Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 May 12.
Published in final edited form as: Prev Med. 2026 Apr 28;208:108589. doi: 10.1016/j.ypmed.2026.108589

Walkable neighborhoods are associated with lower psychological distress through device-measured physical activity: Evidence from the All of Us Research Program, United States

Joonyoung Lee a,*, Yonghua Yan b, Caixia Chen b, Eun Seong Kim a
PMCID: PMC13159079  NIHMSID: NIHMS2171712  PMID: 42061729

Abstract

Objective:

Neighborhood walkability may influence mental health through physical activity, but the behavioral pathway and underlying mechanisms remain unclear. We examined whether walkability affects psychological distress through physical activity, and whether benefits operate via metabolic improvements or direct pathways independent of body mass index (BMI).

Methods:

We analyzed data from 5858 United States All of Us Research Program participants (mean age 58.5 ± 15.6 years, 69% female, data collected 2018–2023) with device-measured physical activity and validated psychological distress assessments. Structural equation modeling (SEM) tested pathways from perceived neighborhood walkability to psychological distress through moderate-to-vigorous physical activity (MVPA) and BMI, controlling for age and neighborhood socioeconomic disadvantage.

Results:

Walkability significantly predicted greater MVPA (β = 0.11, p < .001), which predicted lower psychological distress (β = −0.11, p < .001). This effect was intensity-specific: only MVPA demonstrated associations with both walkability and all distress indicators. Pathway decomposition showed 85% of MVPA’s protective effect operated through direct pathways rather than BMI reduction (15%).

Conclusions:

Neighborhood walkability is associated with lower psychological distress primarily through MVPA-mediated pathways that operate independently of BMI changes. These findings support walkable infrastructure investments as upstream, population-level mental health interventions aligned with health-at-every-size frameworks, with infrastructure designed to facilitate MVPA.

Keywords: Neighborhood walkability, Psychological distress, Moderate-to-vigorous physical activity, Built environment, Mediation analysis

1. Introduction

Mental health disorders affect nearly one in seven people worldwide (World Health Organization, 2025), including over 59 million United States adults annually (National Institute of Mental Health, 2024). Despite growing recognition as a public health priority, significant barriers persist (e.g., costs, workforce shortages, and stigma) limiting clinical interventions’ population-level impact (Alegría et al., 2023; Henking et al., 2023). Consequently, public health research has shifted toward upstream structural determinants, identifying the built environment as a critical lever for mental health promotion (Kousoulis and Goldie, 2021; Shah et al., 2021). Neighborhood socioeconomic disadvantage is a key contextual factor in this relationship, as disadvantaged areas tend to have lower walkability (Conderino et al., 2021) and higher mental health burden (Alegría et al., 2023), making it an essential covariate in walkability–mental health research.

Neighborhood walkability is a modifiable structural feature characterized by high street connectivity, mixed land use, and robust pedestrian infrastructure (Conderino et al., 2021). Unlike individual interventions requiring sustained motivation, walkable design integrates healthy behaviors into daily routines (Sallis et al., 2006, 2016), with natural experiments demonstrating increased physical activity following relocation to walkable areas (Ding et al., 2018). Walkable neighborhoods facilitate 68–89 additional weekly minutes of physical activity across diverse populations (Sallis et al., 2016). Given that regular physical activity is associated with a 17–25% reduction in depression risk and a 26% reduction in anxiety (Pearce et al., 2022; Schuch et al., 2018, 2019), walkability offers a structural strategy for population mental health. Evidence suggests walkability may both increase physical activity and amplify its mental health benefits (Asiamah et al., 2022).

While some studies demonstrate protective effects of walkability on specific mental health outcomes (e.g., depressive symptoms; Berke et al., 2007) and general psychological distress (Sarkar et al., 2013), others found no association (Siqueira Junior et al., 2022). These inconsistencies likely stem from several methodological limitations. First, studies have examined narrow demographic groups—older men only (Sarkar et al., 2013), older adults in single cities (Berke et al., 2007; Siqueira Junior et al., 2022), or women only (James et al., 2017)—limiting generalizability. Second, walkability has been operationalized inconsistently, with varying definitions and assessment approaches hampering cross-study synthesis (Warner et al., 2022). Third, the behavioral pathway linking walkability to mental health through physical activity remains inadequately tested. Studies either omit physical activity measurement (Sarkar et al., 2013) or rely on self-reported measures (Asiamah et al., 2022; Berke et al., 2007; Siqueira Junior et al., 2022), introducing recall bias and preventing intensity-specific analysis. Few studies have used device-measured physical activity to test mediation or identify intensities drive these benefits. In addition, sedentary behavior—distinct from insufficient physical activity— has rarely been examined alongside active behaviors in walkability research. Fourth, walkability–mental health research has focused almost exclusively on depressive symptoms, with limited examination of anxiety or perceived stress. To date, research has rarely assessed walkability’s relationship with depression, anxiety, and stress simultaneously as indicators of a unified psychological distress construct, leaving the broader mental health implications of walkable environments unclear.

Moreover, the mechanisms through which physical activity exerts its protective effects have yet to be established. Two broad categories of pathways have been proposed. The first involves metabolic pathways, grounded in evidence that higher body mass index (BMI) increases depression risk in a dose-response pattern (Luppino et al., 2010). The second involves non-metabolic pathways operating independently of body weight changes. A recent systematic review found strong evidence for non-metabolic mechanisms including psychological pathways (e.g., self-efficacy, affect, social support) and neurobiological pathways (e.g., BDNF upregulation, HPA-axis regulation, inflammation reduction), but only moderate evidence for metabolic pathways like BMI, suggesting non-metabolic mechanisms may predominate (Kandola et al., 2019; White et al., 2024). However, the relative contribution of metabolic versus non-metabolic pathways has not been quantified. Establishing these proportional contributions has critical implications for “health-at-every-size” approaches to mental health promotion, as interventions could emphasize physical activity engagement and direct mental health benefits rather than weight-focused outcomes.

The present study addresses these limitations by leveraging data from the All of Us (AoU) Research Program, a large-scale, nationally representative cohort (All of Us Research Program Investigators, 2019). Unlike previous studies confined to narrow demographic groups, our analysis includes adults across the full lifespan with device-measured physical activity from Fitbit wearables, substantially reducing recall bias. Using structural equation modeling (SEM), we decompose the total effect of perceived neighborhood walkability on psychological distress (depression, anxiety, and stress) into direct pathways and indirect pathways mediated by objective physical activity levels, with BMI examined as a continuous mediator to capture the full dose-response gradient of its association with distress (Luppino et al., 2010). We address three research questions: (1) Is higher neighborhood walkability associated with lower psychological distress? (2) Does moderate-to-vigorous physical activity (MVPA) mediate this relationship? (3) Does MVPA’s protective effect operate primarily through direct pathways or through metabolic improvements indexed by BMI?

2. Methods

2.1. Study design and population

Data were obtained from the AoU Research Program, a National Institutes of Health (NIH)-funded initiative enrolling over 633,000 diverse United States participants between 2018 and 2023 through electronic health records, physical measurements, surveys, and wearable devices (All of Us Research Program Investigators, 2019). The present study used the Controlled Tier Dataset, version 8 (All of Us Research Program, 2023). This study analyzed secondary, de-identified data approved by the NIH Institutional Review Board (IRB) with informed consent obtained from all participants. The authors’ IRB determined this study exempt from review.

Although the AoU Research Program is a longitudinal cohort, this analysis utilized a cross-sectional design in which all survey, anthropometric, and wearable device variables were treated as time-invariant summary measures. The initial cohort included 14,226 adults (≥18 years) with linked Fitbit data. After quality control exclusions for insufficient wear time (<7 valid wear days of ≥600 min of heart rate monitoring per day; Master et al., 2022), missing walkability or distress data, and incomplete covariates, the final analytic sample included 5858 participants (see Fig. 1).

Fig. 1.

Fig. 1.

Open in a new tab

Participant flow diagram for United States adults with Fitbit data in the All of Us Research Program.

Note. This figure illustrates the step-by-step inclusion and exclusion process used to derive the final analytic sample from the All of Us Research Program. HR = heart rate; ADI = Area Deprivation Index; BMI = body mass index; MVPA = moderate-to-vigorous physical activity.

Participants (Table 1) averaged 58.5 ± 15.6 years and were predominantly female (69.1%), White (79.0%), and highly educated (71.1% with a bachelor’s degree or higher). The median of 6559 daily steps (IQR: 4590, 8874) fell below the 8000–10,000 step threshold associated with reduced chronic disease risk (Master et al., 2022).

Table 1.

Demographics and sample characteristics of United States adults with Fitbit data in the All of Us Research Program (N = 5858).a, b, c, d, e

Age, years (mean ± SD) 58.5 ± 15.6
Gender identity, n (%)
Woman 4050 (69.1)
Man 1773 (30.3)
Other/not reporteda 35 (0.6)
Race, n (%)
White 4626 (79.0)
Black or African American 352 (6.0)
Asian 255 (4.4)
More than one race 174 (3.0)
Other racesb 189 (1.7)
Not reported 327 (5.9)
Ethnicity, n (%)
Not Hispanic or Latino 5263 (89.8)
Hispanic or Latino 490 (8.4)
Not reported 106 (1.8)
Education level, n (%)
Advanced degree (Master’s/PhD) 2300 (39.3)
College graduate (bachelor’s) 1865 (31.8)
Some college (1–3 years) 1336 (22.8)
High school diploma/GED 307 (5.2)
Less than high school / not reportedc 50 (0.9)
Marital status, n (%)
Married 3410 (58.2)
Not marriedd 2419 (41.4)
Not reported 29 (0.4)
Employment status, n (%)
Employed 2795 (47.7)
Retired 1692 (28.9)
Self-employed 339 (5.8)
Unable to work 376 (6.4)
Homemaker 171 (2.9%)
Student 220 (3.8)
Other/nor reportede 265 (4.5)
Annual household income, n (%)
< $35,000 1058 (18.1)
$35,000–$74,999 1430 (24.4)
$75,000–$149,999 1928 (32.9)
≥ $150,000 1141 (19.5)
Not reported 305 (5.1)
Neighborhood median income, USD (mean ± SD) 66,851 ± 16,970
Anthropometrics
Weight, kg (mean ± SD) 82.9 ± 22.3
Height, cm (mean ± SD) 167.7 ± 9.2
Device-measured physical activity
Daily steps, median (IQR) 6559 (4590, 8874)
Open in a new tab

Note. Values are presented as mean ± SD or n (%), as appropriate.

SD = standard deviation; GED = General Educational Development; USD = United States dollars; IQR = interquartile range.

a

Other/Not reported includes gender diverse (nonbinary, transgender, and other identities), skipped responses, prefer not to answer, and responses with no matching concept in the source data

b

Other races include Middle Eastern or North African, Native Hawaiian or other Pacific Islander, and none of these.

c

Less than high school/Not reported aggregates grades 9–11, grades 5–8, and missing education responses.

d

Not married includes never married, divorced, widowed, separated, and living with a partner.

e

Other/Not reported employment includes out of work (< 1 year), out of work (≥ 1 year), and missing employment responses.

2.2. Measures

2.2.1. Physical activity

Device-measured physical activity was assessed using Fitbit data from the activity_summary table (All of Us Research Program Investigators, 2019). Fitbit devices demonstrate acceptable validity compared with research-grade accelerometers (Master et al., 2022) and provide daily step counts and minute-level intensity classifications (Bailey et al., 2025). Activity minutes were categorized into four intensity levels based on metabolic equivalent (MET) thresholds: sedentary (< 1.5 METs), lightly active (1.5–3.0 METs), fairly active (3.0–6.0 METs, moderate intensity), and very active (> 6.0 METs, vigorous intensity; Bailey et al., 2025), consistent with definitions underlying World Health Organization physical activity guidelines. Light physical activity (LPA) was defined as lightly active minutes, and MVPA as the sum of fairly and very active minutes (≥3.0 METs). Daily averages were calculated across all valid wear days.

2.2.2. Body mass index

BMI (kg/m2) was calculated from self-reported height and weight.

2.2.3. Neighborhood walkability

Perceived walkability was assessed using five items from the AoU Social Determinants of Health survey (All of Us Research Program Investigators, 2019). Participants rated agreement on a 4-point scale (1 = strongly disagree to 4 = strongly agree): (1) shops and destinations are within walking distance, (2) a transit stop is within a 10–15 min walk, (3) sidewalks are available on most streets, (4) bicycling facilities are available, and (5) recreation facilities are available. Total scores ranged from 5 to 20, with higher scores indicating greater walkability (α = 0.78). Perceived walkability was used as GIS-based measures are unavailable in the AoU dataset; perceived scales may more directly capture the behavioral experience of a neighborhood.

2.2.4. Psychological distress

Psychological distress was assessed using three validated measures: the GAD-7 (anxiety; 7 items, 0–21 range; α = 0.90; Spitzer et al., 2006), PHQ-9 (depression; 9 items, 0–27 range; α = 0.87; Kroenke et al., 2001), and PSS-10 (stress; 10 items, 0–40 range; α = 0.91; Cohen et al., 1983). All use higher scores indicating greater distress. Positively worded PSS-10 items were reverse-coded.

2.2.5. Area deprivation index

Neighborhood disadvantage was assessed using the area deprivation index (ADI; Kind and Buckingham, 2018), a composite measure of 17 Census indicators linked to participants’ ZIP codes. Values range from 0 (least disadvantaged) to 1 (most disadvantaged). ADI was included as a covariate given its associations with both lower walkability and elevated mental health risk (Alegría et al., 2023; Conderino et al., 2021).

2.3. Statistical analysis

Analyses were conducted using Python 3.10 in the Researcher Workbench. Internal consistency was assessed using Cronbach’s alpha (≥0.70 = acceptable; Cronbach, 1951). To address RQ1, Pearson correlations examined associations among walkability, physical activity behaviors (MVPA, LPA, sedentary time), and psychological distress. Physical activity behaviors demonstrating significant associations with both walkability and distress were retained as candidate mediators (see Results). To address RQs 2 and 3, we tested pathways linking walkability to psychological distress through MVPA and BMI using structural equation modeling (SEM) with maximum likelihood estimation in semopy 2.3.8 (Igolkina and Meshcheryakov, 2020). Psychological distress was modeled as a latent variable (GAD-7, PHQ-9, and PSS-10; factor loadings: 0.97–1.00).

We compared Model 1 (sequential: walkability → MVPA → BMI → distress) and Model 2 (parallel: adding direct MVPA → distress pathway). Both controlled for age (predicting MVPA, BMI, distress) and ADI (predicting MVPA, distress). Model fit was evaluated using χ2/df, RMSEA (≤0.08), CFI/TLI (≥0.90), and SRMR (≤0.08; Hu and Bentler, 1999). Models were compared via chi-square difference testing, with Model 2 selected as the final model (see Results). Indirect effects were estimated using 5000 bootstrap iterations: (1) walkability → MVPA → distress and (2) walkability → MVPA → BMI → distress. Listwise deletion was used; missingness resulted primarily from survey nonresponse and insufficient device wear. All tests used a significance level of α = 0.05. Given the large sample size, we interpreted both statistical significance and effect size magnitude, with standardized path coefficients (β) interpreted per Cohen’s (1988) conventions.

3. Results

Bivariate correlations among study variables are presented in Table 2. Regarding RQ1, higher neighborhood walkability showed small but significant association with depression (r = −0.05, p < .001) and stress (r = −0.03, p < .05), though not with anxiety (r = −0.02, p = .103). Notably, walkability was positively associated with MVPA (r = 0.09, p < .001), which in turn showed consistent inverse associations with all three psychological distress indicators (rs = −0.14 to −0.18, ps < 0.001). In contrast, walkability was not associated with LPA (r = −0.01, p = .395) or sedentary behavior (r = 0.02, p = .224). LPA showed inconsistent associations with distress indicators, and sedentary behavior demonstrated only weak positive associations (rs = 0.03 to 0.04, ps < 0.05). Given that only MVPA demonstrated significant associations with both walkability and psychological distress, MVPA was retained as the mediator in subsequent analyses. In addition, MVPA demonstrated a significant inverse association with BMI (r = −0.22, p < .001), and BMI showed significant positive associations with all three distress indicators (rs = 0.12 to 0.19, ps < 0.001), supporting the inclusion of BMI as a potential secondary mediator in the metabolic pathway. Age and ADI demonstrated expected patterns of associations with study variables, justifying their inclusion as covariates.

Table 2.

Pearson correlations among walkability, physical activity, psychological distress, and covariates among United States adults in the All of Us Research Program (N = 5858).

Variable 1 2 3 4 5 6 7 8 9 10
1. Walkability –
2. MVPA (mins/day) 0.09*** –
3. LPA (mins/day) −0.01 0.30*** –
4. Sedentary behavior (mins/day) 0.02 −0.26*** −0.47*** –
5. Depression −0.05*** −0.18*** −0.08*** 0.04** –
6. Anxiety −0.02 −0.14*** 0.01 0.03* 0.79*** –
7. Stress −0.03* −0.17*** −0.02 0.04** 0.77*** 0.80*** –
8. Age −0.07*** 0.08*** −0.14*** 0.04** −0.36*** −0.41*** −0.43*** –
9. BMI −0.10*** −0.22*** −0.19*** 0.13*** 0.19*** 0.12*** 0.13*** −0.06*** –
10. ADI −0.02 −0.05*** −0.01 0.01 0.04*** 0.04** 0.04** −0.06*** 0.07*** –
Mean 14.0 33.6 190.5 867.3 5.5 4.9 13.4 58.5 29.4 0.31
Std. dev. 4.4 28.8 67.6 204.0 5.3 4.8 7.9 15.6 7.4 0.06
Open in a new tab

Note. Values are Pearson correlation coefficients (r). MVPA = moderate-to-vigorous physical activity; LPA = light physical activity; BMI = body mass index; ADI = Area deprivation Index.

***

p < .001,

**

p < .01,

*

p < .05.

To address RQ2 and RQ3, we compared two structural models examining pathways from walkability to psychological distress through MVPA and BMI. Model 1 (sequential mediation) specified that MVPA affects distress exclusively through BMI (walkability → MVPA → BMI → psychological distress). Model 2 (parallel mediation) additionally included a direct MVPA → psychological distress pathway, testing whether MVPA’s protective effect operates through direct psychological distress, BMI, or both. Model 2 demonstrated superior fit to Model 1 across all indices (Model 1: χ2[18] = 366.90, RMSEA = 0.058, CFI = 0.977, TLI = 0.964, SRMR = 0.042; Model 2: χ2[17] = 270.73, RMSEA = 0.050 [90% CI: 0.047, 0.057], CFI = 0.983, TLI = 0.972, SRMR = 0.036; Δχ2 = 96.17, Δdf = 1, p < .001) and was selected as the final model.

Fig. 2 presents the final model (Model 2) with all standardized path coefficients. All hypothesized paths were statistically significant. Higher neighborhood walkability predicted greater MVPA (β = 0.11, p < .001). MVPA, in turn, predicted both lower BMI (β = −0.21, p < .001) and directly predicted lower psychological distress (β = −0.11, p < .001). BMI also predicted higher psychological distress (β = 0.09, p < .001). A significant residual direct effect of walkability on psychological distress remained (β = −0.04, p < .001), indicating additional unmeasured pathways. Among covariates, age demonstrated strong protective effects across all outcomes (βs = 0.09 to −0.39; ps < 0.001), while ADI predicted lower MVPA (β = −0.04, p < .001) but not psychological distress (β = 0.01, p = .531). The model explained 37.2% of variance in psychological distress, 4.9% in BMI, and 1.8% in MVPA.

Fig. 2.

Fig. 2.

Open in a new tab

The final structural equation model (Model 2: parallel mediation) examining pathways from neighborhood walkability to psychological distress among United States adults in the All of Us Research Program (N = 5858).

Note. Standardized path coefficients are shown. The model tests whether MVPA mediates the walkability–psychological distress relationship and whether this effect operates through direct pathways or through BMI reduction. All hypothesized paths were statistically significant. Solid lines represent hypothesized paths; dashed lines represent covariate paths. Depression, Anxiety, and Stress are indicators of the latent psychological distress construct, with factor loadings shown. The model explained 1.8% of the variance in MVPA, 4.9% of the variance in BMI, and 37.2% of the variance in the latent psychological distress construct. MVPA = moderate-to-vigorous physical activity; BMI = body mass index; ADI = Area Deprivation Index. ***p < .001.

Addressing RQ2, the total indirect effect of walkability on psychological distress through MVPA was significant (β = −0.0128, p < .001), confirming MVPA as a significant mediator. Addressing RQ3, pathway decomposition revealed that 85% of this effect operated through direct pathways (walkability → MVPA → psychological distress: β = −0.0109), with only 15% mediated through BMI (walkability → MVPA → BMI → psychological distress: β = −0.0020). Equivalently, of MVPA’s total effect on psychological distress (β = −0.129), 85% operated directly (β = −0.11) rather than through BMI reduction (β = −0.019). Sensitivity analyses confirmed MVPA’s protective effect remained unchanged when controlling for BMI (β = −0.11, p < .001).

4. Discussion

This study examined whether neighborhood walkability influences psychological distress (depression, anxiety, stress) through physical activity behaviors, and whether these benefits operate through metabolic improvements or direct pathways. Three main findings emerged. First, while direct bivariate associations between walkability and psychological distress were small and inconsistent—significant for depression and stress but not anxiety. SEM revealed significant indirect effects through MVPA, confirming it as a behavioral mediator. Second, among device-measured physical activity behaviors, only MVPA, not LPA or sedentary behavior, demonstrated significant associations with both walkability and distress, meeting prerequisites for mediation. Third, decomposition of MVPA’s protective effect revealed that 85% operated through direct pathways, with only 15% mediated through BMI, indicating that physical activity’s mental health benefits are largely independent of BMI changes.

The finding of significant mediation despite small and inconsistent bivariate associations illustrates how indirect pathways can be obscured in bivariate analyses but revealed through SEM. While previous studies have demonstrated protective associations between walkability and mental health, most either omitted physical activity measurement or relied on self-reported measures (Asiamah et al., 2022; Berke et al., 2007; Sarkar et al., 2013; Warner et al., 2022), preventing rigorous pathway quantification and intensity-specific analysis. By quantifying this pathway using device-measured physical activity, our findings provide empirical support for ecological models linking built environments to health outcomes through behavioral mechanisms (Sallis et al., 2006, 2016) and demonstrate that walkability’s mental health benefits operate specifically through MVPA rather than general movement, an intensity-specific distinction undetectable in prior self-report-based studies.

Furthermore, the 85%/15% decomposition provides the first quantitative evidence for White et al.’s (2024) qualitative conclusion that non-metabolic mechanisms predominate. Although higher BMI increases psychological distress risk (Luppino et al., 2010), our findings indicate that BMI accounts for only a small proportion of physical activity’s protective effect, with dominant mechanisms likely involving neurobiological and psychosocial processes (Kandola et al., 2019; White et al., 2024). This has critical implications for intervention design: physical activity promotion programs should emphasize direct mental health benefits rather than weight-focused messaging, aligning with health-at-every-size frameworks.

Our findings also reveal critical intensity-specificity: only MVPA demonstrated associations with both walkability and all distress indicators, consistent with meta-analytic evidence favoring higher-intensity physical activity for mental health benefits (Pearce et al., 2022; Schuch et al., 2018, 2019). For urban planning, infrastructure must support vigorous movement rather than merely enabling leisurely strolling. Such walkability improvements represent upstream, population-level interventions (Kousoulis and Goldie, 2021; Shah et al., 2021), particularly important given persistent sociodemographic disparities in physical activity adherence (Lee et al., 2024) and the role of walkability and social determinants in mental health inequities (Alegría et al., 2023; Conderino et al., 2021). The persistence of a direct walkability effect on psychological distress (β = −0.04) after accounting for MVPA indicates additional unmeasured pathways warrant investigation.

This study’s cross-sectional design limits causal inference and cannot rule out residential self-selection. Although the use of device-measured physical activity is a significant strength, psychological distress was assessed at a single time point, precluding evaluation of longitudinal trajectories and bidirectional effects. Residual confounding from unmeasured factors (e.g., chronic health conditions, lifestyle factors, and neighborhood safety) may remain despite adjustment for age and socioeconomic indicators. Both BMI and perceived neighborhood walkability were self-reported, which may introduce measurement error and underestimate the true associations. Even though approximately 85% of effects were independent of BMI, underlying non-metabolic mechanisms could not be identified due to the absence of neurobiological and psychosocial mediators. Finally, findings from U.S. AoU participants may not generalize to contexts with different urban forms, transportation systems, or walking cultures.

5. Conclusions

This study provides the first quantification of the walkability → MVPA → psychological distress pathway, demonstrating that 85% of MVPA’s protective effects operate through BMI-independent pathways. These benefits are intensity-specific, accruing through MVPA but not LPA, with implications for infrastructure design that must support vigorous physical activity. Given the global mental health crisis and growing recognition that social and environmental determinants drive mental health outcomes, investments in walkable infrastructure represent a promising upstream, population-level mental health intervention scalable to entire communities.

Acknowledgment

We gratefully acknowledge All of Us participants for their contributions, without whom this research would not have been possible. We also thank the National Institutes of Health’s All of Us Research Program for making available the participant data examined in this study. We additionally acknowledge Dr. ConSandra McNeil and Dr. Azad R. Bhuiyan for providing training on the All of Us Research Program.

Funding

This research was supported in part by the Delta Research and Educational Foundation (grant # OT2OD025337, sponsored by the National Institutes of Health). The first and second authors (J.L., Y.Y.) were supported as trainees through this grant. The funder had no role in study design, data collection, analysis, interpretation, or manuscript preparation.

Footnotes

Human subjects approval statement

This study analyzed secondary de-identified data from the AoU Research Program. The protocol was approved by the National Institutes of Health’s Institutional Review Board (IRB), and all participants provided informed consent at enrollment. The Jackson State University IRB determined this study to be exempt from review (Protocol #0105–26).

CRediT authorship contribution statement

Joonyoung Lee: Writing – review & editing, Writing – original draft, Visualization, Resources, Project administration, Methodology, Investigation, Formal analysis, Conceptualization. Yonghua Yan: Writing – review & editing, Supervision, Investigation, Formal analysis. Caixia Chen: Writing – review & editing, Supervision, Formal analysis. Eun Seong Kim: Writing – review & editing, Visualization, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare no competing financial interests or personal relationships that could have influenced this work.

Data availability

This study used data from the All of Us Research Program’s Controlled Tier Dataset version 8, available to authorized users on the Researcher Workbench (https://workbench.researchallofus.org). Access requires institutional signing of a Data Use and Registration Agreement, completion of identity verification, the All of Us Responsible Conduct of Research training, and signing of the Data User Code of Conduct.

References

  1. Alegría M, Alvarez K, Cheng M, Falgas-Bague I, 2023. Recent advances on social determinants of mental health: looking fast forward. Am. J. Psychiatry 180 (7), 473–482. 10.1176/appi.ajp.20230371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. All of Us Research Program, 2023. All of us controlled tier dataset v8 [data set]. National Institutes of Health. https://www.researchallofus.org. [Google Scholar]
  3. All of Us Research Program Investigators, 2019. The “All of Us” research program. N. Engl. J. Med 381 (7), 668–676. 10.1056/NEJMsr1809937. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Asiamah N, Vieira ER, Kouveliotis K, Gasana J, Awuviry-Newton K, Eduafo R, 2022. Associations between older African academics physical activity, walkability and mental health: a social distancing perspective. Health Promot. Int 37 (2), daab093. 10.1093/heapro/daab093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bailey CP, Dodd KW, McClain JJ, Seo I, Wheeler W, Wolff-Hughes DL, 2025. Fitbit physical activity and sleep data in the all of us research program: data exploration and processing considerations for research. Med. Sci. Sports Exerc 57 (12), 2946–2953. 10.1249/MSS.0000000000003804. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Berke EM, Gottlieb LM, Moudon AV, Larson EB, 2007. Protective association between neighborhood walkability and depression in older men. J. Am. Geriatr. Soc 55 (4), 526–533. 10.1111/j.1532-5415.2007.01108.x. [DOI] [PubMed] [Google Scholar]
  7. Cohen J, 1988. Statistical Power Analysis for the Behavioral Sciences, 2nd ed. Lawrence Erlbaum; https://www.utstat.toronto.edu/~brunner/oldclass/378f16/readings/CohenPower.pdf. [Google Scholar]
  8. Cohen S, Kamarck T, Mermelstein R, 1983. A global measure of perceived stress. J. Health Soc. Behav 24, 385–396. [PubMed] [Google Scholar]
  9. Conderino SE, Feldman JM, Spoer B, Gourevitch MN, Thorpe LE, 2021. Social and economic differences in neighborhood walkability across 500 U.S. cities. Am. J. Prev. Med 61 (3), 394–401. 10.1016/j.amepre.2021.03.014. [DOI] [PubMed] [Google Scholar]
  10. Cronbach LJ, 1951. Coefficient alpha and the internal structure of tests. Psychometrika 16 (3), 297–334. 10.1007/BF02310555. [DOI] [Google Scholar]
  11. Ding D, Nguyen B, Learnihan V, Bauman AE, Davey R, Jalaludin B, Gebel K, 2018. Moving to an active lifestyle? A systematic review of the effects of residential relocation on walking, physical activity and travel behaviour. Br. J. Sports Med 52 (12), 789–799. 10.1136/bjsports-2017-098833. [DOI] [PubMed] [Google Scholar]
  12. Henking C, Reeves A, Chrisinger B, 2023. Global inequalities in mental health problems: understanding the predictors of lifetime prevalence, treatment utilisation and perceived helpfulness across 111 countries. Prev. Med 177, 107769. 10.1016/j.ypmed.2023.107769. [DOI] [PubMed] [Google Scholar]
  13. Hu L, Bentler PM, 1999. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct. Equ. Model. Multidiscip. J 6 (1), 1–55. 10.1080/10705519909540118. [DOI] [Google Scholar]
  14. Igolkina AA, Meshcheryakov G, 2020. Semopy: a python package for structural equation modeling. Struct. Equ. Model. Multidiscip. J 27 (6), 952–963. 10.1080/10705511.2019.1704289. [DOI] [Google Scholar]
  15. James P, Hart JE, Hipp JA, Mitchell JA, Kerr J, Hurvitz PM, Glanz K, Laden F, 2017. GPS-based exposure to greenness and walkability and accelerometry-based physical activity. Cancer Epidemiol. Biomarkers Prev 26 (4), 525–532. 10.1158/1055-9965.EPI-16-0925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kandola A, Ashdown-Franks G, Hendrikse J, Sabiston CM, Stubbs B, 2019. Physical activity and depression: towards understanding the antidepressant mechanisms of physical activity. Neurosci. Biobehav. Rev 107, 525–539. 10.1016/j.neubiorev.2019.09.040. [DOI] [PubMed] [Google Scholar]
  17. Kind AJH, Buckingham WR, 2018. Making neighborhood-disadvantage metrics accessible—the neighborhood atlas. N. Engl. J. Med 378 (26), 2456–2458. 10.1056/NEJMp1802313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kousoulis AA, Goldie I, 2021. A visualization of a socio-ecological model for urban public mental health approaches. Front. Public Health 9, 654011. 10.3389/fpubh.2021.654011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kroenke K, Spitzer RL, Williams JBW, 2001. The PHQ-9: validity of a brief depression severity measure. J. Gen. Intern. Med 16 (9), 606–613. 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lee J, Kim ES, Lee H, Huh JH, 2024. Decadal trends in physical activity adherence among Korean older adults: an analysis of national survey of older Korean data, 2011–2020. J. Aging Phys. Act 32 (4), 461–471. 10.1123/japa.2023-0341. [DOI] [PubMed] [Google Scholar]
  21. Luppino FS, de Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BWJH, Zitman FG, 2010. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch. Gen. Psychiatry 67 (3), 220–229. 10.1001/archgenpsychiatry.2010.2. [DOI] [PubMed] [Google Scholar]
  22. Master H, Annis J, Huang S, Beckman JA, Ratsimbazafy F, Marginean K, Carroll R, Natarajan K, Harrell FE, Roden DM, Harris P, Brittain EL, 2022. Association of step counts over time with the risk of chronic disease in the all of us research program. Nat. Med 28 (11), 2301–2308. 10.1038/s41591-022-02012-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. National Institute of Mental Health, 2024. Mental Illness. https://www.nimh.nih.gov/health/statistics/mental-illness.
  24. Pearce M, Garcia L, Abbas A, Strain T, Schuch FB, Golubic R, Kelly P, Khan S, Utukuri M, Laird Y, Mok A, Smith A, Tainio M, Brage S, Woodcock J, 2022. Association between physical activity and risk of depression: a systematic review and meta-analysis. JAMA Psychiatry 79 (6), 550–559. 10.1001/jamapsychiatry.2022.0609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Sallis JF, Cervero RB, Ascher W, Henderson KA, Kraft MK, Kerr J, 2006. An ecological approach to creating active living communities. Annu. Rev. Public Health 27, 297–322. 10.1146/annurev.publhealth.27.021405.102100. [DOI] [PubMed] [Google Scholar]
  26. Sallis JF, Cerin E, Conway TL, Adams MA, Frank LD, Pratt M, Salvo D, Schipperijn J, Smith G, Cain KL, Davey R, Kerr J, Lai P-C, Mitáš J, Reis R, Sarmiento OL, Schofield G, Troelsen J, Van Dyck D, Owen N, 2016. Physical activity in relation to urban environments in 14 cities worldwide: a cross-sectional study. Lancet 387 (10034), 2207–2217. 10.1016/S0140-6736(15)01284-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Sarkar C, Gallacher J, Webster C, 2013. Urban built environment configuration and psychological distress in older men: results from the caerphilly study. BMC Public Health 13 (1), 695. 10.1186/1471-2458-13-695. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Schuch FB, Vancampfort D, Firth J, Rosenbaum S, Ward PB, Silva ES, Hallgren M, Ponce De Leon A, Dunn AL, Deslandes AC, Fleck MP, Carvalho AF, Stubbs B, 2018. Physical activity and incident depression: a meta-analysis of prospective cohort studies. Am. J. Psychiatry 175 (7), 631–648. 10.1176/appi.ajp.2018.17111194. [DOI] [PubMed] [Google Scholar]
  29. Schuch FB, Stubbs B, Meyer J, Heissel A, Zech P, Vancampfort D, Rosenbaum S, Deenik J, Firth J, Ward PB, Carvalho AF, Hiles SA, 2019. Physical activity protects from incident anxiety: a meta-analysis of prospective cohort studies. Depress. Anxiety 36 (9), 846–858. 10.1002/da.22915. [DOI] [PubMed] [Google Scholar]
  30. Shah N, Walker IF, Naik Y, Rajan S, O’Hagan K, Black M, Cartwright C, Tillmann T, Pearce-Smith N, Stansfield J, 2021. National or population level interventions addressing the social determinants of mental health – an umbrella review. BMC Public Health 21 (1), 2118. 10.1186/s12889-021-12145-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Siqueira Junior JDA, Lopes AADS, Godtsfriedt CES, Justina MDD, Paiva KMD, d’Orsi E, Rech CR, 2022. Neighbourhood walkability and mental health in older adults: a cross-sectional analysis from EpiFloripa aging study. Front. Aging 3, 915292. 10.3389/fragi.2022.915292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Spitzer RL, Kroenke K, Williams JB, Löwe B, 2006. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch. Intern. Med 166 (10), 1092–1097. 10.1001/archinte.166.10.1092. [DOI] [PubMed] [Google Scholar]
  33. Warner E, Nadouri D, Orpana H, Wang J, 2022. A scoping review of the definition of walkability and its relationship with depression and anxiety symptoms. Can. J. Commun. Ment. Health 41 (2), 46–67. 10.7870/cjcmh-2022-014. [DOI] [Google Scholar]
  34. White RL, Vella S, Biddle S, Sutcliffe J, Guagliano JM, Uddin R, Burgin A, Apostolopoulos M, Nguyen T, Young C, Taylor N, Lilley S, Teychenne M, 2024. Physical activity and mental health: a systematic review and best-evidence synthesis of mediation and moderation studies. Int. J. Behav. Nutr. Phys. Act 21 (1), 134. 10.1186/s12966-024-01676-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. World Health Organization, 2025. Mental Disorders. https://www.who.int/news-room/fact-sheets/detail/mental-disorders.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

This study used data from the All of Us Research Program’s Controlled Tier Dataset version 8, available to authorized users on the Researcher Workbench (https://workbench.researchallofus.org). Access requires institutional signing of a Data Use and Registration Agreement, completion of identity verification, the All of Us Responsible Conduct of Research training, and signing of the Data User Code of Conduct.

RESOURCES