days, did you ever not eat for a whole day because there wasn’t enough money for food?” Responses were high food security, marginal food security, low food security, and very low food security and dichotomized between and dichotomized between yes (high food and marginal food security) vs. no (low food and very low food security).

Statistical Analyses

We computed descriptive statistics and presented continuous variables as means ± standard error (SE) and categorical variables as weighted percentages after applying direct standardization using the 2010 US Census population as the referent population. We compared the three levels of nSC across sociodemographic, health behavior, and clinical characteristics for all participants.

To test associations between nSC and obesity, we used Poisson regression with robust variance to estimate prevalence ratios (PRs)³¹ and 95% confidence intervals (CIs) of obesity for low and medium vs. high nSC overall, and by race/ethnicity, sex/gender, age, income, and food security status. We adjusted for the following confounders in the overall model: age, sex/gender, race/ethnicity, annual household income, educational attainment, occupational class, food security status, region of residence, and marital status. To test for differences by race/ethnicity across sex/gender, age, annual household income, and food security status, we added respective interaction terms (e.g., nSC*race/ethnicity; nSC*race/ethnicity*age) to the overall model. We used SAS version 9.4 for Windows (Cary, North Carolina) to conduct analyses and determined statistical significance using a two-sided p-value of 0.05.

RESULTS

Study Population Characteristics

Among 154,480 participants, 32%, 33%, and 35% perceived their neighborhoods to have low, medium, and high levels of social cohesion, respectively (Table 1). The mean age was 47.1±0.1 years, and the mean age among those living in low nSC (43.7±0.1) was lower than among those living in high nSC (50.4±0.1). Approximately 51.0% were women, 69.2% self-identified as NH-White, 41.8% had an annual household income ≥$75,000, and 91.0% were food secure. NH-White adults made up 61.8% of those living in low nSC compared to 77.0% of those living in high nSC, while NH-Black and Hispanic/Latinx adults living in low nSC made up 14.0% and 19.1% compared to 7.7% and 10.4%, respectively, of those living in high nSC (Table 1).

Table 1.

Age-standardized Sociodemographic, Health Behavior, and Clinical Characteristics Overall and by Neighborhood Social Cohesion Level, National Health Interview Survey, 2013–2018 (N=154,480)^a

	Neighborhood Social Cohesion
	Low n=49,780 (32%)	Medium n=50,946 (33%)	High n=53,754 (35%)	Overall N=154,480 (100%)
Sociodemographic
Age, mean (S.E.), years	43.7±0.13	47.0±0.13	50.4±0.14	47.1±0.10
18–30	18.7%	16.0%	13.5%	16.2%
31–50	20.1%	23.7%	26.2%	23.4%
≥50	60.3%	60.3%	60.3%	60.3%
Sex/gender
Women	52.3%	48.4%	52.0%	51.0%
Race/ethnicity
NH-White	61.8%	68.2%	77.0%	69.2%
NH-Black	14.0%	11.4%	7.7%	11.0%
Hispanic/Latinx	19.1%	14.2%	10.4%	14.5%
NH-Asian	5.1%	6.2%	4.9%	5.4%
Annual household income
<$35,000	36.7%	25.8%	22.0%	28.0%
$35-$74,999	32.3%	30.2%	28.1%	30.3%
≥$75,000	31.0%	44.0%	50.0%	41.8%
Educational attainment
<High school	13.5%	9.5%	8.2%	10.3%
High school graduate	29.7%	26.0%	25.6%	27.1%
Some college	31.0%	29.7%	29.6%	30.1%
≥College	25.8%	34.8%	36.5%	32.5%
Occupation class
Professional/management	17.6%	23.2%	24.0%	21.8%
Support services	43.6%	43.9%	46.0%	44.6%
Laborers	38.7%	32.9%	30.0%	33.6%
Marital status
Married/living with partner/co-habited	55.5%	62.6%	67.2%	62.0%
Divorced/widowed	24.0%	19.9%	18.2%	20.4%
Single/no live-in partner	20.5%	17.6%	14.5%	17.5%
Food Security Statusb
Food secure	85.7%	92.1%	94.6%	91.0%
Region of residence
Northeast	17.5%	18.6%	17.8%	17.9%
Midwest	21.2%	22.6%	23.8%	22.6%
South	36.7%	35.6%	37.7%	36.7%
West	24.6%	23.1%	20.6%	22.7%
Health Behaviors	Low	Medium	High	Overall
Sleep duration
<6 hours (very short)	11.9%	8.0%	7.5%	9.0%
<7 hours (short)	36.2%	30.5%	28.1%	31.3%
7–9 hours (recommended)	59.3%	65.9%	68.2%	64.7%
>9 hours (long)	4.5%	3.6%	3.8%	4.0%
Smoking status
Never/quit >12 months prior	80.0%	84.5%	85.6%	83.5%
Former	1.5%	1.4%	1.2%	1.4%
Current	18.5%	14.1%	13.2%	15.1%
Alcohol consumption
Never	19.9%	18.0%	18.8%	18.9%
Former	17.4%	14.5%	14.0%	15.1%
Current	62.6%	67.5%	67.3%	66.0%
Leisure-time physical activity (PA)
Never/unable	37.5%	29.9%	28.0%	31.5%
Does not meet PA guidelines	19.6%	19.1%	18.3%	19.0%
Meets PA guidelinesc	42.9%	51.0%	53.7%	50.0%
Clinical Characteristics	Low	Medium	High	Overall
Body Mass Index (BMI)
Underweight (<18.5 kg/m2)	1.6%	1.6%	1.6%	1.6%
Recommended/Normal (18.5-<25 kg/m2)	30.1%	32.3%	33.9%	32.2%
Overweight (25–29.9 kg/m2)	34.2%	36.0%	35.9%	35.4%
Obesity (≥30 kg/m2)	34.1%	30.1%	28.5%	30.7%
Health status
Excellent/very good	49.8%	59.6%	65.6%	58.8%
Good	30.1%	27.4%	23.6%	27.0%
Fair/poor	20.0%	13.0%	10.8%	14.2%
Severe psychological distressd	5.8%	2.7%	2.3%	3.5%
Dyslipidemiae	50.7%	48.8%	50.2%	49.8%
Hypertensionf	38.7%	35.6%	34.0%	35.9%
Prediabetes/diabetesg	20.9%	17.5%	15.6%	17.7%
Cancerh	10.7%	10.9%	11.4%	11.1%

^aNote all estimates are weighted for the survey’s complex sampling design. All estimates are age-standardized to the US 2010 population, except for age. Percentage may not sum to 100 due to missing values or rounding. SE= standard error.

^bFood secure is defined as high and marginal food security whereas food insecure is defined as low and very low food security.

^cMeets PA guidelines defined as ≥150 minutes/week of moderate intensity or ≥75 minutes/week of vigorous intensity or ≥150 minutes/week of moderate and vigorous intensity.

^dKessler 6-psychological distress scale score ≥13.

^eDyslipidemia defined as high cholesterol in the 12 months prior to interview.

^fHypertension defined as ever told by a doctor had hypertension.

^gPrediabetes/diabetes defined as ever told by a doctor had diabetes or prediabetic condition.

^hCancer defined as ever told had cancer.

About one-third (30.7%) of participants were classified as having obesity. Overall, the prevalence of obesity was higher among those who reported living in a neighborhood with low social cohesion (34.1%) compared to medium (30.1%) and high (28.5%) (Table 1). The prevalence of very short (<6 hours) and short sleep (<7 hours) duration was higher among those living in low compared to high nSC (11.9% vs. 7.5% and 36.2% vs. 28.1%, respectively). In contrast, the prevalence of recommended sleep was higher among those living in high compared to low nSC (68.2% vs. 59.3%). Similarly, the prevalence of current smoking was higher in individuals reporting low compared to high nSC (18.5% vs. 13.2%). Additional sociodemographic, health behavior, and clinical characteristics by age, sex/gender, race/ethnicity, income, and food security status are described in Supplemental Tables 2–6. Sociodemographic characteristics by the nSC scale are described in Supplemental Table 7.

Neighborhood Social Cohesion and Obesity Overall, by Race/Ethnicity, and by Sex/Gender

Overall, participants who lived in a neighborhood with low vs. high social cohesion had a 15% (PR=1.15 [95% CI: 1.12–1.18]) higher prevalence of obesity after adjustment (Table 2). Participants who lived in a neighborhood with medium vs. high social cohesion had a 6% (PR=1.06 [95% CI: 1.03–1.08]) higher prevalence of obesity after adjustment. Low vs. high nSC was associated with obesity among NH-White (PR=1.21 [95% CI: 1.17–1.25]), NH-Asian (PR=1.20 [95% CI: 0.96–1.51]), Hispanic/Latinx (PR=1.04 [95% CI: 0.97–1.11]), and NH-Black (PR=1.01 [95% CI: 0.95–1.07]) adults, after adjustment. Medium vs. high nSC was significantly associated with obesity among NH-White adults. Women who lived in a neighborhood with low vs. high social cohesion had a higher prevalence of obesity (PR=1.20 [95% CI: 1.16–1.24]), and this association was more robust than the association among men (PR=1.10 [95% CI: 1.06–1.14]), after adjustment. Likewise, NH-White women who lived in a neighborhood with low vs. high social cohesion had a higher prevalence of obesity (PR=1.26 [95% CI: 1.21–1.31]) with a stronger association observed than the association among NH-White men (PR=1.16 [95% CI: 1.11–1.21]). Hispanic/Latinx women who lived in a neighborhood with low vs. high social cohesion also had a high prevalence of obesity (PR=1.13 [95% CI: 1.03–1.24]) (Table 2). Supplemental analyses considering recommended/normal (18.5-<25 kg/m2), overweight (25–29.9 kg/m2) vs. obesity (≥30 kg/m2) showed similar results (Supplemental Table 8).

Table 2.

Fully Adjusted Prevalence Ratios of Obesity by Low and Medium Compared to High Neighborhood Social Cohesion Among the Overall Study population and across White, Black, Hispanic/Latinx, and Asian Participants Stratified by Sex/Gender, National Health Interview Survey, 2013–2018 (N=154,480)

Neighborhood Social Cohesion	Overall N=154,480	NH-White n=103,920 (67%)	NH-Black n=19,074 (12%)	Hispanic/Latinx n=23,207 (15%)	NH-Asian n=8,279 (5%)
All, N=154,480
Low vs. High	1.15 (1.12 – 1.18)	1.21 (1.17 – 1.25)	1.01 (0.95 – 1.07)	1.04 (0.97 – 1.11)	1.20 (0.96 – 1.51)
Medium Vs. High	1.06 (1.03 – 1.08)	1.08 (1.05 – 1.11)	0.96 (0.91 – 1.03)	0.97 (0.90 – 1.05)	1.07 (0.88 – 1.31)
Women, n=83,369 (54%)
Low vs. High	1.20 (1.16 – 1.24)	1.26 (1.21 – 1.31)	1.01 (0.94 – 1.08)	1.13 (1.03 – 1.24)	1.07 (0.79 – 1.46)
Medium Vs. High	1.09 (1.05 – 1.13)	1.12 (1.08 – 1.17)	0.96 (0.89 – 1.03)	1.02 (0.92 – 1.12)	1.09 (0.82 – 1.46)
Men, n=71,111 (46%)
Low vs. High	1.10 (1.06 – 1.14)	1.16 (1.11 – 1.21)	1.01 (0.90 – 1.12)	0.94 (0.85 – 1.04)	1.32 (0.97 – 1.81)
Medium Vs. High	1.02 (0.99 – 1.06)	1.04 (1.00 – 1.09)	0.98 (0.88 – 1.09)	0.94 (0.84 – 1.04)	1.06 (0.78 – 1.45)

Obesity defined as ≥30 kg/m² vs. without obesity.

All model adjusted for age (18–30, 31–49, 50+ years), sex/gender (women or men), race/ethnicity (NH-White, NH-Black, Hispanic/Latinx, and NH-Asian), annual household income (<$35,000, $35,000-$74,999, $75,000+), educational attainment (<high school, high school graduate, some college, ≥college), occupational class (professional/management, support services, laborers), food security status (food secure or insecure), region of residence (Northeast, Midwest, South, West), and marital/co-habiting status (married/living with partner or cohabitating, divorced/widowed/separated, single/no live-in partner).

Note. All estimates are weighted for the survey’s complex sampling design. Boldface indicates statistically significant results at the 0.05 level.

Neighborhood Social Cohesion and Obesity by Race/Ethnicity-Sex/Gender-Age

Participants ≥50 years old who lived in a neighborhood with low vs. high social cohesion had a higher prevalence of obesity (PR=1.19 [95% CI: 1.15–1.23]), and the association was more substantial than associations among participants aged <50 years old (PR=1.07 [95% CI: 1.03–1.11]), after adjustment (Table 3). Similar patterns emerged by age among men only. Among NH-White adults, low vs. high nSC was associated with a higher prevalence of obesity in those ≥50 (PR=1.21 [95% CI: 1.17–1.26]) than in those <50 years old (PR=1.15 [95% CI: 1.09–1.20]). Among Hispanic/Latinx adults, low vs. high nSC was associated with obesity among participants ≥50 years old (PR=1.15 [95% CI: 1.03–1.29]) but not among those <50 years old (PR=0.96 [95% CI: 0.88–1.04]). Low vs. high nSC was also associated with obesity among Hispanic/Latinx men ≥50 years old (PR=1.20 [95% CI: 1.01–1.42]) and NH-White men ≥50 years old (PR=1.20 [95% CI: 1.13–1.27]) but not among Hispanic/Latinx men <50 years old (PR=0.83 [95% CI: 0.74–0.94]) and NH-White men <50 years old (PR=1.07 [95% CI: 1.00–1.15]) (Table 3).

Table 3.

Fully Adjusted Prevalence Ratios of Obesity by Low and Medium Compared to High Neighborhood Social Cohesion Among the Overall Study Population and across White, Black, Hispanic/Latinx, and Asian Participants Stratified by Sex/Gender and Age, National Health Interview Survey, 2013–2018 (N=154,480)

Neighborhood Social Cohesion	Overall N=154,480		NH-White n=103,920 (67%)		NH-Black n=19,074 (12%)		Hispanic/Latinx n=23,207 (15%)		NH-Asian n=8,279 (5%)
	<50 years n=76,066 (49%)	≥50 years n=78,414 (51%)	<50 years n=45,943 (44%)	≥50 years n=57,977 (56%)	<50 years n=9,664 (50%)	≥50 years n=9,410 (50%)	<50 years n=15,429 (66%)	≥50 years n=7,778 (34%)	<50 years n=5,030 (61%)	≥50 years n=3,249 (39%)
All, N=154,480
Low vs. High	1.07 (1.03 – 1.11)	1.19 (1.15 – 1.23)	1.15 (1.09 – 1.20)	1.21 (1.17 – 1.26)	0.95 (0.87 – 1.03)	1.05 (0.97 – 1.14)	0.96 (0.88 – 1.04)	1.15 (1.03 – 1.29)	1.14 (0.84 – 1.55)	1.26 (0.90 – 1.76)
Medium Vs. High	1.00 (0.96 – 1.03)	1.09 (1.06 – 1.12)	1.03 (0.98 – 1.08)	1.10 (1.06 – 1.14)	0.93 (0.84 – 1.02)	1.00 (0.93 – 1.08)	0.89 (0.81 – 0.98)	1.12 (1.00 – 1.25)	1.09 (0.84 – 1.42)	1.04 (0.75 – 1.44)
Women, n=83,369 (54%)
Low vs. High	1.17 (1.11 – 1.23)	1.18 (1.13 – 1.23)	1.23 (1.15 – 1.32)	1.23 (1.16 – 1.30)	0.98 (0.89 – 1.09)	1.02 (0.93 – 1.12)	1.13 (1.00 – 1.27)	1.11 (0.96 – 1.27)	1.11 (0.70 – 1.74)	1.05 (0.68 – 1.61)
Medium Vs. High	1.07 (1.02 – 1.13)	1.09 (1.04 – 1.13)	1.11 (1.03 – 1.19)	1.12 (1.06 – 1.18)	0.97 (0.86 – 1.09)	0.95 (0.87 – 1.04)	1.00 (0.88 – 1.14)	1.03 (0.88 – 1.20)	1.16 (0.77 – 1.86)	1.05 (0.69 – 1.58)
Men, n=71,111 (46%)
Low vs. High	0.99 (0.93 – 1.04)	1.19 (1.13 – 1.25)	1.07 (1.00 – 1.15)	1.20 (1.13 – 1.27)	0.91 (0.78 – 1.05)	1.11 (0.95 – 1.31)	0.83 (0.74 – 0.94)	1.20 (1.01 – 1.42)	1.15 (0.78 – 1.70)	1.71 (0.99 – 2.95)
Medium Vs. High	0.93 (0.89 – 0.99)	1.09 (1.05 – 1.15)	0.97 (0.91 – 1.04)	1.09 (1.03 – 1.14)	0.89 (0.76 – 1.05)	1.08 (0.93 – 1.26)	0.82 (0.72 – 0.94)	1.19 (1.02 – 1.39)	1.03 (0.72 – 1.48)	1.06 (0.62 – 1.81)

Obesity defined as ≥30 kg/m² vs. without obesity.

All model adjusted for sex/gender (women or men), race/ethnicity (NH-White, NH-Black, Hispanic/Latinx, and NH-Asian), annual household income (<$35,000, $35,000-$74,999, $75,000+), educational attainment (<high school, high school graduate, some college, ≥college), occupational class (professional/management, support services, laborers), food security status (food secure or insecure), region of residence (Northeast, Midwest, South, West), and marital/co-habiting status (married/living with partner or cohabitating, divorced/widowed/separated, single/no live-in partner).

Note. All estimates are weighted for the survey’s complex sampling design. Boldface indicates statistically significant results at the 0.05 level.

Neighborhood Social Cohesion and Obesity by Race/Ethnicity-Sex/Gender-Annual Household Income

Participants with annual household incomes ≥$75,000 who lived in a neighborhood with low vs.high social cohesion had a similar prevalence of obesity (PR=1.12 [95% CI: 1.07–1.18]) to those with annual household incomes <$75,000 (PR=1.11 [95% CI: 1.08–1.15]), after adjustment (Table 4). Among those with annual household incomes of ≥$75,000, positive associations between low vs. high nSC and higher prevalence of obesity were stronger among women (PR=1.27 [95% CI: 1.18–1.36]) compared to men (PR=1.02 [95% CI: 0.96–1.10]), after adjustment. In contrast, the prevalence of obesity was similar among men and women who earned annual incomes <$75,000.

Table 4.

Fully Adjusted Prevalence Ratios of Obesity by Low and Medium Compared to High Neighborhood Social Cohesion Among the Overall Population and across White, Black, Hispanic/Latinx, and Asian Participants Stratified by Sex/Gender and Income, National Health Interview Survey, 2013–2018 (N=154,480)

Neighborhood Social Cohesion	Overall N=154,480		NH-White n=103,920, 67%		NH-Black n=19,074, 12%		Hispanic/Latinx n=23,207, 15%		NH-Asian n=8,279, 5%
	<$75k n=104,911(67%)	≥$75k n=49,569(33%)	<$75k n=65,352(62%)	≥$75k n=38,568(38%)	<$75k n=15,928(83%)	≥$75k n=3,146)(17%)	<$75k n=18,818(81%)	≥$75k n=4,389(19%)	<$75k n=4,813(57%)	≥$75k n=3,466(43%)
All, N=154,480
Low vs. High	1.11 (1.08 – 1.15)	1.12 (1.07 – 1.18)	1.17 (1.13 – 1.21)	1.17 (1.11 – 1.23)	0.98 (0.91 – 1.05)	1.04 (0.91 – 1.20)	1.05 (0.98 – 1.12)	0.92 (0.78 – 1.09)	1.42 (1.03 – 1.96)	0.95 (0.68 – 1.34)
Medium Vs. High	1.03 (1.00 – 1.06)	1.06 (1.01 – 1.18)	1.05 (1.01 – 1.08)	1.09 (1.04 – 1.14)	0.99 (0.92 – 1.06)	0.90 (0.79 – 1.03)	0.99 (0.91 – 1.08)	0.91 (0.79 – 1.06)	1.05 (0.77 – 1.44)	1.08 (0.83 – 1.42)
Women, n=83,369 (54%)
Low vs. High	1.11 (1.07 – 1.16)	1.27 (1.18 – 1.36)	1.18 (1.12 – 1.24)	1.30 (1.20 – 1.41)	0.96 (0.89 – 1.04)	1.10 (0.91 – 1.32)	1.06 (0.97 – 1.16)	1.27 (0.98 – 1.62)	1.12 (0.73 – 1.71)	0.93 (0.57 – 1.51)
Medium Vs. High	1.04 (1.00 – 1.08)	1.13 (1.05 – 1.20)	1.08 (1.03 – 1.14)	1.15 (1.06 – 1.23)	0.98 (0.90 – 1.05)	0.88 (0.73 – 1.06)	0.95 (0.86 – 1.06)	1.16 (0.91 – 1.46)	0.80 (0.51 – 1.24)	1.34 (0.89 – 2.03)
Men, n=71,111 (46%)
Low vs. High	1.11 (1.06 – 1.16)	1.02 (0.96 – 1.10)	1.16 (1.09 – 1.22)	1.08 (1.00 – 1.17)	1.00 (0.88 – 1.15)	1.00 (0.80 – 1.24)	1.02 (0.92 – 1.14)	0.77 (0.64 – 0.94)	1.83 (1.13 – 2.97)	1.01 (0.65 – 1.57)
Medium Vs. High	1.02 (0.97 – 1.07)	1.01 (0.95 – 1.06)	1.01 (0.95 – 1.06)	1.05 (0.99 – 1.11)	1.01 (0.89 – 1.15)	0.93 (0.77 – 1.12)	1.02 (0.90 – 1.16)	0.80 (0.68 – 0.95)	1.35 (0.83 – 2.19)	0.91 (0.60 – 1.36)

Obesity defined as ≥30 kg/m² vs. without obesity.

All model adjusted for age (18–30, 31–49, 50+ years), sex/gender (women or men), race/ethnicity (NH-White, NH-Black, Hispanic/Latinx, and NH-Asian), educational attainment (<high school, high school graduate, some college, ≥college), occupational class (professional/management, support services, laborers), food security status (food secure or insecure), region of residence (Northeast, Midwest, South, West), and marital/co-habiting status (married/living with partner or cohabitating, divorced/widowed/separated, single/no live-in partner).

Note. All estimates are weighted for the survey’s complex sampling design. Boldface indicates statistically significant results at the 0.05 level.

Among NH-White women, stronger associations were among those with an annual income of ≥$75,000 compared to <$75,000: low vs. high nSC was associated with a higher prevalence of obesity in those who earned annual incomes of ≥$75,000 (PR=1.30 [95% CI: 1.20–1.41]) than those with annual household incomes of <$75,000 (PR=1.18 [95% CI: 1.12–1.24]). A similar pattern emerged between medium vs. high nSC. Among Hispanic/Latinx adults who earned annual incomes of ≥$75,000, low vs. high nSC was associated with a higher prevalence of obesity in women (PR=1.27 [95% CI: 0.98–1.62]) and a lower prevalence of obesity in men (PR=0.77 [95% CI: 0.64–0.94]). Finally, among NH-Asians who earned <$75,000 (PR=1.42 [95% CI: 1.03–1.96]), and specifically NH-Asian men who earned <$75,000 (PR=1.83 [95% CI: 1.13–2.97]), living in low vs. high nSC was associated with a higher prevalence of obesity (Table 4).

Neighborhood Social Cohesion and Obesity by Race/Ethnicity-Sex/Gender-Food Security Status

Participants who were food secure and lived in a neighborhood with low vs. high social cohesion had a higher prevalence of obesity (PR=1.13 [95% CI: 1.10–1.16]), but there was no association among those who were food insecure (PR=1.03 [95% CI: 0.96–1.11]) (Table 5). Participants who were food secure and lived in a neighborhood with medium vs. high social cohesion had a higher prevalence of obesity (PR=1.05 [95% CI: 1.02–1.07]), but there was no association among those who were food insecure (PR=0.99 [95% CI: 0.91–1.07]). NH-White adults who were food secure living in a neighborhood with low vs. high nSC had a higher prevalence of obesity (PR=1.18 [95% CI: 1.14–1.22]) while there was a marginal association among those who were not food secure (PR=1.09 [95% CI: 0.99–1.20]). Women who were food secure living in a neighborhood with low vs. high social cohesion had a higher prevalence of obesity (PR=1.18 [95% CI: 1.14–1.23]), and this association was more robust than the association observed among men who were food secure (PR=1.07 [95% CI: 1.03–1.12]). Women who were food secure living in a neighborhood with medium vs. high social cohesion had a higher prevalence of obesity (PR=1.08 [95% CI: 1.04–1.12]) while there was a null association among men (PR=1.01 [95% CI: 0.98–1.05]). Among NH-White adults who were food secure, low vs. high nSC was associated with a higher prevalence of obesity in women (PR=1.25 [95% CI: 1.19–1.30]), and associations were more robust than those among men (PR=1.12 [95% CI: 1.07–1.18]). Among NH-White men who were food insecure, low vs. high nSC was associated with a higher prevalence of obesity (PR=1.18 [95% CI: 1.00–1.39]). Similar patterns were seen among Hispanic/Latinx food-secure women (PR=1.13 [95% CI: 1.02–1.26]). Among NH-Asian women who were food insecure, low vs. high nSC (PR=0.44 [95% CI: 0.20–0.97]) and medium vs. high nSC (PR=0.20 [95% CI: 0.08–0.52]) was associated with a lower prevalence of having obesity (Table 5).

Table 5.

Fully Adjusted Prevalence Ratios of Obesity by Low and Medium Compared to High Neighborhood Social Cohesion Among the Overall Study Population and across White, Black, Hispanic/Latinx, and Asian Participants Stratified by Sex/Gender and Food Security Status, National Health Interview Survey, 2013–2018 (N=154,480)

Neighborhood Social Cohesion	Overall N=154,480		NH-White n=103,920, 67%		NH-Black n=19,074, 12%		Hispanic/Latinx n=23,207, 15%		NH-Asian n=8,279, 5%
	Food Insecure n=16,308(10%)	Food Secure n=138,172(90%)	Food Insecure n=8,067(8%)	Food Secure n=95,853(92%)	Food Insecure n=3,904(20%)	Food Secure n=15,170(80%)	Food Insecure n=3,901(17%)	Food Secure n=19,306(83%)	Food Insecure n=436(5%)	Food Secure n=7,843(95%)
All, N=154,480
Low vs. High	1.03 (0.96 – 1.11)	1.13 (1.10 – 1.16)	1.09 (0.99 – 1.20)	1.18 (1.14 – 1.22)	0.99 (0.87 – 1.13)	0.99 (0.92 – 1.06)	0.98 (0.84 – 1.13)	1.02 (0.94 – 1.10)	0.69 (0.35 – 1.34)	1.22 (0.97 – 1.54)
Medium Vs. High	0.99 (0.91 – 1.07)	1.05 (1.02 – 1.07)	1.00 (0.90 – 1.11)	1.07 (1.04 – 1.10)	1.02 (0.89 – 1.17)	0.95 (0.89 – 1.01)	0.91 (0.76 – 1.09)	0.97 (0.89 – 1.06)	0.63 (0.29 – 1.33)	1.09 (0.89 – 1.34)
Women, n=83,369 (54%)
Low vs. High	1.02 (0.94 – 1.10)	1.18 (1.14 – 1.23)	1.05 (0.93 – 1.17)	1.25 (1.19 – 1.30)	1.01 (0.88 – 1.17)	0.98 (0.90 – 1.06)	0.99 (0.82 – 1.19)	1.13 (1.02 – 1.26)	0.44 (0.20 – 0.97)	1.07 (0.77 – 1.48)
Medium Vs. High	0.98 (0.90 – 1.07)	1.08 (1.04 – 1.12)	1.01 (0.89 – 1.15)	1.12 (1.07 – 1.17)	1.01 (0.87 – 1.18)	0.94 (0.86 – 1.02)	0.91 (0.72 – 1.15)	1.02 (0.92 – 1.14)	0.20 (0.08 – 0.52)	1.15 (0.86 – 1.54)
Men, n=71,111 (46%)
Low vs. High	1.07 (0.94 – 1.21)	1.07 (1.03 – 1.12)	1.18 (1.00 – 1.39)	1.12 (1.07 – 1.18)	0.93 (0.72 – 1.22)	1.01 (0.90 – 1.15)	0.96 (0.75 – 1.22)	0.92 (0.82 – 1.03)	0.86 (0.30 – 2.46)	1.38 (1.00 – 1.90)
Medium Vs. High	1.00 (0.87 – 1.14)	1.01 (0.98 – 1.05)	1.00 (0.84 – 1.20)	1.03 (0.99 – 1.08)	1.03 (0.78 – 1.35)	0.97 (0.86 – 1.09)	0.93 (0.71 – 1.21)	0.93 (0.82 – 1.00)	1.28 (0.47 – 3.52)	1.05 (0.76 – 1.45)

Obesity defined as ≥30 kg/m² vs. without obesity.

All model adjusted for age (18–30, 31–49, 50+ years), sex/gender (women or men), race/ethnicity (NH-White, NH-Black, Hispanic/Latinx, and NH-Asian), annual household income (<$35,000, $35,000-$74,999, $75,000+), educational attainment (<high school, high school graduate, some college, ≥college), occupational class (professional/management, support services, laborers), region of residence (Northeast, Midwest, South, West), and marital/co-habiting status(married/living with partner or cohabitating, divorced/widowed/separated, single/no live-in partner).

Note. All estimates are weighted for the survey’s complex sampling design. Boldface indicates statistically significant results at the 0.05 level.

DISCUSSION

Among a large, nationally representative sample of adults in the US, we found that low and medium perceived nSC was associated with a higher prevalence of obesity. Consistent with our hypothesis, results suggest that participants who perceived their neighborhood as having low social cohesion had a higher prevalence of obesity. However, we found that these associations were stronger among NH-White compared to NH-Black and Hispanic/Latinx adults, which did not align with our hypothesis. Nonetheless, NH-Black and Hispanic/Latinx adults with obesity comprised a larger proportion of the population in neighborhoods with low vs. high social cohesion, which could translate into higher health consequences or burdens³². Furthermore, our results showed differences by sex/gender and age, with stronger associations among women compared to men and among older compared to younger adults. While we did not observe differences by annual household income in the overall population, we found differences by income across sex/gender (e.g., women living in low vs. high nSC had a higher prevalence of having obesity) and across sex/gender by race/ethnicity (e.g., NH-Asian men living in low vs. high nSC who earned lower income had a higher prevalence of having obesity). Our results regarding differences in food security status varied.

Consistent with prior studies^4,33, our study showed that individuals who perceived their neighborhood social cohesion as low or medium vs. high had a higher prevalence of obesity. A recent systematic review reported that 15 of 22 studies observed that higher levels of social capital, including higher social cohesion, were associated with lower levels of obesity, also measured via BMI⁴. Further, 12 of 17 studies found that social capital protected against high BMI through social cohesion (e.g., trust, collective efficacy)⁴. This emphasizes the potential positive impact that high nSC may have on lowering the prevalence of obesity.

NH-Black and Hispanic/Latinx adults with obesity comprised a larger proportion of the population in neighborhoods with low vs. high social cohesion could translate into higher health consequences or burdens. Specifically, we observed a higher percentage of NH-Black participants with obesity living in lower socially cohesive neighborhoods (41.4%) than NH-White participants (33.9%). While our model results suggest that perceived low vs. high nSC was associated with a higher prevalence of obesity, more so among NH-White adults vs. minoritized adults, which is similar to prior studies³⁴, there is a study conducted in South Carolina among an NH-Black sample that found higher levels of social cohesion were associated with lower levels of BMI³⁵. As such, this illustrates the need for public health interventions tailored for minoritized individuals. Also, future research is warranted to examine the association between nSC and obesity among a large sample of minoritized racial/ethnic groups, given the consistently higher prevalence of high BMI among NH-Black adults¹.

Our finding that perceived lower nSC was associated with a higher prevalence of obesity among women compared to men was consistent with prior studies^16,36. A different study conducted in Brazil found that living in the least vs. most socially cohesive neighborhoods had higher odds of obesity among women but not among men²⁹. Similarly, South Asian women living in California and neighborhoods with low vs. high social cohesion had higher BMI levels³⁰. Prior literature also has documented that women compared to men are more influenced by social factors, including social norms, support networks, and perception of neighborhood safety^16,37. Since women have a higher prevalence of obesity than men¹, these results also demonstrate the importance of intervening in nSC to address sex/gender disparities. Our study further expands on the literature by showing a stronger association between low vs. high nSC and obesity among NH-White women compared to NH-White men.

Furthermore, older adults (≥50 years old) who perceived lower nSC had a higher prevalence of obesity, which corroborated prior studies³⁶. We also observed this association regardless of sex/gender, reaffirming the importance of the neighborhood environment for older adults. Findings demonstrated that NH-White older adults and Hispanic/Latinx older adults in neighborhoods with lower social cohesion had a higher prevalence of obesity. However, another study using NHIS data did not find a significant association between nSC and obesity among racially/ethnically diverse older adults²⁵. Since that study used the mean of nSC rather than three levels, this may explain differences in our findings. Further, a novel finding of our study includes the association between low vs. high nSC and obesity among NH-White older men and Hispanic/Latinx, older men. These findings suggest that nSC may be important for older adults and racial/ethnic and racial/ethnic-sex/gender groups.

Similar to a prior study using NHIS data²⁵, we did not find differences between nSC and obesity by annual household income level in the overall population. However, we did observe differences by income across sex/gender groups where women with higher household incomes (≥$75,000) living in low vs. high nSC had stronger associations with a higher level of obesity compared to men with higher household incomes. Likewise, we observed this relationship among NH-White women and Hispanic/Latinx adults. Conversely, we found NH-Asian men living in low vs. high nSC who earned lower income (<$75,000) had a higher prevalence of obesity, which represented a strong association and may demonstrate the importance of income to NH-Asians across sex/gender given that NH-Asian adults appear, on average, less likely to have obesity compared to other racial/ethnic groups¹.

Our results indicated that in the overall population, those who were food secure and lived in neighborhoods with lower social cohesion had a higher prevalence of obesity, which is inconsistent with our hypothesis. We also observed this association among women and NH-White adults. However, NH-White men who were food insecure and living in neighborhoods with lower social cohesion had a higher prevalence of obesity. Given that food-insecure participants comprised a small portion of the study population (<10%), it is possible that we did not have enough power to detect meaningful differences. Nevertheless, this is aligned with a previous study that reported an association between low community social capital and food insecurity²². Social capital may improve access to social support in times of need and thus improve food security status and, ultimately prevalence of obesity²¹. Similarly, a longitudinal study found that perceptions of lower nSC were related to food insecurity among families living in low-income neighborhoods³⁸. A more recent study in Philadelphia among mothers of preschool-aged children concluded that lower food insecurity prevalence was associated with higher perceived nSC³⁹. Another study in California suggested that nSC may be relevant for reducing the risk of food insecurity among minoritized racial/ethnic mothers, including Latina mothers⁴⁰. These results, taken together, suggest that having trusting neighborhood relationships may provide a resource for those who are food insecure and that nSC likely improves health by sharing resources, knowledge, and information⁴¹.

Our results support the ecological model of obesity, which posits that the neighborhood environment and social factors contribute to conditions that cause obesity (e.g., individual lifestyle choices)⁴². nSC, and the neighborhood environment in general, are hypothesized to influence obesity through a biological stress response pathway⁴³. Biological mechanisms of energy metabolism are affected by different features of the built, social, and economic environment that produce a given distribution of eating and physical activity⁹. These conditions enable or constrain eating and physical activity and are embodied in biological systems to affect these behaviors⁹. For instance, people who live in less socially cohesive neighborhoods may perceive less safety in their surroundings^44,45. This can lead to prolonged activation of the sympathetic nervous system and increased secretion of stress hormones, which may lead to chronic stress. Chronic stress is directly related to having obesity⁴⁶. nSC may also influence obesity by reinforcing unhealthy norms, such as promoting the consumption of unbalanced diets or not consuming foods in moderation to cope with stress⁴⁷. Further study of potential mechanisms is warranted.

Our study has limitations. First, we cannot infer the causality of associations between nSC and obesity, given our study’s cross-sectional design. Individuals with higher levels of BMI may report lower social cohesion in their neighborhood environment due to stigma, for example, which may lead to reverse causation⁴⁸. Second, self-reported data may bias our results; however, self-reported weight and height to calculate BMI has been validated^49,50. Similarly, included participants had a similar prevalence of obesity to excluded participants, and included participants had higher socioeconomic status (e.g., annual household income) and better health parameters (e.g., met PA guidelines). Also, excluded participants were more likely to be minoritized racial/ethnic groups. Therefore, selection bias is possible. Third, although it is convenient, the use of BMI to capture obesity has been noted to have limitations (e.g., it does not distinguish fat adequately from the fat-free mass; initially created for Belgians and thus may not translate to other populations⁵¹) where other measures (e.g., waist circumference⁵²) may be more appropriate; however, this was not measured in the NHIS. Fourth, while it is important to measure the totality of people’s neighborhood environment (e.g., work environments where people spend most of their day), such neighborhood measures were not available in the NHIS. A fifth and final limitation to note is the NHIS use of a binary sex/gender definition that excludes transgender and non-binary individuals.

Our study also has several noteworthy strengths. Our assessment of potential modification of the nSC-obesity relationship by race/ethnicity, sex/gender, age, annual household income, and food security status helps fill a gap in the literature by examining associations across these stratified groups. Food security, in particular, has not been reviewed extensively by prior literature²² may be an important confounder considering its association with food access and consumption and, consequently, obesity. Further, our large sample size allowed us to conduct robust stratification among these important potential modifiers selected a priori. Our results are generalizable to NH-White, NH-Black, Hispanic/Latinx, and NH-Asian adults in the US. While not all racial/ethnic groups were included in this study due to sample size limitations, these are the first findings, to our knowledge, that can be generalized to a diverse group in the US. Lastly, using self-reported data for nSC is important, given that perceived measures of the neighborhood social environment are more strongly associated with health outcomes than objective measures⁵³.

These results can inform future interventions focused on nSC funneling resources into neighborhoods to enhance social cohesion, which may help mitigate the impact of obesity in the US population. For example, neighborhoods can foster social cohesion through the built environment, such as community centers, parks, and recreational facilities. A cost-effective approach may be joint-use agreements, which schools and communities have developed to allow community members to access school recreational facilities after school hours. These can foster social cohesion by enhancing opportunities for social engagement via physical activity. Further, community-based participatory interventions demonstrate moderate evidence of preventing obesity⁵⁴. For example, the Shape Up Somerville Program engages multiple community members to collaborate on issues related to walkability, fitness, and more⁵⁴. While initially focused on preventing weight gain among children, this program may be replicated and expanded to foster social cohesion among adults, specifically sociodemographic groups such as women, to reduce obesity levels. Another initiative, which utilizes the principles and practices of community organizing, successfully obtained more than 70 policy wins across 21 communities of color to alter the underlying social conditions driving childhood obesity in communities of color⁵⁵. Furthermore, another avenue is 15-minute communities, where all essential needs (e.g., work, grocery stores, entertainment, community gathering places, education, and healthcare) are within a fifteen-minute walk of one’s home⁵⁶. By providing infrastructure (e.g., housing, businesses, and public spaces) that meets all needs without having to drive, it is possible to improve neighborhood walkability and increase space for sustainable projects (e.g., community gardens, outdoor sporting events, charitable fundraisers), which can enhance cohesion⁵⁷. Since previous research has highlighted the existing social inequalities in 15-minute walkable neighborhoods⁵⁷, it is essential for future interventions to address the consequences of current interventions (e.g., displacement). It is also important to note that barriers to adhering to lifestyle interventions aimed at weight loss exist, including environmental and societal pressures, socioeconomic constraints, and lack of time¹⁰. Therefore, a macro-level approach is needed that recognizes the historical and existing context of economic and public policies, which has led to the industrialization of the food industry¹². Interventions like food taxation, advertisement restrictions, or banning items are insufficient unless they address existing agricultural policies and political structures¹².

Ultimately, our findings suggest that the neighborhood social environment is associated with obesity, underscoring the importance of this upstream determinant of obesity disparities and that the neighborhood social environment may be a point of intervention for improving health. Additionally, policies and interventions that enhance favorable contextual factors, including nSC, will likely contribute to solving the obesity epidemic⁹. Understanding that obesity is a complex system in which behavior is affected by multiple individual-level factors and socioenvironmental factors, a systems-oriented approach is needed to address the various factors and levels⁹.

HIGHLIGHTS.

• Prevalence of having obesity was highest in NH-Black and Hispanic/Latinx

• Low neighborhood social cohesion was associated with higher obesity in US adults

• Association between low cohesion and having obesity was strong in older participants

• Association between low cohesion and having obesity was strong in women

• Efforts to address nSC may improve health and address health disparities

ABBREVIATIONS

BMI

Body Mass Index

CI

Confidence Interval

NH

Non-Hispanic

nSC

Neighborhood social cohesion

PR

Prevalence Ratio

US

United States