Community-Level Social Support Infrastructure and Adult Onset of Major Depressive Disorder in a South Asian Postconflict Setting

William G Axinn; Karmel W Choi; Dirgha J Ghimire; Faith Cole; Sabrina Hermosilla; Corina Benjet; Melany C Morgenstern; Younga H Lee; Jordan W Smoller

doi:10.1001/jamapsychiatry.2021.4052

. 2022 Jan 26;79(3):1–7. doi: 10.1001/jamapsychiatry.2021.4052

Community-Level Social Support Infrastructure and Adult Onset of Major Depressive Disorder in a South Asian Postconflict Setting

William G Axinn ^1,^✉, Karmel W Choi ^2,^3,⁴, Dirgha J Ghimire ^1,⁵, Faith Cole ¹, Sabrina Hermosilla ¹, Corina Benjet ⁶, Melany C Morgenstern ¹, Younga H Lee ^2,^3,⁴, Jordan W Smoller ^2,^3,⁴

¹Population Studies Center, Institute for Social Research, University of Michigan, Ann Arbor

²Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston

³Psychiatric and Neurodevelopmental Genetics Unit, Center for Genomic Medicine, Massachusetts General Hospital, Boston

⁴Stanley Center for Psychiatric Research, Broad Institute, Boston, Massachusetts

⁵Institute for Social and Environmental Research–Nepal, Chitwan, Nepal

⁶Department of Epidemiology and Psychosocial Research, National Institute of Psychiatry Ramón de la Fuente Muñiz, Mexico City, Mexico

Accepted for Publication: November 24, 2021.

Published Online: January 26, 2022. doi:10.1001/jamapsychiatry.2021.4052

^✉

Corresponding Author: William G. Axinn, PhD, Institute for Social Research, University of Michigan, 426 Thompson St, Ann Arbor, MI 48104-2321 (baxinn@umich.edu).

Author Contributions: Dr Axinn had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Axinn, Choi, Ghimire, Cole, Benjet, Morgenstern, Lee.

Acquisition, analysis, or interpretation of data: Axinn, Choi, Ghimire, Cole, Hermosilla, Benjet, Morgenstern, Smoller.

Drafting of the manuscript: Axinn, Choi, Ghimire, Cole, Morgenstern, Lee.

Critical revision of the manuscript for important intellectual content: Axinn, Choi, Cole, Hermosilla, Benjet, Morgenstern, Lee, Smoller.

Statistical analysis: Axinn, Cole, Morgenstern.

Obtained funding: Axinn, Ghimire, Smoller.

Administrative, technical, or material support: Axinn, Hermosilla, Morgenstern, Lee.

Supervision: Axinn, Ghimire, Smoller.

Conflict of Interest Disclosures: Dr Choi reported receiving personal fees from Depression and Anxiety (Deputy Editor honorarium) outside the submitted work. Dr Ghimire reported receiving other support as director of the Institute for Social and Environmental Research–Nepal (ISER-N), which collected the data for the research reported here, and reported that his potential conflict of interest management plan is approved and monitored by the Regents of the University of Michigan. No other disclosures were reported.

Funding/Support: This work was supported by the National Institute of Mental Health (grant number R01MH110872). We also gratefully acknowledge use of the services and facilities of the Population Studies Center at the University of Michigan, funded by a National Institute of Child Health and Human Development (NICHD) Center Grant (grant number P2CHD041028).

Role of the Funder/Sponsor: The National Institute of Mental Health and NICHD 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.

Additional Contributions: We thank the survey staff of the Institute for Social and Environmental Research–Nepal for collecting the data reported here; the staff of the Survey Research Operations unit of the University of Michigan Survey Research Center for development and support of the technical systems that made the fieldwork in Nepal possible; the World Mental Health Consortium leadership and staff at Harvard University for their input into the design and all subsequent steps of collecting and analyzing the data reported here; and the respondents of the CVFS, whose generous contributions made this research possible. The authors alone remain responsible for any errors or omissions in this manuscript.

^✉

Corresponding author.

PMCID: PMC8792786 PMID: 35080609

This cohort study examines whether living in a neighborhood with social support groups is associated with the development of major depressive disorder among young adults in high-stress settings.

Key Points

Question

Is community-level social infrastructure associated with reduced incidence of major depressive disorder (MDD) among adults in high-risk settings?

Findings

In this cohort study of 1917 adults in 149 high-poverty neighborhoods that have recently experienced armed conflict, living in neighborhoods with available social programs was associated with subsequent reduced incidence of MDD.

Meaning

Because local social support infrastructure may reduce incidence of adult-onset MDD in settings of high exposure to potentially traumatic experiences, such infrastructure should be prioritized for population-level interventions.

Abstract

Importance

Individual-level social support protects against major depressive disorder (MDD) among adults exposed to trauma. Little is known about the consequences of community-level interventions in the general population.

Objective

To determine the potential consequences of neighborhood social infrastructure on incident MDD in a high-risk general population.

Design, Setting, and Participants

This longitudinal, multilevel study estimated associations between a neighborhood-level program in a case-control design and subsequent individual outcomes across 10 years (2006-2015) in a cohort of young adults. Exogenously placed social programs simulate natural experiment conditions in a high-poverty population experiencing armed conflict (1998-2006). The western Chitwan valley in Nepal has a general population at high risk of MDD, with neighborhoods exposed to interventions to improve social support. From a random sample (response rate 93%) selected to represent the general population in 2016, participants aged 25 to 34 years in 2006 were studied. These individuals resided within 149 neighborhoods that varied in their availability of active social support programs. The analyses were conducted between October 2020 and November 2021.

Exposures

The Small Farmers Development Program (SFDP) uses shared, joint liability financial credit among neighbors to build social capital and cohesion within neighborhoods.

Main Outcomes and Measures

Onset of DSM-IV MDD after the conflict, assessed by the Nepal-specific, clinically validated World Mental Health Composite International Diagnostic Interview with a life history calendar. The hypothesis tested was that exposure to SFDP reduced adult onset of MDD.

Results

Of the 1917 survey participants, 886 (46.2%) were women, and 856 (44.7%) were of Brahmin or Chhetri ethnicity. Of the 149 neighborhoods, 21 had an active SFDP group, and 156 of 1917 (8.1%) participants experienced MDD between 2006 and 2015. Discrete-time hazard models showed participants living in neighborhoods with an SFDP experienced incident MDD at nearly half the rate as others (odds ratio = 0.55; 95% CI, 0.30-1.02; P = .06). A multivariate, multilevel matching analysis showed the incidence of MDD among adults living in neighborhoods with an SFDP was 19 of 256 (7.4%), compared with 33 of 256 (12.9%) in the matched sample with no SFDP (z = 2.05; P = .04).

Conclusions and Relevance

Living in a neighborhood with community-level social support infrastructure was associated with reduced subsequent rates of adult-onset MDD, even in this high-risk population. Investments in such infrastructure may reduce population-level MDD, supporting clinical focus on potentially unpreventable cases.

Introduction

At the individual level, social support has been shown to reduce the risk of developing depression,¹ but we know little about the potential of exogenously created community-level infrastructure to promote social support among individuals. Identification of specific forms of community infrastructure that promote social support and decrease the risk of major depressive disorder (MDD) is a high priority for population-scale interventions to reduce both the incidence of MDD and the need for treatment. MDD is widespread in the general population and a leading cause of disability worldwide.² Moreover, community infrastructure to prevent MDD may be particularly relevant for communities recovering from conflict and crisis. Mitigating the need for treatment of cases that can be prevented with community-level interventions reduces the overall population-scale treatment gap and allows treatment services to focus on other unpreventable cases.

Social theory points to 2 powerful and complementary mechanisms of social support: social capital and social cohesion. Social capital refers to the emotional and instrumental resources that come from the structure of an individual’s relationships to others, including both family and nonfamily relationships.³ Under the right circumstances, just like human capital, social capital can be transformed into financial resources. The creation of local, nonfamily organizations provides an actionable opportunity to create new forms of social capital. The concept of social cohesion refers broadly to the quality and strength of supportive relationships between members of a group.^4,5 Prior research in a military cohort found that horizontal social cohesion (specifically, support and sense of solidarity between peers) may protect against the risk of developing depression after stressful experiences like deployment⁶ and that social cohesion may reduce the risk of depression even among people who are genetically at increased risk.⁷ However, these findings come from a special population: the US Army. The only evidence of a community-level social support intervention reducing depression risk also comes from a special population, elderly people, and does not account for other available community factors (eg, medical or educational resources) likely to promote health.⁸ General population evidence of the potential for community-level programs to reduce depression risk is a high priority for prevention at a population scale. Programs that stimulate horizontal social cohesion among groups of individuals have strong potential to provide a generalizable strategy for increasing social support at the community level.

High-stress settings are high-priority demonstration contexts for depression prevention through promotion of social support. Nepal is a setting with a general population exposed to high levels of stress, including widespread poverty and a 6-year, medium-intensity armed conflict,^9,10 with the majority of people experiencing 2 or more potentially traumatic experiences in their lifetime.¹¹ Otherwise, Nepal’s population is similar to agrarian populations throughout Asia.^12,13,14 Serendipitously, this setting also provides a natural experiment with exogenously introduced community-level cooperative programs that promote social support through both social capital and social cohesion. The Small Farmers Development Program (SFDP) was originally created by the Food and Agriculture Organization of the United Nations to benefit the lowest-income farmers. Nepalese officials chose SFDP sites from a large set of similar high-poverty neighborhoods, outside of the control of the local population. The SFDP features joint liability for agricultural loans as a substitute for collateral, creating social capital for obtaining financial resources. Once established, the SFDP provides a new context for neighbors to help each other, focusing on economic success.¹⁵

Although designed to promote social capital for economic success, prior evidence shows the high degree of horizontal social cohesion produced within the SFDP groups also yields multiple noneconomic consequences. The best documented example is the diffusion of health behavior changes through health education programs (eg, on contraceptive use) integrated into the SFDP, yielding significant changes in behavior.¹⁵ Although these consequences of SFDP group participation are documented, the potential for this community program to benefit mental health has never been evaluated.

We use unique new mental health data from a long-term panel study of communities to investigate the association between neighborhood (not individual) exposure to the SFDP and incidence of MDD. Exposure refers to living in neighborhoods with an active SFDP group. The 25-year Chitwan Valley Family Study (CVFS)^16,17 is a longitudinal, multilevel study conducted in Nepal that follows up whole populations of neighborhoods before, during, and after Nepal’s 6-year, medium-intensity armed conflict.¹⁰ The CVFS also measures whether or not each neighborhood has an active SFDP group across time (active means the group is holding at least 1 meeting every 6 months on any topic). We hypothesized that young adults living in neighborhoods with an active SFDP group would be less likely to experience MDD over the recovery years, after the armed conflict, than those living in neighborhoods with no SFDP groups, even after controlling for other relevant individual and neighborhood characteristics. Although not randomly assigned, the SFDP group placement was exogenously determined to include the lowest-resourced neighborhoods (with high-poverty households) within a setting of many similar neighborhoods. Adjoining neighborhoods with equally low resources were not selected, simulating a natural experiment (eMethods in the Supplement). Using this variability in location of SFDP groups, we also performed a novel multivariate, multilevel matching analytic strategy^18,19 to reduce sensitivity to unmeasured confounding.

Methods

Ethics and Consent

All procedures involving human participants were approved by the University of Michigan Health Sciences and Behavioral Sciences institutional review board and the Nepal Health Research Council. Written or verbal (witnessed and formally recorded) informed consent was obtained from all participants. Participation was voluntary, and participants did not receive compensation.

Study Design and Population

The CVFS launched in 1995 with a random selection, equal probability, population-representative sample of neighborhoods in western Chitwan, Nepal.^1,16 Data included in this analysis were collected from January 2015 through December 2018. Neighborhood characteristics, including SFDP in the neighborhood, were measured from 2006 to 2015 using observation and neighborhood history calendars.²⁰ Lifetime prevalence and onset of MDD were measured in 2016 to 2018 among all participants aged 15 to 59 years in CVFS neighborhoods (n = 10 623, response rate 93%²). We focused on the cohort of young adults, aged 25 to 34 years at the end of the armed conflict (in 2006, n = 2135): a cohort experiencing substantial disruption in their transitions to adulthood but responsible for the recovery of their households starting in 2006.¹⁰ Prior site-specific ethnography indicated this age group was at particular risk and most likely to benefit from social support in this setting.²¹ This produced a prospective analysis of incident MDD in a population-representative sample. Refer to the eMethods in the Supplement for more detail.

Measures

Neighborhood-Level Measures

The CVFS combines direct observation, focus group interviewing, and examination of written records to measure neighborhood-level events and characteristics in each of the CVFS neighborhoods over time (eMethods, eResults, and eFigure in the Supplement).^20,21 Innovative neighborhood history calendar methods measured neighborhood residents’ access to the nearest school and health service in walking time and recorded neighborhood presence of an active SFDP group, which changes annually.^12,13,20,22 The calendar indexed neighborhood-level availability (not individual engagement) of all services, including SFDP. We coded dichotomous, time-varying indicators of whether (1) or not (0) an SFDP was actively holding meetings in the neighborhood in each year from 2006 to 2015 and also whether (1) or not (0) a neighborhood had access to a school or health service within a 5-minute walk in that year.

Individual-Level Measures

Individual MDD was assessed with the Nepal-specific World Mental Health Composite International Diagnostic Instrument 3.0,²³ a clinically validated, structured diagnostic instrument.²³ This instrument was paired with a life history calendar to improve recall of disorders and timing of onset.²⁴ For the hazard of incident MDD, only person-years after the armed conflict (2006-2015) were analyzed, eliminating respondents with lifetime MDD onset before or during the conflict. The remaining respondents were coded with a 1 in the year they first met diagnostic criteria according to DSM-IV for MDD and 0 in all other years.

Our analysis estimated the total association between neighborhood exposure to the SFDP and individual-level MDD. We included age (years, time-varying), self-reported gender, and ethnicity, which capture key sources of individual variation established at birth. This approach allows variability in all other potential individual-level mechanisms connecting SFDP to MDD (eg, income, education, nutrition) within this otherwise low-resource population.¹⁶ Ethnicity was assessed by asking the full names of each individual in the household, which in Chitwan, Nepal, includes a detailed ethnicity–identifying family name. In analysis, these detailed ethnicities were then grouped into the following categories: Brahmin or Chhetri (high-caste Hindu individuals, reference category), Dalit (low-caste Hindu individuals), Hill Janajati, Newar, and Terai Janajati.¹⁶

Statistical Analyses

First, we conducted multilevel, discrete-time hazard analyses of MDD, with person-years as the unit of analysis, estimated with multivariable logistic regression (coefficients presented as odds ratios). We used a multilevel technique, initially designed for the CVFS and now widely used,²⁵ that adjusts for individuals clustered within neighborhoods. All models controlled for the sociodemographic measures described above, and we investigated estimates of associations with the presence of an SFDP group with and without adjustment for schools and health services. Having an active SFDP group in the neighborhood in year x is used to predict whether or not the individual transitions to MDD in year x + 1. All significance tests were evaluated with 2-sided tests (P values for asymptotic z ratios at P < .10 are indicated in the Tables). We performed data analysis using Stata version 16 between October 2020 and November 2021.

Second, to simulate an experimental design, we further used optimal multilevel matching methods created by Zubizarreta and colleagues.^26,27 The treatment-control matching process tested for the presence of a neighborhood-level treatment effect of an SFDP. Such matching methods are applied to observational data to approximate the results of a true randomized experiment. Beginning with a generally homogenous population represented in CVFS, the process used the full sample of 1917 to find 256 adults without exposure to the SFDP that match the 256 adults who did have exposure to an SFDP group.³ First, we used a form of cardinality matching,²⁶ which simultaneously mean-balances and/or exact-matches on key covariates (on both the neighborhood and individual level) in their original form, without needing to estimate propensity scores or any other summary measure of the covariates.²⁷ This process resulted in 2 matched samples (eTable 1 in the Supplement) of individuals who were marginally mean-balanced within 0.05 SD on 2 neighborhood-level dimensions (school or health service located within a 5-minute walk) and on gender (individual-level). Second, we exact-matched on cohort and a binary version of ethnicity (Brahmin/Chhetri and Newar [51% of the sample] or not) because of the key role this ethnic divide plays in shaping mental disorders within Nepal.¹⁶ Third, we used a 2-sample z test to test the association between the treatment condition (SFDP) and MDD incidence after the armed conflict.

Results

Neighborhood-Level Characteristics

The analytic sample (N = 1917) included participants from 149 neighborhoods that had a mean (SD) size of 90.8 (40.9) residents. Of the initial cohort, 218 young adults were excluded because they met diagnostic criteria for MDD before 2006. In 2006, only 21 of 149 (14.1%) neighborhoods had an SFDP group; 74 of 149 (49.7%) had a school and 56 of 149 (37.6%) had a health service within a 5-minute walk.

Sample Individual-Level Characteristics

The 1917 participants aged 25 to 34 years in 2006 contributed 18 448 person-years at risk of MDD after the armed conflict. Table 1 shows the individual descriptive characteristics of the sample. Participants reporting Brahmin or Chhetri ethnicity composed the largest ethnic group (856/1917, or 44.7% of the sample). Lifetime prevalence of adult-onset MDD at any point from 2006 to 2015 was 156 of 1917 (8.1%). In 2006, half of the participants lived in a neighborhood with a school (987/1917, 51.5%), and more than a third lived in a neighborhood with a health service (708/1917, 36.9%). However, only 226 of 1917 (11.8%) individuals lived in a neighborhood with an SFDP group. Associations of sociodemographic characteristics with MDD are described elsewhere.¹⁶

Table 1. Descriptive Sample Characteristics for Adults Aged 25 to 34 Years in 2006 (N = 1917)^a.

Characteristic	Participants, No. (%)			χ² Test	P value^c
Characteristic	Overall (N = 1917)	SFDP group in neighborhood (n = 226)^b	No SFDP group in neighborhood (n = 1691)^b	χ² Test	P value^c
Experienced major depressive disorder, 2006-2015	156 (8.14)	13 (5.75)	143 (8.46)	1.95	.16
Gender				1.12	.29
Female	886 (46.22)	97 (42.92)	789 (46.66)
Male	1031 (53.78)	129 (57.08)	902 (53.34)
Ethnicity
Brahmin or Chhetri	856 (44.65)	121 (53.54)	735 (43.47)	8.19	.004
Dalit	198 (10.33)	30 (13.27)	168 (9.93)	2.40	.12
Hill Janajati	413 (21.54)	36 (15.93)	377 (22.29)	4.78	.03
Newar	120 (6.26)	4 (1.77)	116 (6.86)	8.80	.003
Terai Janajati	330 (17.21)	35 (15.49)	295 (17.45)	0.54	.46
Neighborhood context for individuals in 2006
Lived in a neighborhood with a school within a 5-min walk	987 (51.49)	90 (39.82)	897 (53.05)	13.95	<.001
Lived in a neighborhood with a health service within a 5-min walk	708 (36.93)	132 (58.41)	576 (34.06)	50.72	<.001
Lived in a neighborhood with an SFDP group	226 (11.79)	226 (100)	0	NA	NA

Open in a new tab

Abbreviations: NA, not applicable; SFDP, Small Farmers Development Program.

^{^a}

The analytic sample includes all residents of study neighborhoods aged 25 to 34 years at the end of the armed conflict (2006). Individuals who met diagnostic criteria for major depressive disorder before 2006 are excluded from the analytic sample.

^{^b}

The existence of SFDP in Chitwan Valley Family Study neighborhoods changes over time. In the discrete-time hazard models, SFDP group in neighborhood was included as a time-varying variable. In this table, the values are from 2006.

^{^c}

The χ² test is for the difference between those living in neighborhoods with an SFDP group and those living in neighborhoods without an SFDP group in 2006. All P values are 2-sided.

Multilevel Associations With MDD

Table 2 shows 2 nested discrete-time hazard models. Model 1 includes school availability and health service availability within a 5-minute walk as neighborhood-level covariates, controlling for sociodemographic confounders, whereas model 2 adds SFDP in the neighborhood. All models were adjusted for individual-level covariates (gender, ethnicity, age, and age squared). Without adjusting for SFDP in the neighborhood (model 1), neither school nor health service within a 5-minute walk was significantly associated with MDD onset. By contrast, having an SFDP group in the neighborhood was significantly associated with reduced odds of MDD (model 2: 0.55; 95% CI, 0.30-1.02; P = .06). The magnitude of this estimate was also large, nearly cutting the odds of MDD in half. We confirm this estimate below, by applying a rigorous multilevel matching test.

Table 2. Multilevel Discrete-Time Hazard of First Diagnosis of Major Depressive Disorder Among Adults Aged 25 to 34 Years in 2006, Estimated With Multivariable Logistic Regression^a.

Variable	Adjusted odds ratio (95% CI)
Variable	Model 1	Model 2
Neighborhood-level measures^b
SFDP group in neighborhood	NA	0.55 (0.30 to 1.02)
P value	NA	.06
School within a 5-min walk	0.77 (0.53 to 1.12)	0.76 (0.52 to 1.11)
Health service within a 5-min walk	0.97 (0.66 to 1.42)	1.03 (0.70 to 1.53)
Gender
Male	1 [Reference]	1 [Reference]
Female	4.47 (3.12 to 6.40)	4.42 (3.09 to 6.32)
P value	<.001	<.001
Ethnicity
Brahmin or Chhetri	1 [Reference]	1 [Reference]
Dalit	2.19 (1.28 to 3.75)	2.18 (1.29 to 3.70)
P value	.004	.004
Hill Janajati	1.33 (0.86 to 2.05)	1.28 (0.82 to 2.01)
Newar	0.95 (0.41 to 2.19)	0.89 (0.39 to 2.07)
Terai Janajati	0.87 (0.51 to 1.48)	0.89 (0.53 to 1.49)
Age^b
Age	0.90 (0.53 to 1.52)	0.90 (0.53 to 1.53)
Age squared^c	1.00 (0.99 to 1.01)	1.00 (0.99 to 1.01)
Observations, person-years	18 448	18 448
Individuals, No.	1917	1917
−2 Log likelihood	1711.81	1707.19
AIC	1731.81	1729.19
BIC	1810.03	1815.24

Open in a new tab

Abbreviations: AIC, Akaike information criterion; BIC, Bayesian information criterion; NA, not applicable; SFDP, Small Farmers Development Program.

^{^a}

The analytic sample included all residents of study neighborhoods aged 25 to 34 years at the end of the armed conflict (2006). The time period analyzed was 2006 to 2015, and individuals started contributing person-years to the model at age 12 years. Individuals who met diagnostic criteria for major depressive disorder before 2006 were excluded from the analytic sample.

^{^b}

Neighborhood-level measures, age, and age squared were treated as time-varying covariates.

^{^c}

Age squared is a time-varying indicator of the decaying effect of age.

Other results were consistent with findings reported previously.¹⁶ Female gender was significantly associated with higher odds of MDD in both models, as was Dalit ethnicity (the most disadvantaged ethnic group).

Female gender was significantly associated with higher odds of MDD in both models (model 1: odds ratio = 4.47; 95% CI, 3.12 to 6.40; P < .001; model 2: odds ratio = 4.42; 95% CI, 3.09 to 6.32; P < .001), as was Dalit ethnicity, the most disadvantaged ethnic group (model 1: odds ratio = 2.19; 95% CI, 1.28 to 3.75; P = .004; model 2: odds ratio = 2.18; 95% CI, 1.29 to 3.70; P = .004). Unadjusted estimates for the association between neighborhood-level covariates and MDD onset are shown in Table 3. Gender-stratified estimates were neither substantively nor statistically different by gender (not shown in Tables). In addition, we estimated models among other age cohorts and from 1998 to 2006, the time period during the armed conflict (eResults, eTable 2, and eTable 3 in the Supplement).

Table 3. Unadjusted Multilevel Discrete-Time Hazard of First Diagnosis of Major Depressive Disorder Among Adults Aged 25 to 34 Years in 2006, Estimated With Logistic Regression^a.

Variable	Odds ratio (95% CI)
Variable	Model 1	Model 2	Model 3
Neighborhood-level measures^b
SFDP group in neighborhood	0.52 (0.29 to 0.92)^c	NA	NA
School within a 5-min walk	NA	0.78 (0.57 to 1.07)	NA
Health service within a 5-min walk	NA	NA	0.93 (0.67 to 1.29)
Observations, person-years	18 448	18 448	18 448
Individuals, No.	1917	1917	1917
−2 Log likelihood	1793.80	1797.40	1799.60
AIC	1797.74	1801.44	1803.60
BIC	1813.39	1817.08	1819.24

Open in a new tab

Abbreviations: AIC, Akaike information criterion; BIC, Bayesian information criterion; NA, not applicable; SFDP, Small Farmers Development Program.

^{^a}

^{^b}

Neighborhood-level measures were treated as time-varying covariates.

^{^c}

P = .02.

The multivariate, multilevel matching methods revealed consistent findings with the hazard models. There were 256 young adults living in neighborhoods that had an SFDP group and 256 who matched these young adults on neighborhood- and individual-level characteristics but did not live in neighborhoods with an SFDP group. The incidence of MDD among those living in a neighborhood with an SFDP group was 19 of 256 (7.4%) compared with 33 of 256 (12.9%) among the matched, unexposed sample (z = 2.05; P = .04). This alternative analytic approach, simulating random assignment, provides strong corroborating evidence that the presence of SFDP is associated with reduced incidence of subsequent MDD in this population.

Discussion

This longitudinal, population-representative study of young adults in a high-stress setting found that living in a neighborhood with community-level social support infrastructure (in the form of the SFDP) was associated with reduced subsequent development of MDD, even after controlling for individual and neighborhood characteristics. This is consistent with the hypothesis that exogenously introduced social support groups, delivering both social capital and social cohesion, can be protective against the onset of MDD. Our findings using standard time-varying hazard analysis were confirmed with innovative multivariate, multilevel matching to reduce unmeasured bias and approximate a randomized design with a treatment or no-treatment condition. Both analyses were performed using measures of MDD collected in a large-scale, rigorous, and clinically validated general population survey.²⁴ Both analyses pointed to large reductions in MDD among individuals exposed to neighborhood-level social support, showing nearly 50% lower incidence of MDD compared with their unexposed counterparts.

Our results indicate the potential that social support fostered at a community level can have in preventing incident depression. These results have direct implications for population-scale targeting of interventions to reduce the risks of psychiatric disorders. Prior studies have implicated social support and social cohesion as important protective factors against the development of depression,¹ even among people at higher risk,^7,8,28 but these studies have focused almost exclusively on individual-level social connections or, even when related to group-level social climate, have relied on aggregated perceptions.⁶ In contrast, our study suggests that investment in social support structures (which might be called the built social environment), which can be objectively observed at the community level, may also be an effective strategy to improve mental health outcomes in populations. It is also notable that the effects of these exogenously created social support groups were observed above and beyond other available community-level resources (ie, health services, schools). This suggests that the consequences of community cooperative groups are unique (for example, via building social cohesion and trust) and not simply attributable to socioeconomic or structural benefits. Investigation of the precise mechanisms was beyond the scope of the work reported here but does warrant further research.

Strengths and Limitations

Our study has several notable strengths. The multilevel, longitudinal CVFS provided an opportunity to investigate prospective consequences of social support over nearly a decade. In addition, we used clinically validated diagnostic outcomes with high response rates (93%). Furthermore, CVFS features direct observation of schools and health services, which are known to promote population health. This is a crucial advance over studies that use individual reports of social cohesion or ignore other key community characteristics.^8,29 Another strength includes confirmation of the findings using a novel approach to simulate an experimental design by matching on both individual and community characteristics to control for unmeasured confounding.

Our results should be interpreted in light of several limitations. First, because of retrospective assessment, the incidence of MDD is likely to be underreported and the reported ages at onset may be biased.³⁰ We attempted to reduce this bias by using the validated life history calendar approach to improve retrospective reporting of MDD and timing of onset.²⁴ Second, this examination of neighborhood social capital and cohesion focuses on only one type of social program. The SFDP is particularly likely to produce consequences for those living nearby because the program has proved so powerful with respect to both social capital and social cohesion,¹⁵ but other social support programs may also have important consequences for mitigating vulnerability to MDD. Third, we focus here on young adults, aged 25 to 34 years in 2006, who survived a 6-year armed conflict without experiencing MDD during the conflict itself. These people may be more representative of those who were relatively resilient, and findings may not apply to other cohorts who developed MDD earlier or who experienced armed conflict or other traumatic stressors at different points in their life course.³¹ However, more research is necessary to establish the breadth of the influences of other social support infrastructure on MDD. Fourth, this study did not test specific mechanisms (eg, instrumental support, interpersonal cohesion, or neighborhood-level resources) that may underlie the association between SFDP availability and MDD risk. Because of the breadth of potential consequences of the SFDP intervention, these specific mechanisms are a high priority for future investigation. Fifth, as with all observational research, there remains the possibility that some unmeasured factors shape the association between SFDP and MDD and should be investigated further. Sixth, this study was conducted in the general population of Nepal. Although this population represents an important demonstration, given that individuals are at high risk for MDD owing to high prevalence of multiple risk factors, performing epidemiological and trial-based tests of community-level social support interventions like SFDP in a broader range of settings is a high priority.

Conclusions

This large, multilevel, prospective panel provides the strongest evidence to date that introducing community-level forms of social support, designed to foster both social capital and social cohesion, may have the potential to reduce the adult onset of MDD above and beyond other community-level resources, even in a setting with many risk factors. This evidence suggests that population-level interventions prioritizing social support mechanisms may have a powerful effect on prevention efforts in population mental health.

Supplement.

eMethods

eResults

eFigure. Neighborhood-level continuous histories of exposure to support services (1953-2015)

eTable 1. Descriptive sample characteristics of the multilevel matching analysis

eTable 2. Multilevel discrete-time hazard of first diagnosis of major depressive disorder among each cohort, estimated with multivariable logistic regression

eTable 3. Multilevel discrete-time hazard of first diagnosis of major depressive disorder among each cohort, from 1998 to 2006, estimated with multivariable logistic regression

eReferences

Click here for additional data file.^{(365.4KB, pdf)}

References

1.Gariépy G, Honkaniemi H, Quesnel-Vallée A. Social support and protection from depression: systematic review of current findings in Western countries. Br J Psychiatry. 2016;209(4):284-293. doi: 10.1192/bjp.bp.115.169094 [DOI] [PubMed] [Google Scholar]
2.Global Burden of Disease 2017 Disease and Injury Incidence and Prevalence Collaborators . Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392(10159):1789-1858. doi: 10.1016/S0140-6736(18)32279-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Coleman JS. Foundations of Social Theory. Harvard University Press; 1990. [Google Scholar]
4.Friedkin NE. Social cohesion. Annu Rev Sociol. 2004;30:409-425. doi: 10.1146/annurev.soc.30.012703.110625 [DOI] [Google Scholar]
5.Kawachi I, Berkman LF. Social Capital, Social Cohesion, and Health. Oxford University Press; 2015. doi: 10.1093/med/9780195377903.003.0008 [DOI] [Google Scholar]
6.Campbell-Sills L, Flynn PJ, Choi KW, et al. Unit cohesion during deployment and post-deployment mental health: is cohesion an individual- or unit-level buffer for combat-exposed soldiers? Psychol Med. Published online June 10, 2020. doi: 10.1017/S0033291720001786 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Choi KW, Chen CY, Ursano RJ, et al. ; Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium . Prospective study of polygenic risk, protective factors, and incident depression following combat deployment in US Army soldiers. Psychol Med. 2020;50(5):737-745. doi: 10.1017/S0033291719000527 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Miao J, Wu X, Sun X. Neighborhood, social cohesion, and the elderly’s depression in Shanghai. Soc Sci Med. 2019;229:134-143. doi: 10.1016/j.socscimed.2018.08.022 [DOI] [PubMed] [Google Scholar]
9.Axinn WG, Ghimire D, Williams NE. Collecting survey data during armed conflict. J Off Stat. 2012;28(2):153-171. [PMC free article] [PubMed] [Google Scholar]
10.Williams NE, Ghimire DJ, Axinn WG, Jennings EA, Pradhan MS. A micro-level event-centered approach to investigating armed conflict and population responses. Demography. 2012;49(4):1521-1546. doi: 10.1007/s13524-012-0134-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Axinn WG, Chardoul S. Improving reports of health risks: life history calendars and measurement of potentially traumatic experiences. Int J Methods Psychiatr Res. 2021;30(1):e1853. doi: 10.1002/mpr.1853 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Axinn WG, Yabiku ST. Social change, the social organization of families, and fertility limitation. Am J Sociol. 2001;106(5):1219-1261. doi: 10.1086/320818 [DOI] [Google Scholar]
13.Brauner-Otto SR, Axinn WG, Ghimire DJ. The spread of health services and fertility transition. Demography. 2007;44(4):747-770. doi: 10.1353/dem.2007.0041 [DOI] [PubMed] [Google Scholar]
14.Ghimire DJ, Axinn WG, Yabiku ST, Thornton A. Social change, premarital nonfamily experience, and spouse choice in an arranged marriage society. AJS. 2006;111(4):1181-1218. doi: 10.1086/498468 [DOI] [Google Scholar]
15.Axinn WG. Rural income-generating programs and fertility limitation: evidence from a microdemographic study in Nepal. Rural Sociology. 1992;57(3):396-413. doi: 10.1111/j.1549-0831.1992.tb00472.x [DOI] [Google Scholar]
16.Scott KM, Zhang Y, Chardoul S, Ghimire DJ, Smoller JW, Axinn WG. Resilience to mental disorders in a low-income, non-Westernized setting. Psychol Med. Published online June 1, 2020. doi: 10.1017/S0033291720001464 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.University of Michigan . Welcome to the Chitwan Valley Family Study! Chitwan Valley Family Study. Accessed February 1, 2021. https://cvfs.isr.umich.edu/
18.de los Angeles Resa M, Zubizarreta JR. Direct and stable weight adjustment in non-experimental studies with multivalued treatments: analysis of the effect of an earthquake on post-traumatic stress. J R Stat Soc Series A Stat Soc. 2020;183(4):1387-1410. doi: 10.1111/rssa.12561 [DOI] [Google Scholar]
19.Zubizarreta JR, Keele L. Optimal multilevel matching in clustered observational studies: a case study of the effectiveness of private schools under a large-scale voucher system. J Am Stat Assoc. 2017;112(518):547-560. doi: 10.1080/01621459.2016.1240683 [DOI] [Google Scholar]
20.Axinn WG, Barber JS, Ghimire DJ. The neighborhood history calendar: a data collection method designed for dynamic multilevel modeling. Sociol Methodol. 1997;27(1):355-392. doi: 10.1111/1467-9531.271031 [DOI] [PubMed] [Google Scholar]
21.Axinn WG, Pearce LD. Mixed Method Data Collection Strategies. Cambridge University Press; 2006. doi: 10.1017/CBO9780511617898 [DOI] [Google Scholar]
22.Axinn WG, Ghimire DJ, Williams NE, Scott KM. Associations between the social organization of communities and psychiatric disorders in rural Asia. Soc Psychiatry Psychiatr Epidemiol. 2015;50(10):1537-1545. doi: 10.1007/s00127-015-1042-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Kessler RC, Ustün TB. The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Int J Methods Psychiatr Res. 2004;13(2):93-121. doi: 10.1002/mpr.168 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Axinn WG, Chardoul S, Gatny H, et al. Using life history calendars to improve measurement of lifetime experience with mental disorders. Psychol Med. 2020;50(3):515-522. doi: 10.1017/S0033291719000394 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Barber JS, Murphy SA, Axinn WG, Maples J. Discrete-time multilevel hazard analysis. Sociol Methodol. 2000;30(1):201-235. doi: 10.1111/0081-1750.00079 [DOI] [Google Scholar]
26.Zubizarreta JR, Paredes RD, Rosenbaum PR. Matching for balance, pairing for heterogeneity in an observational study of the effectiveness of for-profit and not-for-profit high schools in Chile. Ann Appl Stat. 2014;8(1):204-231. doi: 10.1214/13-AOAS713 [DOI] [Google Scholar]
27.de Los Angeles Resa M, Zubizarreta JR. Evaluation of subset matching methods and forms of covariate balance. Stat Med. 2016;35(27):4961-4979. doi: 10.1002/sim.7036 [DOI] [PubMed] [Google Scholar]
28.Choi KW, Stein MB, Nishimi KM, et al. ; 23andMe Research Team; Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium . An exposure-wide and Mendelian randomization approach to identifying modifiable factors for the prevention of depression. Am J Psychiatry. 2020;177(10):944-954. doi: 10.1176/appi.ajp.2020.19111158 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Lê F, Tracy M, Norris FH, Galea S. Displacement, county social cohesion, and depression after a large-scale traumatic event. Soc Psychiatry Psychiatr Epidemiol. 2013;48(11):1729-1741. doi: 10.1007/s00127-013-0698-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Wells JE, Horwood LJ. How accurate is recall of key symptoms of depression? a comparison of recall and longitudinal reports. Psychol Med. 2004;34(6):1001-1011. doi: 10.1017/S0033291703001843 [DOI] [PubMed] [Google Scholar]
31.Benjet C, Axinn WG, Hermosilla S, et al. Exposure to armed conflict in childhood vs older ages and subsequent onset of major depressive disorder. JAMA Netw Open. 2020;3(11):e2019848-e2019848. doi: 10.1001/jamanetworkopen.2020.19848 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Thornton A, Ghimire D, Young-DeMarco L, Bhandari P.. The reliability and stability of measures about individual’s values and beliefs concerning developmental idealism in Nepal. Sociol Dev. 2019;5:314-336. doi: 10.1525/sod.2019.5.3.314 [DOI] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplement.

eMethods

eResults

eFigure. Neighborhood-level continuous histories of exposure to support services (1953-2015)

eTable 1. Descriptive sample characteristics of the multilevel matching analysis

eTable 2. Multilevel discrete-time hazard of first diagnosis of major depressive disorder among each cohort, estimated with multivariable logistic regression

eTable 3. Multilevel discrete-time hazard of first diagnosis of major depressive disorder among each cohort, from 1998 to 2006, estimated with multivariable logistic regression

eReferences

Click here for additional data file.^{(365.4KB, pdf)}

[yoi210083r1] 1.Gariépy G, Honkaniemi H, Quesnel-Vallée A. Social support and protection from depression: systematic review of current findings in Western countries. Br J Psychiatry. 2016;209(4):284-293. doi: 10.1192/bjp.bp.115.169094 [DOI] [PubMed] [Google Scholar]

[yoi210083r2] 2.Global Burden of Disease 2017 Disease and Injury Incidence and Prevalence Collaborators . Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392(10159):1789-1858. doi: 10.1016/S0140-6736(18)32279-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r3] 3.Coleman JS. Foundations of Social Theory. Harvard University Press; 1990. [Google Scholar]

[yoi210083r4] 4.Friedkin NE. Social cohesion. Annu Rev Sociol. 2004;30:409-425. doi: 10.1146/annurev.soc.30.012703.110625 [DOI] [Google Scholar]

[yoi210083r5] 5.Kawachi I, Berkman LF. Social Capital, Social Cohesion, and Health. Oxford University Press; 2015. doi: 10.1093/med/9780195377903.003.0008 [DOI] [Google Scholar]

[yoi210083r6] 6.Campbell-Sills L, Flynn PJ, Choi KW, et al. Unit cohesion during deployment and post-deployment mental health: is cohesion an individual- or unit-level buffer for combat-exposed soldiers? Psychol Med. Published online June 10, 2020. doi: 10.1017/S0033291720001786 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r7] 7.Choi KW, Chen CY, Ursano RJ, et al. ; Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium . Prospective study of polygenic risk, protective factors, and incident depression following combat deployment in US Army soldiers. Psychol Med. 2020;50(5):737-745. doi: 10.1017/S0033291719000527 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r8] 8.Miao J, Wu X, Sun X. Neighborhood, social cohesion, and the elderly’s depression in Shanghai. Soc Sci Med. 2019;229:134-143. doi: 10.1016/j.socscimed.2018.08.022 [DOI] [PubMed] [Google Scholar]

[yoi210083r9] 9.Axinn WG, Ghimire D, Williams NE. Collecting survey data during armed conflict. J Off Stat. 2012;28(2):153-171. [PMC free article] [PubMed] [Google Scholar]

[yoi210083r10] 10.Williams NE, Ghimire DJ, Axinn WG, Jennings EA, Pradhan MS. A micro-level event-centered approach to investigating armed conflict and population responses. Demography. 2012;49(4):1521-1546. doi: 10.1007/s13524-012-0134-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r11] 11.Axinn WG, Chardoul S. Improving reports of health risks: life history calendars and measurement of potentially traumatic experiences. Int J Methods Psychiatr Res. 2021;30(1):e1853. doi: 10.1002/mpr.1853 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r12] 12.Axinn WG, Yabiku ST. Social change, the social organization of families, and fertility limitation. Am J Sociol. 2001;106(5):1219-1261. doi: 10.1086/320818 [DOI] [Google Scholar]

[yoi210083r13] 13.Brauner-Otto SR, Axinn WG, Ghimire DJ. The spread of health services and fertility transition. Demography. 2007;44(4):747-770. doi: 10.1353/dem.2007.0041 [DOI] [PubMed] [Google Scholar]

[yoi210083r14] 14.Ghimire DJ, Axinn WG, Yabiku ST, Thornton A. Social change, premarital nonfamily experience, and spouse choice in an arranged marriage society. AJS. 2006;111(4):1181-1218. doi: 10.1086/498468 [DOI] [Google Scholar]

[yoi210083r15] 15.Axinn WG. Rural income-generating programs and fertility limitation: evidence from a microdemographic study in Nepal. Rural Sociology. 1992;57(3):396-413. doi: 10.1111/j.1549-0831.1992.tb00472.x [DOI] [Google Scholar]

[yoi210083r16] 16.Scott KM, Zhang Y, Chardoul S, Ghimire DJ, Smoller JW, Axinn WG. Resilience to mental disorders in a low-income, non-Westernized setting. Psychol Med. Published online June 1, 2020. doi: 10.1017/S0033291720001464 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r17] 17.University of Michigan . Welcome to the Chitwan Valley Family Study! Chitwan Valley Family Study. Accessed February 1, 2021. https://cvfs.isr.umich.edu/

[yoi210083r18] 18.de los Angeles Resa M, Zubizarreta JR. Direct and stable weight adjustment in non-experimental studies with multivalued treatments: analysis of the effect of an earthquake on post-traumatic stress. J R Stat Soc Series A Stat Soc. 2020;183(4):1387-1410. doi: 10.1111/rssa.12561 [DOI] [Google Scholar]

[yoi210083r19] 19.Zubizarreta JR, Keele L. Optimal multilevel matching in clustered observational studies: a case study of the effectiveness of private schools under a large-scale voucher system. J Am Stat Assoc. 2017;112(518):547-560. doi: 10.1080/01621459.2016.1240683 [DOI] [Google Scholar]

[yoi210083r20] 20.Axinn WG, Barber JS, Ghimire DJ. The neighborhood history calendar: a data collection method designed for dynamic multilevel modeling. Sociol Methodol. 1997;27(1):355-392. doi: 10.1111/1467-9531.271031 [DOI] [PubMed] [Google Scholar]

[yoi210083r21] 21.Axinn WG, Pearce LD. Mixed Method Data Collection Strategies. Cambridge University Press; 2006. doi: 10.1017/CBO9780511617898 [DOI] [Google Scholar]

[yoi210083r22] 22.Axinn WG, Ghimire DJ, Williams NE, Scott KM. Associations between the social organization of communities and psychiatric disorders in rural Asia. Soc Psychiatry Psychiatr Epidemiol. 2015;50(10):1537-1545. doi: 10.1007/s00127-015-1042-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r23] 23.Kessler RC, Ustün TB. The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Int J Methods Psychiatr Res. 2004;13(2):93-121. doi: 10.1002/mpr.168 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r24] 24.Axinn WG, Chardoul S, Gatny H, et al. Using life history calendars to improve measurement of lifetime experience with mental disorders. Psychol Med. 2020;50(3):515-522. doi: 10.1017/S0033291719000394 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r25] 25.Barber JS, Murphy SA, Axinn WG, Maples J. Discrete-time multilevel hazard analysis. Sociol Methodol. 2000;30(1):201-235. doi: 10.1111/0081-1750.00079 [DOI] [Google Scholar]

[yoi210083r26] 26.Zubizarreta JR, Paredes RD, Rosenbaum PR. Matching for balance, pairing for heterogeneity in an observational study of the effectiveness of for-profit and not-for-profit high schools in Chile. Ann Appl Stat. 2014;8(1):204-231. doi: 10.1214/13-AOAS713 [DOI] [Google Scholar]

[yoi210083r27] 27.de Los Angeles Resa M, Zubizarreta JR. Evaluation of subset matching methods and forms of covariate balance. Stat Med. 2016;35(27):4961-4979. doi: 10.1002/sim.7036 [DOI] [PubMed] [Google Scholar]

[yoi210083r28] 28.Choi KW, Stein MB, Nishimi KM, et al. ; 23andMe Research Team; Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium . An exposure-wide and Mendelian randomization approach to identifying modifiable factors for the prevention of depression. Am J Psychiatry. 2020;177(10):944-954. doi: 10.1176/appi.ajp.2020.19111158 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r29] 29.Lê F, Tracy M, Norris FH, Galea S. Displacement, county social cohesion, and depression after a large-scale traumatic event. Soc Psychiatry Psychiatr Epidemiol. 2013;48(11):1729-1741. doi: 10.1007/s00127-013-0698-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r30] 30.Wells JE, Horwood LJ. How accurate is recall of key symptoms of depression? a comparison of recall and longitudinal reports. Psychol Med. 2004;34(6):1001-1011. doi: 10.1017/S0033291703001843 [DOI] [PubMed] [Google Scholar]

[yoi210083r31] 31.Benjet C, Axinn WG, Hermosilla S, et al. Exposure to armed conflict in childhood vs older ages and subsequent onset of major depressive disorder. JAMA Netw Open. 2020;3(11):e2019848-e2019848. doi: 10.1001/jamanetworkopen.2020.19848 [DOI] [PMC free article] [PubMed] [Google Scholar]

[yoi210083r32] 32.Thornton A, Ghimire D, Young-DeMarco L, Bhandari P.. The reliability and stability of measures about individual’s values and beliefs concerning developmental idealism in Nepal. Sociol Dev. 2019;5:314-336. doi: 10.1525/sod.2019.5.3.314 [DOI] [Google Scholar]

PERMALINK

Community-Level Social Support Infrastructure and Adult Onset of Major Depressive Disorder in a South Asian Postconflict Setting

William G Axinn, PhD

Karmel W Choi, PhD

Dirgha J Ghimire, PhD

Faith Cole, BA

Sabrina Hermosilla, PhD

Corina Benjet, PhD

Melany C Morgenstern, BA

Younga H Lee, PhD

Jordan W Smoller, MD, ScD

Key Points

Question

Findings

Meaning

Abstract

Importance

Objective

Design, Setting, and Participants

Exposures

Main Outcomes and Measures

Results

Conclusions and Relevance

Introduction

Methods

Ethics and Consent

Study Design and Population

Measures

Neighborhood-Level Measures

Individual-Level Measures

Statistical Analyses

Results

Neighborhood-Level Characteristics

Sample Individual-Level Characteristics

Table 1. Descriptive Sample Characteristics for Adults Aged 25 to 34 Years in 2006 (N = 1917)a.

Multilevel Associations With MDD

Table 2. Multilevel Discrete-Time Hazard of First Diagnosis of Major Depressive Disorder Among Adults Aged 25 to 34 Years in 2006, Estimated With Multivariable Logistic Regressiona.

Table 3. Unadjusted Multilevel Discrete-Time Hazard of First Diagnosis of Major Depressive Disorder Among Adults Aged 25 to 34 Years in 2006, Estimated With Logistic Regressiona.

Discussion

Strengths and Limitations

Conclusions

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Table 1. Descriptive Sample Characteristics for Adults Aged 25 to 34 Years in 2006 (N = 1917)^a.

Table 2. Multilevel Discrete-Time Hazard of First Diagnosis of Major Depressive Disorder Among Adults Aged 25 to 34 Years in 2006, Estimated With Multivariable Logistic Regression^a.

Table 3. Unadjusted Multilevel Discrete-Time Hazard of First Diagnosis of Major Depressive Disorder Among Adults Aged 25 to 34 Years in 2006, Estimated With Logistic Regression^a.