951006721632Health PlaceHealth PlaceHealth & place1353-82921873-205434116496832891710.1016/j.healthplace.2021.102587NIHMS1707612ArticlePathways Linking Census Tract Typologies with Subjective Neighborhood Disorder and Depressive Symptoms in the Black Women’s Experiences Living with Lupus (BeWELL) StudyMartzConnor D.a*HunterEvelyn A.bKramerMichael R.cWangYijiedChungKaraeBrownMichaelfDrenkardCristinacgLimS. SamcgChaeDavid H.eDepartment of Human Development and Family Science, Auburn University, 203 Spidle Hall, Auburn, AL 36849 USADepartment of Special Education, Rehabilitation, and Counseling, Auburn University, 2084 Haley Center, Auburn, AL 36849 USADepartment of Epidemiology, Emory University Rollins School of Public Health, 1518 Clifton Rd. NE, Atlanta, GA 30322 USADepartment of Human Development and Family Studies, Michigan State University, 552 W. Circle Dr, East Lansing, MI 48824 USADepartment of Global Community Health and Behavioral Sciences, Tulane University School of Public Health and Tropical Medicine, 1440 Canal St, New Orleans, LA 70112 USASchool of Kinesiology, Auburn University, 301 Wire Rd., Auburn, AL 36849 USADepartment of Medicine, Division of Rheumatology, Emory University School of Medicine, 1658 Clifton Rd. A, Atlanta, GA 30322 USA

Author Contributions: CDM had full access to the data in the study and takes responsibility for the accuracy of the data analysis. Study concept and design: CDM, MK, DHC. Acquisition of data: DHC, SSL, CD. Analysis and interpretation of data: CDM, EAH, YW, MK, DHC. Drafting of the manuscript: CDM. Critical revisions of the manuscript for important intellectual content: CDM, EAH, MK, YW, KC, MB, CD, SSL, DHC.

Corresponding Author. Address: 203 Spidle Hall, Auburn, AL 36849 USA. connordmartz@auburn.edu2652021096202172021017202270102587102587

Depression is a common comorbidity among Black women with systemic lupus erythematosus (SLE), an understudied autoimmune disease characterized by major racial and gender inequities. Research is needed that examines how area-level factors influence risk of depression in this population. Latent profile analysis revealed four neighborhood typologies among metropolitan Atlanta, Georgia census tracts that participants (n=438) in the Black Women’s Experiences Living with Lupus (BeWELL) Study were living in: Integrated/High-SES, Moderately Segregated/Mid-SES, Highly Segregated/Mid-SES, and Highly Segregated/Low-SES. Structural equation models indicated that highly segregated census tracts were associated with the greatest levels of depression via increased subjective assessments of neighborhood disorder. Policies that invest in segregated areas and address physical and social aspects of the environment that contribute to neighborhood disorder may promote mental health among Black women with SLE.

systemic lupus erythematosusdepressionracial residential segregationneighborhood disorderstructural racismlatent profile analysis

Systemic lupus erythematosus (SLE) is a chronic autoimmune condition that involves several organ systems and unpredictable periods of disease activity that wax and wane [1]. Symptoms of SLE are heterogeneous and potentially debilitating, such as chronic pain, fatigue, face and skin rashes, ulcers, hair loss, and joint and chest pain [1]. Up to 262,000 people in the US live with SLE and Black women have the highest prevalence rate [2–5]. Black women are also burdened with faster disease progression, worse disease outcomes (including mortality) and at earlier ages than White women [6, 7]. Recent evidence suggests racial disparities in disease outcomes are widening, including in the prevalence of lupus comorbidities, such as depression, which has been associated with worse disease outcomes [8–12]. It is estimated that 24–35% of people living with SLE are diagnosed with major depressive disorder, making it one of the most common comorbid conditions observed in SLE patients [13–15]. The etiology of depression within SLE is complex and manifold, including clinical, inflammatory, and psychosocial mechanisms as well as their interactions [7, 16]. Growing evidence suggests that qualitatively unique psychosocial stressors experienced by Black Americans may contribute to racial disparities in SLE severity, which may be partly driven by comorbid depression [7, 17, 18]. For example, psychosocial stressors related to trauma, racism, and their intersections with gender–all risk factors for depression–have been associated with disease incidence and worse outcomes among women with SLE [17, 19–23]. Chronic exposure to psychosocial stress can produce elevated levels of inflammation which cause or exacerbate depressive symptoms, in turn accelerating the SLE disease course [24–28]. Given that Black women are more likely than White women to experience chronic psychosocial stress throughout the lifecourse, the comorbid occurrence of depression with SLE may hasten disease progression for Black women and contribute to racial SLE disparities [29–32].

Neighborhood factors and area-level characteristics, including aspects of the physical, service, and social environment (e.g., quality of the built environment, neighborhood poverty, and neighborhood disorder) are associated with mental health outcomes [33, 34]. For example, census tract poverty and various aspects of physical and social disorder (vandalism, crime, low social cohesion) have been associated with greater risk and severity of depression [35–37]. The inequitable allocation of these neighborhood health-promoting versus health-damaging factors are structured and reinforced by racial residential segregation. Racial residential segregation is characteristic of major metropolitan areas in the US, and has been intentionally designed by meso-level actors to maintain White supremacy [38, 39]. One major consequence has been the creation of predominantly Black neighborhoods that on average have greater systematic disadvantages compared to White neighborhoods [38, 40].

There is a paucity of research on neighborhood-level factors and their association with depression specifically among people living with SLE. Prior research on SLE and neighborhood characteristics is limited and most empirical studies focus on associations between area-level poverty and outcomes of organ damage and mortality [41, 42]. For example, geographic variation in SLE incidence and mortality are concentrated in areas with high rates of poverty and high proportions of racial and ethnic minorities [44]. Recent qualitative work identified that fear of crime and exposure to neighborhood disorder is a stressor in the lives of SLE patients, particularly for those living in areas of concentrated poverty [45]. Aspects of neighborhood disorder, such as crime, physical decay, and low social cohesion have also been associated with depressive symptoms and poorer SLE outcomes among patients living in Egypt [46]. Taken together, this evidence suggests that contextual neighborhood attributes, including racial residential segregation and poverty concentration, are linked to area-level stressors that may partially account for racial SLE disparities through mental health mechanisms [75]. However, there are few studies that examine this relationship, and research among Black women is particularly sparse.

Emerging methods in area-level research have begun to incorporate neighborhood-centered techniques which address methodological limitations of more traditional variable-centered approaches [47, 48]. Latent profile analysis (LPA) is one method that can be used to identify practically meaningful and distinct groupings of types of neighborhoods based on heterogeneity in objective area-level characteristics [49]. Past studies have used LPA to identify neighborhood typologies based on various area-level indicators, such as socioeconomic disadvantage, alcohol availability, food environments, green space, and air pollution [49].

Black and White Americans reside in qualitatively distinct neighborhoods largely due to the legacy and persistence of racial residential segregation, concurrent economic segregation, and contemporary racial discrimination in housing and lending [38, 40]. In the current study, we: (1) use LPA to identify distinct census tract typologies that Black women with SLE live in; and (2) examine associations between census tract typologies, subjective neighborhood disorder, and depressive symptoms. We hypothesize that participants living in tract typologies associated with greater perceived neighborhood disorder will report more depressive symptoms.

MethodsSample

Data are from the Black Women’s Experiences Living with Lupus (BeWELL) Study. Details of the BeWELL sample have been previously described [19]. Briefly, the BeWELL Study consists of 438 Black women living in metropolitan Atlanta, GA with validated SLE based on guidelines set by the American College of Rheumatology (≥ 4 criteria or three criteria with documented diagnosis by board-certified rheumatologist) [50]. Participants were recruited from April 2015 to May 2017 from the Georgians Organized Against Lupus (GOAL) cohort. The GOAL cohort was largely derived from the Georgia Lupus Registry (GLR), a population-based SLE registry funded by the Centers of Disease Control and Prevention [2, 51]. The GLR was designed to estimate the incidence and prevalence of SLE in Atlanta, GA and ascertained cases from various sources including hospitals, health care providers, state and regional laboratories, and electronic medical records [2]. In addition to the assessment of psychosocial factors, SLE outcomes, and other health conditions, the BeWELL Study geocoded participant addresses to US census tracts in metropolitan Atlanta.

VariablesArea-Level

Sociodemographic and socioeconomic data on Atlanta Metropolitan Statistical Area (MSA) census tracts were obtained from American Community Survey five-year estimates (2014–2018) [52]. The following tract-level socioeconomic indicators were matched to geocoded participant addresses and used to characterize latent tract typologies: (1) percent poverty, measured as the percent of individuals living at or below the federal poverty threshold; (2) percent unemployment is the percent of individuals who are unemployed and actively looking for employment; (3) percent on Supplemental Nutritional Assistance Program (SNAP) is the percent of individuals receiving federal assistance or SNAP; and (4) percent with a high school degree or less measured as the percent of individuals without any post-secondary education.

The evenness dimension of racial residential segregation was assessed using divergence scores which quantify the difference between Black-White racial composition at the tract-level to the broader Atlanta MSA [53]. Divergence values for individual tracts range from 0 (no segregation) to a maximum value dependent on the proportions of each group in the MSA. Greater relative divergence within an MSA indicates that a single race predominates the tract above and beyond the average MSA tract. We use the following formula to calculate divergence values for tract i in the BeWELL sample: Di=∑m=1Mπimlogπimπm where πim is race m’s proportion of the total Black-White population in tract i, and πm is race m’s proportion of the overall Black-White MSA population [53]. Values of Di do not indicate the racial composition of census tracts, only their divergence from the racial composition of the overall MSA. Percent of Black individuals at the census tract-level was used to examine the racial composition of census tracts; when used for participants all residing within a single MSA, it is considered to be an indicator of racial isolation [54].

Individual-LevelSubjective Neighborhood Disorder.

Items measuring subjective neighborhood disorder were from the Perceived Neighborhood Disorder Scale [55]. Four items assessed neighborhood social disorder. Example items are: “There are too many people hanging around on the streets near my home” and “There is a lot of crime in my neighborhood”. Participants indicated their agreement to each item, with response choices ranging from 1 (strongly disagree) to 5 (strongly agree). We examined the mean score across items (α=.90).

Physical disorder

Physical disorder was measured using six items that assessed physical conditions of participants’ neighborhoods, including the presence of graffiti, abandoned buildings, vandalism, and cleanliness. We examined the mean of response choices, which also ranged from 1 (strongly disagree) to 5 (strongly agree) (α=.88).

Depressive Symptoms.

Six items from the Patient-Reported Outcomes Measurement Information System (PROMIS) Short Form v1.0 - Depression 6a instrument were used to assess depressive symptomology [56, 57]. PROMIS short-form instruments for depression were developed by the National Institutes of Health, are validated, precise, and reliable measures, and have been correlated with other established measures, such as the Center for Epidemiological Studies Depression (CES-D) Scale [58]. Participants reported how frequently they experienced feeling the following in the past seven days: worthless, helpless, depressed, hopeless, like a failure, and unhappy. Response options ranged from “never” (1) to “always” (5). Scores were summed and converted to standardized T-Scores, where a T-score of 50 and a standard deviation of 10 represents the general population average for depressive symptoms. T-Scores range from 38.4–80.3 with higher scores indicating greater depressive symptoms.

Covariates.

Demographic covariates were age in years and SLE duration in years. Socioeconomic covariates included categories of educational attainment (less than high school, high school degree, some college or associates/technical degree, bachelor’s or graduate or professional degree), work status (full time, part time, out of labor force, or unable to work), insurance status (public, private, or none), and income-to-poverty ratio measured continuously. Health-related covariates included body mass index (measured continuously as weight (kg) divided by height (m2)), smoking status (current vs. not current), and the following categories of current SLE medication use, each measured dichotomously (yes vs. no): steroids, hydroxychloroquine, and immunomodulators.

Analytic Strategy

There were a total of 257 census tracts represented in the sample. Most (89.0%) of the census tracts in the sample included only one or two participants (Mean=1.7, max=5) which is insufficient clustering for use of multilevel models [59]. Hence, census-derived variables were examined as individual-level variables. Descriptive statistics and bivariate correlations for all study variables were examined prior to conducting analyses. Individual-level missing data was minimal (<1% for any study variable) and accounted for using Full Information Maximum Likelihood (FIML) estimation within Mplus Version 8 [60, 61].

Latent profile analysis (LPA) was used to identify subgroups or typologies of census tracts based on the covariance structure of four continuous tract-level socioeconomic characteristics (percent poverty, high school degree or less, SNAP, and unemployment) and two indicators of racial residential segregation (divergence and percent Black). LPA models of increasing number of profiles (1–6) were conducted to identify the best fitting solution. The following fit indices were used to evaluate models: Bayesian Information Criteria (BIC), sample-size adjusted BIC (ABIC), entropy, and a parametric bootstrapped likelihood ratio test (BLRT). The smallest values of BIC and ABIC and highest entropy values are considered to represent the best fitting model [48]. Elbow plots of BIC and ABIC values were examined to illustrate relative gains in model fit of additional profiles [62]. The BLRT statistically compares the fit of k vs k-1 profile solutions, with a significant (p<.05) BLRT indicating the higher-profile solution represents a better fit of the data [48]. In addition to fit indices, theory in neighborhoods and health guided decisions to identify conceptually meaningful profiles and select a model of best fit [48].

ANOVA and chi-square tests examined whether census tract typologies significantly differed on sociodemographic and socioeconomic characteristics, subjective social and physical neighborhood disorder, and depressive symptoms. The most likely typology assignments were then merged with individual-level participant data for structural equation models (SEM). First, a measurement model was specified which indexed the two measures of perceived neighborhood factors as a latent construct of subjective neighborhood disorder. Standardized factor loadings for the measurement model were evaluated for measurement quality and good model fit (CFI>.95; RMSEA<.10, p<.0001; SRMR<.10) [63]. Next, a structural model was specified to assess direct and indirect effects between tract typologies, subjective neighborhood disorder, and depressive symptoms. Standard errors for indirect effects were estimated using the Delta method in Mplus. SEM estimates for paths linking census tract typologies to subjective neighborhood disorder, census tract typologies to depressive symptoms, and subjective neighborhood disorder to depressive symptoms adjusted for all covariates. Post-hoc sensitivity analyses were conducted to compare model estimates using FIML versus listwise deletion for observations with missing individual-level data on any study variable (n=7), and to test the direction of indirect effects modeled in primary analyses.

ResultsDescriptive Statistics

Means and correlations among study variables in Table 1. Participants in the sample were an average of 46.8 years old (SD=12.3) with a mean disease duration of 16.0 years (SD=10.4). The mean depression T-Score for our sample was 51.7 (SD=9.9) which is 1.7 units greater than the average for the US general population (M=50, SD=10).

Characteristics of census tracts represented in the BeWELL Study and the broader Atlanta MSA are also presented in Table 1. With the exception of percent of individuals with a high school degree or less, significant mean differences exist between tracts in BeWELL and the Atlanta MSA. Compared to all populated tracts in the Atlanta MSA (n=948), tracts in the BeWELL sample (n=257) have significantly higher rates of poverty, SNAP recipients, and unemployment. Tracts in the BeWELL sample also have, on average, greater divergence scores and higher proportions of Black individuals compared to the MSA, indicating that BeWELL participants live in more segregated census tracts than the average Atlanta resident.

Latent Profile Analysis Results

Results from the LPA indicated that a 4-profile solution best characterized census tracts in the BeWELL sample (Table 2). Posterior probabilities ranged from .94 to .95 indicating high concordance in the likelihood of latent profile membership. Although the 5- and 6-profiles solution had lower BIC and ABIC, higher entropy, and significant BLRT, these profiles were not conceptually distinct from one another (i.e., additional racially integrated and mid-SES typologies). Therefore, we determined that the 4-profile solution best characterized the sample and represented unique neighborhood typologies.

The first latent tract profile of the 4-profile solution was defined by several indicators of socioeconomic affluence and racially integrated demographics (29.9% Black, divergence=0.08) (Table 3). We named this tract typology Integrated/High-SES. The second profile indicated moderate socioeconomic disadvantage and a significantly higher percent of Black individuals compared to Profile 1 (44.1% Black, divergence=0.13). We named this typology Moderately Segregated/Mid-SES. The third tract profile was also characterized by moderate levels of socioeconomic disadvantage but was predominantly Black and racially segregated (86.0% Black, divergence=0.63). We classified this tract typology as Highly Segregated/Mid-SES. The final latent profile had a similar racial composition as Profile 3 (predominantly Black (84.7%) and racially segregated (divergence=0.62)), but had significantly greater socioeconomic disadvantage relative to other profiles. This tract typology was described as Highly Segregated/Low-SES. Figure 1 depicts standardized mean values for all indicators across unique tract typologies.

Of the 257 tracts in the sample, Highly Segregated/Mid-SES tract typologies were most prevalent (33.9%), followed by Integrated/High-SES (28.8%), Highly Segregated/Low-SES (20.2%), and Moderately Segregated/Mid-SES typologies (17.1%). Participants in the BeWELL Study (n=438) primarily lived in census tracts characterized as Highly Segregated/Mid-SES (47.0%), followed by Integrated/High-SES (19.4%) and Moderately Segregated/Mid-SES (18.5%). The remaining 15.1% of BeWELL participants lived in tracts classified as Highly Segregated/Low-SES.

Table 3 presents results of ANOVA and chi-square tests which indicated that Highly Segregated/Low-SES tracts had significantly higher rates of socioeconomic disadvantage compared to other typologies. In contrast, Integrated/High-SES tracts had significantly lower rates of socioeconomic disadvantage compared to all other typologies. There were significant, yet inconsistent, differences in most socioeconomic characteristics between Highly Segregated/Mid-SES and Moderately Segregated/Mid-SES tracts typologies. For instance, Highly Segregated/Mid-SES tracts had higher percent unemployment, but lower percent of individuals with up to a high school degree compared to the Moderately Segregated/Mid-SES tract typology.

Mean perceived social and physical disorder was lowest in Integrated/High-SES tracts (M=1.29, SD=0.43; M=1.23, SD=0.43, respectively) and highest in Highly Segregated/Low-SES tracts (M=2.11, SD=0.82; M=1.98, SD=0.64). There were similar levels of mean perceived social and physical disorder across Moderately Segregated/Mid-SES (M=1.60, SD=0.73; M=1.50, SD=0.67) and Highly Segregated/Mid-SES tracts (M=1.79, SD=0.74; M=1.66, SD=0.56). Participants in Integrated/High-SES tracts had the highest individual-level income-to-poverty ratio (M=2.63, SD=1.84), were most likely to have at least some college education (89.4%) and work full-time (38.8%), and were least likely to be uninsured (5.9%). Women who lived in Highly Segregated/Low-SES tracts reported the lowest income-to-poverty ratio (M=1.42, SD=1.26) and were most likely to have a high school degree or less (39.2%), be unable to work (65.4%), and be uninsured (14.8%). Mean levels of depressive symptoms did not significantly differ across tract typologies (F=0.71, p=0.54). Additional descriptive statistics for participant-level variables across tract typologies are presented in Table 3.

Structural Equation Model Results

Measurement model factor loadings for perceived social and physical disorder that indexed the latent construct of neighborhood disorder were 0.850 and 0.826 (p<.0001), respectively, indicating high measurement quality [63].

Initial models examined standardized direct path coefficients of tract typologies predicting subjective neighborhood disorder and depressive symptoms, separately (Fig. 2 and Table 4). Subsequent models tested indirect effects of the relationship between tract typologies and depressive symptoms through subjective neighborhood disorder (Table 5). Highly Segregated/Low-SES was the referent group for Integrated/High-SES estimates, and Integrated/High SES served as the referent group for all other typologies. Direct and indirect path estimates for additional referent group combinations are included as supplementary material (Tables S1–S6).

Direct Effects.

Model fit is good (RMSEA=0.02; CFI=0.98; SRMR=0.02) [63]. Relative to Integrated/High-SES tracts, both Highly Segregated/Low-SES (β=0.495, SE=0.059, 95% Confidence Interval (CI)=0.384, 0.624) and Highly Segregated/Mid-SES tracts (β=0.402, SE=0.052, 95% CI=0.302, 0.504) were most strongly associated with subjective neighborhood disorder after adjusting for demographic, socioeconomic, and health-related covariates. Integrated/High-SES tracts were associated with less subjective neighborhood disorder (β=−0.332, SE=0.040, 95% CI=−0.412, −0.254) compared to Highly Segregated/Low-SES tracts. Moderately Segregated/Mid-SES tracts were associated with higher subjective neighborhood disorder (β=0.184, SE=0.053, 95% CI=0.088, 0.314) than Integrated/High-SES tracts, but less disorder compared to both Highly Segregated/Mid-SES (β=−0.104, SE=0.057, 95% CI=−0.215, −0.006) and Highly Segregated/Low-SES tracts (β=−0.272, SE=0.067, 95% CI=−0.414,−0.154) (Tables S2–S3). Subjective neighborhood disorder was positively associated with depressive symptoms (β=0.134, SE=0.062, 95% CI=0.021, 0.259). No significant direct effects between tract typologies and depressive symptoms were found.

Indirect Effects.

Standardized estimates for indirect paths between tract typologies, subjective neighborhood disorder, and depressive symptoms are presented in Table 5. Integrated/High-SES tracts were indirectly associated with fewer depressive symptoms through decreased subjective neighborhood disorder, relative to Highly Segregated/Low-SES tracts (β=−0.040, SE=0.020, 95% CI=−0.088, −0.007). Compared to Integrated/High-SES tracts, all other typologies were indirectly associated with greater depressive symptoms through increased exposure to subjective neighborhood disorder. Specifically, both Highly Segregated/Low-SES (β=0.066, SE=0.032, 95% CI=0.013, 0.144) and Highly Segregated/Mid-SES tracts (β=0.054, SE=0.026, 95% CI=0.012, 0.120) were indirectly associated with elevated depressive symptoms via subjective neighborhood disorder. Moderately Segregated/Mid-SES tracts were also associated with greater depressive symptoms through subjective neighborhood disorder (β=0.025, SE=0.014, 95% CI=0.005, 0.063) compared to Integrated/High-SES tracts. In comparison to Highly Segregated/Low-SES tracts, all other tract typologies were associated with less depressive symptomology through subjective neighborhood disorder (Table S6).

Sensitivity Analyses

Results did not substantively differ when comparing models estimated using FIML (n=438) to models estimated using listwise deletion (n=431). Standardized estimates for FIML models were nearly identical to models using listwise deletion and did not meaningfully alter the interpretation of results. Post-hoc sensitivity analyses tested an alternative causal pathway (e.g., Tract Typology → Depressive Symptoms → Subjective Neighborhood Disorder). Indirect effect estimates were nonsignificant and provide support for the causal order of effects presented in the main analyses (Table S7).

Discussion

The objectives of this study were to identify latent typologies of metropolitan Atlanta, GA census tracts that BeWELL Study participants live in and examine subjective neighborhood disorder as a pathway linking tract typologies to depressive symptomology. We identified four unique tract typologies based on sociodemographic and socioeconomic characteristics: 1) Integrated/High-SES, 2) Moderately Segregated/Mid-SES, 3) Highly Segregated/Mid-SES, and 4) Highly Segregated/Low-SES. Most of the Black women in our sample (62.1%) lived in a highly segregated tract typology, reflecting the stark racial residential segregation in Atlanta, GA. An overwhelming majority of individuals living in those tracts are Black (M=85%), which significantly diverges from the percent of the Black population in the overall Atlanta MSA (33.9%). We found that participants who live in highly segregated census tracts are at greatest risk of depression through their increased exposure to neighborhood disorder, compared to racially integrated tracts. This was also the case for moderately segregated tracts, albeit to a lesser degree. These findings could offer partial explanation for racial disparities in SLE outcomes given that depression is a risk factor for disease severity among Black women [8–11].

An important conclusion from our study is that structural racism, in the form of racial residential segregation, shapes area-level socioeconomic resources, exposure to neighborhood disorder, and subsequent risk of depression among Black women in our sample. Racial residential segregation and neighborhood disadvantage is the intentional byproduct of institutional gatekeepers (e.g., lenders, politicians, executives, police and courts) who perpetuate the political economy of racial capitalism by shaping access to opportunity and resources based on the reification of race [39, 64]. Thus, the underlying cause of area-level socioeconomic disadvantage and neighborhood disorder observed in our sample is the action and inaction of these meso-level actors [39].

Participants living in Integrated/High-SES tract typologies reported less subjective neighborhood disorder which was associated with fewer depressive symptoms compared to moderately and highly segregated tracts. Among the highly segregated tract typologies, women in moderate SES tracts reported lower levels of neighborhood disorder compared to those living in low SES tracts, in concordance with prior research [65–68]. High levels of physical disorder (e.g., abandoned buildings, graffiti, litter) and social disorder (e.g., lack of trust, fear of crime, and hostility and conflict) may trigger or worsen depressive symptomology through stress-mediated pathways [37, 69]. Perceptions of neighborhood disorder also reduce predictability and controllability in neighborhoods, and can lead to residents feeling a lack of control which is a risk factor for depression [36]. These unique sources of stress should be considered in context of our sample of Black women with SLE who face inherent stressors of disease management, unpredictability of disease course, and racism [70]. Living in neighborhoods with high physical and social disorder may proliferate concurrent stressors and increase risk of experiencing depressive symptoms [35, 36, 71], which can lead to heightened disease activity through multiple pathways, such as compromised disease management [30, 72, 73, 74].

Our findings are consistent with the relatively few studies that have examined associations between sociocontextual factors and health outcomes among SLE patients. Low social cohesion and poor neighborhood aesthetics have been associated with depressive symptoms and adverse SLE outcomes, and general perceived stress has been shown to partially account for the association between high area-level poverty and increased SLE-related organ damage and mortality [10, 39, 47, 48]. Our study advances this line of research and provides evidence for the significance of racial residential segregation and subjective neighborhood conditions in shaping risk of depression–a common comorbidity in SLE–specifically among Black women.

Notably, despite relatively similar area-level socioeconomic characteristics and levels of perceived social and physical disorder, tracts in the moderately segregated middle-class typology were associated with fewer depressive symptoms compared to highly segregated middle-class tracts. These results suggest that in metropolitan Atlanta, GA, additional structural deficits that accompany high levels of segregation and which stem from meso-level factors promote neighborhood disorder [38, 76]. For example, aspects of the service environment (e.g., presence of parks, recreational facilities, and green space; access to healthcare; healthy food availability) may contribute to heightened or reduced neighborhood disorder but were not captured with the indicators (i.e., area-level SES) used in the LPA. An important direction for future research will involve defining neighborhoods with more comprehensive and encompassing measures of neighborhood social, physical, and service environment.

Structural equation models did not indicate direct associations between tract typologies and depressive symptoms, nor did we observe statistical differences in mean levels of depression across typologies. These findings should be interpreted in the unique geographic context of Atlanta which has many affluent Black neighborhoods relative to other large metropolitan areas in the US. As such, there are several health-promoting resources associated with predominantly Black neighborhoods in Atlanta, such as Black political power, high quality Black educational institutions, and social capital, which are not as salient in other highly segregated cities [40, 77, 78]. Heterogeneity in these factors across tract typologies in the BeWELL Study may explain the lack of a direct association between typologies and depressive symptoms. For instance, health-protective characteristics of living in a racially homogeneous neighborhood may offset risk factors for depressive symptoms caused by neighborhood disorder. In contrast, racially integrated neighborhoods may mitigate risk of depression through reduced neighborhood disorder but increase exposure to risk factors for depression, such as racial discrimination [79].

Results from our study suggest that efforts to address neighborhood context, including both area-level socioeconomic deprivation and neighborhood disorder, should be considered as part of a comprehensive effort to improve mental health among Black women living with SLE [35, 80]. Effective solutions for addressing these issues and promoting health equity occur at the structural level [81]. For example, recent evidence suggests that interventions that address over-policing and police violence may reduce subjective neighborhood disorder among Black women [82]. The enactment and strict enforcement of laws prohibiting housing discrimination and racial residential segregation may reduce the concentration of Black women with SLE in high-disorder neighborhoods. Housing vouchers that facilitate transitions from segregated to integrated neighborhoods are considered a cost-effective approach to reduce exposure to neighborhood disorder and promote mental health [83–86]. Similarly, homeownership, affordable housing, and anti-racist tenant agreements may reduce exposure to psychosocial stress associated with neighborhood disorder [87, 88]. Economic and public investment towards urban planning, land use, and quality housing in primarily Black communities also reduces social and physical disorder, and is beneficial to the mental health of neighborhood residents [89–94]. At the clinician-level, SLE-providers should screen for depressive symptoms and discuss mental health services if needed. Clinicians should also leverage the political power of professional medical associations to advocate for evidence-based policies that reduce physical and social neighborhood disorder and invest in racially segregated communities.

Limitations

This study used cross-sectional data and results should not be interpreted with causality. Our data does not allow us to distinguish the extent to which sociocontextual factors cause depressive symptoms or how an SLE diagnosis forces Black women to live in segregated neighborhoods with high levels of disorder, which therefore exacerbates poor mental health. However, post-hoc analyses of alternative indirect pathways provide support for the direction of effects in this study, and our results are consistent with past longitudinal studies on neighborhood socioeconomic conditions, subjective neighborhood disorder, and mental health outcomes that provide more robust causal interpretations [34, 95]. Our findings are limited to census tracts represented in the BeWELL Study throughout metropolitan Atlanta and may not be generalizable to other regions or populations, including rural areas and predominantly Black affluent neighborhoods. Compared to all tracts in the Atlanta MSA, tracts in the BeWELL sample had higher rates of socioeconomic disadvantage and greater residential segregation. Furthermore, the use of census tracts to describe participant “neighborhoods” is a limitation as census tracts are relatively large (approximately 4,000 residents/tract) and obscure within-tract heterogeneity in area-level characteristics. Future analyses should examine contextual factors, depression, and SLE at more finite levels and in non-urban geographies.

Indicators used in latent profile analyses were limited to sociodemographic and socioeconomic characteristics. There are likely other area-level characteristics, such as aspects of the service environment, that affect mental health and the SLE disease course which should be incorporated in forthcoming studies. Future research should also examine area-level mechanisms that link protective factors associated with living in a predominately Black neighborhood to depression. Another limitation is the self-reported nature of neighborhood disorder and depressive symptoms. However, subjective neighborhood assessments have been highly correlated with objective measures [96], and our measure of depressive symptoms is reliable and validated for epidemiologic research [57].

These limitations should be considered in context of the study’s methodological strengths. Our sample of 438 Black women with validated SLE was derived from an enumerated population-based registry and is among the largest to examine areal-level sociocontextual factors and depression–an important consideration given Black woman are most burdened yet underrepresented in existing areal-level SLE research. Latent profile analysis enabled us to identify unique configurations of segregation and SES census tract typologies that exist among Black women living with SLE in metropolitan Atlanta. This approach represents another methodological strength and advances area-level SLE research. Rather than examining area-level characteristics singularly, as in multivariable regression analysis, latent profile analysis allows for the simultaneous examination of multiple conceptually important influences on health [49].

Conclusions

This study addresses gaps in SLE research on area-level factors and depression, an outcome that by itself is important for those living with the disease, but which also has been shown to increase disease severity, and potentially may have a role in widening racial disparities in SLE outcomes. Our findings reveal nuances in the types of geographies that Black women living with SLE reside in, and how these are related to neighborhood disorder and depressive symptoms. Racial residential segregation–a pernicious facet of structural racism–may shape exposure to neighborhood disorder and subsequent depression, which is a risk factor for accelerated disease progression. Findings from this study highlight potential policy interventions, such as investment in racially segregated neighborhoods, to reduce neighborhood disorder and advance equitable health outcomes for Black women with SLE.

Supplementary MaterialAcknowledgments

We would like to thank Olivia Nichols, MS for providing helpful and thoughtful comments on manuscript drafts.

Sponsors Role:

The National Institute of Arthritis and Musculoskeletal and Skin Disease (NIAMS) had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The views expressed in this article are those of the authors and do not necessarily reflect those of Auburn University or the NIAMS.

Funding:

This work was supported by the National Institute of Arthritis and Musculoskeletal and Skin Disease (F31AR076234, R01AR065493).

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declarations of interest: none

ReferencesFavaA, PetriM (2019) Systemic lupus erythematosus: Diagnosis and clinical management. J Autoimmun 96:1–13. 10.1016/j.jaut.2018.11.00130448290LimSS, BayaklyAR, HelmickCG, (2014) The incidence and prevalence of systemic lupus erythematosus, 2002–2004: The Georgia Lupus Registry. Arthritis Rheumatol 66:357–368. 10.1002/art.3823924504808SomersEC, MarderW, CagnoliP, (2014) Population-based incidence and prevalence of systemic lupus erythematosus: The Michigan Lupus Epidemiology and Surveillance Program. Arthritis Rheumatol 66:369–378. 10.1002/art.3823824504809Dall’EraM, CisternasMG, SnipesK, (2017) The incidence and prevalence of systemic lupus erythematosus in San Francisco County, California: The California Lupus Surveillance Project. Arthritis Rheumatol 69:1996–2005. 10.1002/art.4019128891237IzmirlyPM, WanI, SahlS, (2017) The incidence and prevalence of systemic lupus erythematosus in New York County (Manhattan), New York: The Manhattan Lupus Surveillance Program. Arthritis Rheumatol 69:2006–2017. 10.1002/art.4019228891252LimSS (2019) Racial Disparities in Mortality Associated with Systemic Lupus Erythematosus — Fulton and DeKalb Counties, Georgia, 2002–2016. MMWR Morb Mortal Wkly Rep 68:. 10.15585/mmwr.mm6818a4DrenkardC, LimSS (2019) Update on lupus epidemiology: advancing health disparities research through the study of minority populations. Curr Opin Rheumatol 31:689. 10.1097/BOR.000000000000064631436582PalaginiL, MoscaM, TaniC, (2013) Depression and systemic lupus erythematosus: a systematic review. Lupus 22:409–416. 10.1177/096120331347722723427220ParkD-J, KangJ-H, LeeK-E, (2018) Association of depression with socioeconomic status, anticardiolipin antibodies, and organ damage in patients with systemic lupus erythematosus: results from the KORNET registry. Clin Exp Rheumatol 36:627–63529465349CalderónJ, FloresP, AguirreJ, (2017) Impact of cognitive impairment, depression, disease activity, and disease damage on quality of life in women with systemic lupus erythematosus. Scand J Rheumatol 46:273–280. 10.1080/03009742.2016.120661727701937JordanJ, ThompsonNJ, Dunlop-ThomasC, (2019) Relationships among organ damage, social support, and depression in African American women with systemic lupus erythematosus. Lupus 28:253–260. 10.1177/096120331881557330482093YenEY, ShaheenM, WooJMP, (2017) 46-Year trends in systemic lupus erythematosus mortality in the United States, 1968 to 2013: A nationwide population-based study. Ann Intern Med 167:777–785. 10.7326/M17-010229086801LimGY, TamWW, LuY, (2018) Prevalence of Depression in the Community from 30 Countries between 1994 and 2014. Sci Rep 8:2861. 10.1038/s41598-018-21243-x29434331ZhangL, FuT, YinR, (2017) Prevalence of depression and anxiety in systemic lupus erythematosus: a systematic review and meta-analysis. BMC Psychiatry 17:70. 10.1186/s12888-017-1234-128196529MoustafaAT, MoazzamiM, EngelL, (2020) Prevalence and metric of depression and anxiety in systemic lupus erythematosus: A systematic review and meta-analysis. Semin Arthritis Rheum 50:84–94. 10.1016/j.semarthrit.2019.06.01731303437Figueiredo-BragaM, CornabyC, CortezA, (2018) Depression and anxiety in systemic lupus erythematosus. Medicine (Baltimore) 97:. 10.1097/MD.0000000000011376HunterEA, SpearsEC, MartzCD, (2020) Racism-related stress and psychological distress: Black Women’s Experiences Living with Lupus study. J Health Psychol 135910532091308. 10.1177/1359105320913085VinaER, HausmannLRM, UtsetTO, (2015) Perceptions of racism in healthcare among patients with systemic lupus erythematosus: A cross-sectional study. Lupus Sci Med 2:e000110. 10.1136/lupus-2015-00011026322238ChaeDH, MartzCD, Fuller-RowellTE, (2019) Racial Discrimination, Disease activity, and organ damage: the Black Women’s Experiences Living With Lupus (BeWELL) Study. Am J Epidemiol 188:1434–1443. 10.1093/aje/kwz10531062841MartzCD, AllenAM, Fuller-RowellTE, (2019) Vicarious Racism Stress and Disease Activity: the Black Women’s Experiences Living with Lupus (BeWELL) Study. J Racial Ethn Health Disparities 6:1044–1051. 10.1007/s40615-019-00606-831215018FeldmanCH, MalspeisS, LeatherwoodC, (2019) Association of Childhood Abuse with Incident Systemic Lupus Erythematosus in Adulthood in a Longitudinal Cohort of Women. J Rheumatol 46:1589–1596. 10.3899/jrheum.19000931092723RobertsAL, MalspeisS, KubzanskyLD, (2017) Association of Trauma and Posttraumatic Stress Disorder With Incident Systemic Lupus Erythematosus in a Longitudinal Cohort of Women. Arthritis Rheumatol 69:2162–2169. 10.1002/art.4022228929625BergerM, SarnyaiZ (2015) “More than skin deep”: Stress neurobiology and mental health consequences of racial discrimination. Stress 18:1–10. 10.3109/10253890.2014.98920425407297McEwenBS (2006) Stress, Adaptation, and Disease: Allostasis and Allostatic Load. Ann N Y Acad Sci 840:33–44. 10.1111/j.1749-6632.1998.tb09546.xSuD-L, LuZ-M, ShenM-N, (2012) Roles of pro- and anti-inflammatory cytokines in the pathogenesis of SLE. J Biomed Biotechnol 2012:347141. 10.1155/2012/34714122500087EudyAM, VinesAI, DooleyMA, (2014) Elevated C-reactive protein and self-reported disease activity in systemic lupus erythematosus. Lupus 23:1460–1467. 10.1177/096120331454391525057037LiuCS, AdibfarA, HerrmannN, (2016) Evidence for Inflammation-Associated Depression. In: DantzerR, CapuronL (eds) Inflammation-Associated Depression: Evidence, Mechanisms and Implications. Springer International Publishing, Cham, pp 3–30FelgerJC, LotrichFE (2013) Inflammatory cytokines in depression: Neurobiological mechanisms and therapeutic implications. Neuroscience 246:199–229. 10.1016/j.neuroscience.2013.04.06023644052BertsiasGK, BoumpasDT (2010) Pathogenesis, diagnosis and management of neuropsychiatric SLE manifestations. Nat Rev Rheumatol 6:358–367. 10.1038/nrrheum.2010.6220458332MackayM (2015) Lupus brain fog: a biologic perspective on cognitive impairment, depression, and fatigue in systemic lupus erythematosus. Immunol Res 63:26–37. 10.1007/s12026-015-8716-326481913MeszarosZS, PerlA, FaraoneSV (2012) Psychiatric symptoms in systemic lupus erythematosus: a systematic review. J Clin Psychiatry 73:993–1001. 10.4088/JCP.11r0742522687742ElmelechY, LuH-H (2004) Race, ethnicity, and the gender poverty gap. Soc Sci Res 33:158–182. 10.1016/S0049-089X(03)00044-9BravemanP, EgerterS, WilliamsDR (2011) The Social Determinants of Health: Coming of Age. Annu Rev Public Health 32:381–398. 10.1146/annurev-publhealth-031210-10121821091195PérezE, BraënC, BoyerG, (2020) Neighbourhood community life and health: A systematic review of reviews. Health Place 61:102238. 10.1016/j.healthplace.2019.10223831735517HastingsJF, SnowdenLR (2019) African Americans and Caribbean Blacks: Perceived neighborhood disadvantage and depression. J Community Psychol 47:227–237. 10.1002/jcop.2211730706518BlairA, RossNA, GariepyG, SchmitzN (2014) How do neighborhoods affect depression outcomes? A realist review and a call for the examination of causal pathways. Soc Psychiatry Psychiatr Epidemiol 49:873–887. 10.1007/s00127-013-0810-z24414031HillTD, RossCE, AngelRJ (2005) Neighborhood Disorder, Psychophysiological Distress, and Health. J Health Soc Behav 46:170–186. 10.1177/00221465050460020416028456WilliamsDR, CollinsC (2001) Racial residential segregation: A fundamental cause of racial disparities in health. Public Health Rep 116:404–41612042604SewellAA (2016) The Racism-Race Reification Process: A Mesolevel Political Economic Framework for Understanding Racial Health Disparities. Sociol Race Ethn 2:402–432. 10.1177/2332649215626936KramerMR, HogueCR (2009) Is segregation bad for your health? Epidemiol Rev 31:178–194. 10.1093/epirev/mxp00119465747YelinE, TrupinL, YazdanyJ (2017) A prospective study of the impact of current poverty, history of poverty, and exiting poverty on accumulation of disease damage in systemic lupus erythematosus. Arthritis Rheumatol 69:1612–1622. 10.1002/art.4013428480630ShahAK, MullainathanS, ShafirE (2012) Some Consequences of Having Too Little. Science 338:682–685. 10.1126/science.122242623118192YelinE, YazdanyJ, TrupinL (2017) Relationship between poverty and mortality in systemic lupus erythematosus. Arthritis Care Res 70:1101–1106. 10.1002/acr.23428WalshSJ, DechelloLM (2001) Geographical variation in mortality from systemic lupus erythematosus in the United States. Lupus 10:637–646. 10.1191/09612030168243023011678453YelinE, TrupinL, BundeJ, YazdanyJ (2019) Poverty, Neighborhoods, Persistent Stress, and Systemic Lupus Erythematosus Outcomes: A Qualitative Study of the Patients’ Perspective. Arthritis Care Res 71:398–405. 10.1002/acr.23599Abdul- SattarAB, MagdSAE (2017) Association of perceived neighborhood characteristics, socioeconomic status and rural residency with health outcomes in Egyptian patients with systemic lupus erythematosus: one center study. Int J Rheum Dis 20:2045–2052. 10.1111/1756-185X.1233124646073HowardMC, HoffmanME (2018) Variable-Centered, Person-Centered, and Person-Specific Approaches: Where Theory Meets the Method. Organ Res Methods 21:846–876. 10.1177/1094428117744021CollinsLM, LanzaST (2009) Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences. John Wiley & SonsLekkasP, StankovI, DanielM, PaquetC (2019) Finite mixture models in neighbourhoods-to-health research: A systematic review. Health Place 59:102140. 10.1016/j.healthplace.2019.05.01831374380HochbergMC (1997) Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 40:1725. 10.1002/1529-0131(199709)40:9<1725::AID-ART29>3.0.CO;2-YDrenkardC, RaskKJ, EasleyKA, (2013) Primary preventive services in patients with systemic lupus erythematosus: Study from a population-based sample in Southeast U.S. Semin Arthritis Rheum 43:209–216. 10.1016/j.semarthrit.2013.04.00323731530US Census Bureau (2019) American Community Survey 2014–2018 5-Year Estimates. US Census Bureau, Suitland, MDRobertoE (2016) The Divergence Index: A Decomposable Measure of Segregation and InequalityMasseyDS, DentonNA (1988) The Dimensions of Residential Segregation. Soc Forces 67:281–315. 10.1093/sf/67.2.281RossCE, JangSJ (2000) Neighborhood Disorder, Fear, and Mistrust: The Buffering Role of Social Ties with Neighbors. Am J Community Psychol 28:401–420. 10.1023/A:100513771333210965384PilkonisPA, ChoiSW, ReiseSP, (2011) Item Banks for Measuring Emotional Distress From the Patient-Reported Outcomes Measurement Information System (PROMIS®): Depression, Anxiety, and Anger. Assessment 18:263–283. 10.1177/107319111141166721697139PilkonisPA, YuL, DoddsNE, (2014) Validation of the depression item bank from the Patient-Reported Outcomes Measurement Information System (PROMIS®) in a three-month observational study. J Psychiatr Res 56:112–119. 10.1016/j.jpsychires.2014.05.01024931848ChoiSW, SchaletB, CookKF, CellaD (2014) Establishing a common metric for depressive symptoms: Linking the BDI-II, CES-D, and PHQ-9 to PROMIS Depression. Psychol Assess 26:513–527. 10.1037/a003576824548149ClarkeP (2008) When can group level clustering be ignored? Multilevel models versus single-level models with sparse data. J Epidemiol Community Health 62:752–758. 10.1136/jech.2007.06079818621963GrahamJW (2009) Missing Data Analysis: Making It Work in the Real World. Annu Rev Psychol 60:549–576. 10.1146/annurev.psych.58.110405.08553018652544MuthénL, MuthénB (2017) Mplus Version 8. Muthén & Muthén, Los Angeles, CAMorinAJS, MarshHW (2015) Disentangling Shape from Level Effects in Person-Centered Analyses: An Illustration Based on University Teachers’ Multidimensional Profiles of Effectiveness. Struct Equ Model Multidiscip J 22:39–59. 10.1080/10705511.2014.919825KlineRB (2011) Principles and practice of structural equation modeling, 3rd ed. Guilford Press, New YorkRileyAR (2018) Neighborhood Disadvantage, Residential Segregation, and Beyond—Lessons for Studying Structural Racism and Health. J Racial Ethn Health Disparities 5:357–365. 10.1007/s40615-017-0378-528573643BoothJM, TeixeiraS, ZuberiA, WallaceJM (2018) Barrios, ghettos, and residential racial composition: Examining the racial makeup of neighborhood profiles and their relationship to self-rated health. Soc Sci Res 69:19–33. 10.1016/j.ssresearch.2017.10.00229169532SampsonRJ, RaudenbushSW (2005) Neighborhood stigma and the perception of disorder. Focus 24:7–11FranziniL, CaughyMO, NettlesSM, O’CampoP (2008) Perceptions of disorder: Contributions of neighborhood characteristics to subjective perceptions of disorder. J Environ Psychol 28:83–93. 10.1016/j.jenvp.2007.08.003ChangVW, HillierAE, MehtaNK (2009) Neighborhood Racial Isolation, Disorder and Obesity. Soc Forces 87:2063–2092. 10.1353/sof.0.018820179775CutronaCE, WallaceG, WesnerKA (2006) Neighborhood Characteristics and Depression: An Examination of Stress Processes. Curr Dir Psychol Sci 15:188–192. 10.1111/j.1467-8721.2006.00433.x18185846WilliamsEM, BrunerL, AdkinsA, (2016) I too, am America: a review of research on systemic lupus erythematosus in African-Americans. Lupus Sci Med 3:e000144. 10.1136/lupus-2015-00014427651918PearlinLI, SchiemanS, FazioEM, MeersmanSC (2005) Stress, Health, and the Life Course: Some Conceptual Perspectives. J Health Soc Behav 46:205–219. 10.1177/00221465050460020616028458CarrF, NicassioP, IshimoriM, (2011) Depression predicts self-reported disease activity in systemic lupus erythematosus. Lupus 20:80–84. 10.1177/096120331037867220937622HeimanE, LimSS, BaoG, DrenkardC (2018) Depressive Symptoms Are Associated With Low Treatment Adherence in African American Individuals With Systemic Lupus Erythematosus: JCR J Clin Rheumatol 24:368–374. 10.1097/RHU.000000000000079429912774TonnerC, TrupinL, YazdanyJ, (2010) Role of community and individual characteristics in physician visits for persons with systemic lupus erythematosus. Arthritis Care Res 62:888–895. 10.1002/acr.20125TrupinL, TonnerMC, YazdanyJ, (2008) The role of neighborhood and individual socioeconomic status in outcomes of systemic lupus erythematosus. J Rheumatol 35:1782–178818634153DoDP, LocklarLRB, FlorsheimP (2019) Triple jeopardy: the joint impact of racial segregation and neighborhood poverty on the mental health of black Americans. Soc Psychiatry Psychiatr Epidemiol 54:533–541. 10.1007/s00127-019-01654-530671599HuntMO, WiseLA, JipguepM-C, (2007) Neighborhood Racial Composition and Perceptions of Racial Discrimination: Evidence From the Black Women’s Health Study. Soc Psychol Q 70:272–289. 10.1177/019027250707000306HutchinsonRN, PuttMA, DeanLT, (2009) Neighborhood racial composition, social capital and black all-cause mortality in Philadelphia. Soc Sci Med 1982 68:1859–1865. 10.1016/j.socscimed.2009.02.005EnglishD, LambertSF, EvansMK, ZondermanAB (2014) Neighborhood Racial Composition, Racial Discrimination, and Depressive Symptoms in African Americans. Am J Community Psychol 54:219–228. 10.1007/s10464-014-9666-y24969707PanaiteV, BowersoxNW, ZivinK, (2019) Individual and neighborhood characteristics as predictors of depression symptom response. Health Serv Res 54:586–591. 10.1111/1475-6773.1312730829408WilliamsDR, LawrenceJA, DavisBA (2019) Racism and Health: Evidence and Needed Research. Annu Rev Public Health 40:105–125. 10.1146/annurevpublhealth-040218-04375030601726SewellAA, FeldmanJM, RayR, (2020) Illness spillovers of lethal police violence: the significance of gendered marginalization. Ethn Racial Stud 1–26. 10.1080/01419870.2020.1781913MilesDRB, SamuelsB, PollackCE (2017) Leveraging Housing Vouchers to Address Health Disparities. Am J Public Health 107:238–240. 10.2105/AJPH.2016.30356528075648LeventhalT, Brooks-GunnJ (2003) Moving to Opportunity: an Experimental Study of Neighborhood Effects on Mental Health. Am J Public Health 93:1576–1582. 10.2105/AJPH.93.9.157612948983LindbergRA, ShenassaED, Acevedo-GarciaD, (2010) Housing Interventions at the Neighborhood Level and Health: A Review of the Evidence. J Public Health Manag Pract 16:S44–S52. 10.1097/PHH.0b013e3181dfbb7220689374ZafariZ, MuennigP (2020) The cost-effectiveness of limiting federal housing vouchers to use in low-poverty neighborhoods in the United States. Public Health 178:159–166. 10.1016/j.puhe.2019.08.01631698138LindbladMR, ManturukKR, QuerciaRG (2013) Sense of Community and Informal Social Control Among Lower Income Households: The Role of Homeownership and Collective Efficacy in Reducing Subjective Neighborhood Crime and Disorder. Am J Community Psychol 51:123–139. 10.1007/s10464-012-9507-922484395CascianoR, MasseyDS (2012) Neighborhood disorder and anxiety symptoms: New evidence from a quasi-experimental study. Health Place 18:180–190. 10.1016/j.healthplace.2011.09.00221963216Brooks HollidayS, TroxelW, HaasA, (2020) Do investments in low-income neighborhoods produce objective change in health-related neighborhood conditions? Health Place 64:102361. 10.1016/j.healthplace.2020.10236132838886TatianPA, KingsleyGT, ParillaJ, PendallR (2012) Building successful neighborhoods. Wash DC Urban InstGalsterG, TatianP, AccordinoJ (2006) Targeting Investments for Neighborhood Revitalization. J Am Plann Assoc 72:457–474. 10.1080/01944360608976766Dulin-KeitaA, ClayO, WhittakerS, (2015) The influence of HOPE VI neighborhood revitalization on neighborhood-based physical activity: A mixed-methods approach. Soc Sci Med 139:90–99. 10.1016/j.socscimed.2015.06.00226164364TwinamT (2017) Danger zone: Land use and the geography of neighborhood crime. J Urban Econ 100:104–119. 10.1016/j.jue.2017.05.006CarterSP, CarterSL, DannenbergAL (2003) Zoning Out Crime and Improving Community Health in Sarasota, Florida: “Crime Prevention Through Environmental Design.” Am J Public Health 93:1442–1445. 10.2105/AJPH.93.9.144212948960BaranyiG, SieberS, CullatiS, (2020) The Longitudinal Associations of Perceived Neighborhood Disorder and Lack of Social Cohesion With Depression Among Adults Aged 50 Years or Older: An Individual-Participant-Data Meta-Analysis From 16 High-Income Countries. Am J Epidemiol 189:343–353. 10.1093/aje/kwz20931573028CurryA, LatkinC, Davey-RothwellM (2008) Pathways to depression: The impact of neighborhood violent crime on inner-city residents in Baltimore, Maryland, USA. Soc Sci Med 67:23–30. 10.1016/j.socscimed.2008.03.00718396366

Standardized means of census tract characteristics for latent tract typologies in the BeWELL Study (n=257).

Structural equation model of standardized direct effects among census tract typologies, subjective neighborhood disorder, and depressive symptoms in the BeWELL Study (n=438).

Model Fit: RMSEA=0.02, CFI=0.98, SRMR=0.02. Notes: Estimates are adjusted for all sociodemographic, socioeconomic, and health-related covariates. Solid lines indicate significant association at p<.05.

1referent: Highly Segregated/Low-SES 2referent: Integrated/High-SES.

Descriptive Statistics and Bivariate Correlations of Main Study Variables in the BeWELL Study (Ntracts = 257) and Atlanta, GA Metropolitan Statistical Area (MSA) (Ntracts = 948)

BeWELLAtlanta MSA
MeanSDMeanSD123456789
Individual-Level (n=438)
1) Depression T-Score51.709.85---
2) Subjective Social Disorder1.730.75--0.14-
3) Subjective Physical Disorder1.610.62--0.120.70-
Tract-Level
4) Percent Poverty20.6712.2015.3411.390.060.300.35-
5) Percent ≤ High School12.857.6411.909.14−0.010.210.260.66-
6) Percent SNAP19.8611.7213.4111.230.020.330.360.830.65-
7) Percent Unemployed10.205.367.474.780.020.300.330.620.380.71-
8) Percent Black61.1528.7933.9129.99−0.010.310.330.370.130.550.61-
9) Divergence0.370.30--−0.020.310.340.360.120.540.590.89-

Note: bolded correlations are significant at p<.05

Model Fit Statistics for Latent Profile Analyses of Census Tract Characteristics in the BeWELL Study (n=257)

Profile SizeProfile ProportionsBICABICEntropyBLRT
11.0010005.59967.4--
2.52/.489301.39241.00.9560.000
3.46/.37/.169125.19042.60.9210.000
4.34/.29/.20/.178995.48890.80.9010.000
5.30/.23/.18/.14/.158909.08782.10.9140.000
6.23/.21/.18/.16/.14/.098851.88702.80.9100.000

Tract-Level and Participant Characteristics by Tract Typology in the BeWELL Study

Typology 1Typology 2Typology 3Typology 4
Integrated/ High-SESModerately Segregated/Mid-SESHighly Segregated/Mid-SESHighly Segregated/Low-SES
(Ntracts = 74)(Ntracts = 52)(Ntracts = 87)(Ntracts = 44)
(Nparticpants = 85)(Nparticpants = 66)(Nparticpants = 206)(Nparticpants = 81)
VariableMSDMSDMSDMSDF or χ2P
Tract-Level
 Percent Poverty9.09a4.7022.50b7.4520.02b7.9239.28c8.16168.59<.0001
 Percent ≤ High School6.81a4.0117.55b7.4210.94c3.8621.22d7.5379.18<.0001
 Percent SNAP7.09a3.9819.86b5.6721.47b6.2338.17c6.77281.41<.0001
 Percent Unemployed5.74a2.208.48b3.5811.64c3.5916.90d5.8289.27<.0001
 Percent Black29.90a15.9644.14b16.4485.99c8.7584.72c13.43298.57<.0001
 Divergence0.08a0.080.13a0.110.63b0.160.62b0.20291.42<.0001
Participant-Level
Depressive Symptoms T-Score*51.5010.5552.679.9351.119.2852.6310.520.710.5400
Perceived Physical Disorderl.23a0.431.50b0.671.66b0.561.98c0.6425.32<.0001
Perceived Social Disorder1.29a0.431.60b0.731.79b0.742.11C0.8220.3<.0001
Age (years)45.51ab12.0843.68b11.5148.65a12.1346.17ab13.093.400.0178
Disease Duration (years)14.989.7714.628.8616.8110.9115.8710.671.080.3556
Income-Poverty Ratio2.63a1.841.68b1.352.08bc1.751.42b1.268.47<.0001
Education (n, %)43.81<.0001
 < High School55.8869.09115.371518.52
 High School44.711319.704220.492020.69
 Some College3541.183350.009445.853543.21
 College Degree4148.241421.215828.291113.58
Work Status (n, %)18.930.0257
 Full-time3338.821624.245627.182024.69
 Part-time1011.761116.672612.6278.64
 Out of labor force67.0600.00157.2811.23
 Unable to work3642.353959.0910952.915365.43
Insurance Status (n, %)23.090.008
 Private4350.592030.307938.351518.52
 Public3743.534162.1210149.035466.67
 Uninsured55.8857.582612.621214.81
BMI28.837.2430.918.8931.048.4031.077.720.500.6843
Current Smoker (n, %)1214.29710.612713.171923.466.140.1049
Steroid Use (n, %)4856.473959.0911656.314150.621.210.7513
Hydroxychloroquine Use (n, %)6778.825177.2713666.026580.259.280.0258
Immunosuppressant Use (n, %)4047.062740.919345.153543.210.660.8828

Note: the F-statistic is presented for results of one-way ANOVAs for continuous dependent variables. The χ2 statistic is presented for results of chi-squared tests for categorical dependent variables. Means that do not share subscripts differ by p<.05.

T-scores are in reference to the general U.S. population, which has mean of 50 and standard deviation of 10.

Direct effects among tract typologies, subjective neighborhood disorder, and depression in the BeWELL Study (n=438).

Tract TypologyStandardized Path Coefficients
Subjective Neighborhood DisorderDepressive Symptoms
βSE95% CIβSE95% CI
Integrated/High-SES1−0.3320.040(−0.412, −0.254)0.0720.056(−0.033, 0.185)
Moderately Segregated/Mid-SES20.1840.053(0.088, 0.314)−0.0250.062(−0.142, 0.095)
Highly Segregated/Mid-SES20.4020.052(0.302, 0.504)−0.1050.072(−0.258, 0.029)
Highly Segregated/Low-SES20.4950.059(0.384, 0.624)−0.1000.064(−0.273, 0.036)

Note: Estimates adjusted for all sociodemographic, socioeconomic, and health-related covariates.

referent: Highly Segregated/Low-SES,

referent: Integrated/High-SES

Indirect effects among tract typologies, subjective neighborhood disorder, and depressive symptoms in the BeWELL Study (n=438).

Indirect PathsStandardized Path Coefficients
βSE95% CI
Integrated/High-SES1→Neighborhood Disorder→Depressive Symptoms−0.0400.020(−0.088, −0.007)
Moderately Segregated/Mid-SES2→Neighborhood Disorder →Depressive Symptoms0.0250.014(0.005, 0.063)
Highly Segregated/Mid-SES2→Neighborhood Disorder→Depressive Symptoms0.0540.026(0.012, 0.120)
Highly Segregated/Low-SES2→Neighborhood Disorder→Depressive Symptoms0.0660.032(0.013, 0.144)

Note: Estimates adjusted for all sociodemographic, socioeconomic, and health-related covariates.

referent: Highly Segregated/Low-SES,

referent: Integrated/High-SES

Highlights

Neighborhood disorder is a product of structural racism and residential segregation

Perceived neighborhood disorder was associated with greater depression

Structural racism and neighborhood factors may contribute to depression in lupus