Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 1.
Published in final edited form as: Psychiatry Res. 2024 Feb 2;333:115766. doi: 10.1016/j.psychres.2024.115766

Neighborhood-Level Economic Characteristics and Depression and PTSD Symptoms Among Houstonians Who Have Experienced Hurricane Harvey and COVID-19

Gregory H Cohen 1,*, Ruochen Wang 1, Samuel B Rosenberg 1, Laura Sampson 2, Sarah R Lowe 3, Howard Cabral 4, Kenneth Ruggiero 5, Sandro Galea 6
PMCID: PMC10964477  NIHMSID: NIHMS1968122  PMID: 38335779

Abstract

Little is known about how neighborhood economic characteristics relate to risk of depression and Posttraumatic Stress Disorder (PTSD) in the context of multiple disasters. We sampled 88 super neighborhoods in Houston, Texas and surveyed 872 residents who were living in Houston during Hurricane Harvey and COVID-19 and lived in the same residence since Hurricane Harvey, about their demographics and symptoms of depression and PTSD. Using data from the American Community Survey, we estimated neighborhood-level unemployment, median income, and income inequality (i.e., Gini coefficient). We investigated whether these underlying neighborhood socioeconomic factors were associated with the mental health consequences of mass traumatic events. We examined associations between neighborhood-level constructs and individual-level depression and PTSD, using multilevel linear models. Partially adjusted multilevel models showed that lower neighborhood median income was associated with higher symptom scores of PTSD, while greater neighborhood income inequality was associated with higher symptom scores of depression and PTSD. However, fully adjusted models showed that these associations are better accounted for by event-specific stressors and traumas. These findings suggest that in the context of multiple large scale traumatic events, neighborhood socioeconomic context may structure individual-level exposure to stressful and traumatic events.

Keywords: Disasters, Depression, Posttraumatic Stress Disorder, Neighborhoods, Socioeconomic Factors, Income Inequality

1. INTRODUCTION

Houstonians, that is residents of Houston, Texas, USA, have experienced multiple disasters since 2017. First, Houston residents experienced Hurricane Harvey, which made landfall in Texas in late August 2017 as a category 4 hurricane and was one of the wettest and costliest hurricanes in US history. Hurricane Harvey resulted in approximately $125 billion in damages and displaced 40,000 flood victims who took refuge in shelters (Blake and Zelinsky, 2017; FEMA, 2018). Overall, the hurricane directly resulted in 68 deaths, with about half of these deaths occurring in the Houston metropolitan area (Blake and Zelinsky, 2017; FEMA, 2018). Second, Houstonians have experienced the COVID-19 pandemic, which was declared a global pandemic on March 11, 2020. As of February 23, 2023, the city of Houston has cumulatively experienced 540,914 confirmed COVID-19 cases and 4,806 COVID-19 related deaths (“Case Data | Harris County COVID-19 Data Hub prod2,” n.d.).

Depression and Posttraumatic Stress Disorder (PTSD) are 2 of the most common mental health conditions that follow disasters (Goldmann and Galea, 2014), and have been observed following both Hurricane Harvey (Bevilacqua et al., 2020; Fitzpatrick, 2021; Schwartz et al., 2018) and the emergence of COVID-19 (Abdalla et al., 2021; Ettman et al., 2020). These conditions are associated with a substantial burden of disease and disability (GBD 2019 Mental Disorders Collaborators, 2022), including years lived with disability, years of life lost, and unemployment. Population mental health, including depression and PTSD, are also sensitive to economic conditions and resources. In addition to increasing in the context of disasters, these mental health ailments also have been shown to emerge in the context of economic recessions and downturns such as that following the COVID-19 pandemic (Margerison-Zilko et al., 2016; Modrek et al., 2015).

Common measures of economic conditions and resources include unemployment and income, each of which has been linked to a wide range of health indicators, including mental health. Unemployment has been linked to depression across a variety of contexts (Amiri, 2022), and to persistence of PTSD following disasters (Nandi et al., 2004). Low income has also been linked to depression across a range of contexts (Costello et al., 2003; Ettman et al., 2020; Zimmerman and Katon, 2005), and financial strain has been linked to PTSD in the context of disaster exposure, including COVID-19 (Abdalla et al., 2021; Neria et al., 2008). Beyond individual-level factors, neighborhood context is an important predictor of mental health, as shown by prior research (Lowe et al., 2015). For example, income inequality – a neighborhood-level construct, has been found to be associated with both depression (Ahern and Galea, 2006; Patel et al., 2018) and PTSD (Pabayo et al., 2017).

While the connections between individual-level economic conditions and PTSD and depression are well understood generally, less is known about how these factors are associated with PTSD and depression at the neighborhood-level. Findings on the relationship between neighborhood-level socioeconomic conditions and depression have been mixed (Richardson, 2015), and there are few studies on neighborhood socioeconomic conditions and PTSD. In particular, there are no studies of neighborhood socioeconomic conditions and mental health among survivors of Hurricane Harvey in Houston. Moreover, it is not well understood how these economic factors affect depression and PTSD symptoms among people who have experienced multiple disasters. It is unclear, for example, whether these underlying neighborhood-level advantages may be associated with protection against the consequences of multiple mass traumatic events. Moreover, it is unclear whether identified associations between neighborhood factors and symptoms of depression and PTSD will remain after adjustment for individual-level stressful and traumatic events from each of these disasters. Houstonians who have experienced not only Hurricane Harvey, but also now COVID-19, represent a unique population with which to address these questions. Therefore, in this study, we examine associations between neighborhood-level unemployment, median income, and income inequality and individual-level symptoms of depression and PTSD.

2. METHODS

2.1. Study Sample

The Houston Trauma and Recovery Study (HTRS) is an observational survey study that used address-based sampling to randomly sample 88 Houston neighborhoods, referred to by the city of Houston as super-neighborhoods (City of Houston, n.d.). We sent recruitment letters to 12,009 addresses, 899 of which were not valid. A total of 11,110 letters were received by eligible participants, and 1266 people participated in the study for an overall response rate 11.4%. Study recruitment began in November 2020 and ended in August 2021. Of the 1266 who participated in the study, 1167 were living in Houston at the time of Hurricane Harvey, and among those, 872 participants were living at the same address as they were living at during Hurricane Harvey, constituting our final study sample for this analysis. This final sample allowed us to temporally match study exposures to participant neighborhoods of residence. Participants were allowed to fill out a mailed paper survey or complete the survey on a web-based platform. The survey was available in both English and Spanish. The institutional review board at Boston University Medical Campus approved of this study, and all study methods and practice were performed in accordance with the relevant guidelines and regulations. Informed consent was obtained from all subjects.

We created 2 sets of complex survey weights – (1) neighborhood-level weights that address probability of (a) random selection into the survey, (b) probability of survey response, (c) neighborhood size, and (d) number of adults per household; and (2) individual-level weights that take neighborhood-level weights and then rake (Deville et al., 1993) them to match the individual-level distributions of sex, race, ethnicity, age and educational attainment in the Houston Metro Area as of 2020 (Bureau, n.d.). Additionally, to account for missing data, we performed multivariable multiple imputation of 15 data sets, which were analyzed individually, and the results were then combined per Rubin’s Rules (Rubin, 2004).

2.2. Study Measures

Demographics.

Participants were asked several questions about their demographic characteristics. They were asked about their sex – male, female, or “other.” Participants were asked about their age in years, which we report as a continuous measure. They provided information on educational attainment which was coded dichotomously as high school or less versus some college or higher. Current marital status was asked, and participants were categorized as married, divorced/separated, or never married. Participants were separately asked about their ethnicity (Hispanic, non-Hispanic) and their race (White, Black, Asian, American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander), and were classed into the following race/ethnicity categories: White non-Hispanic, Black non-Hispanic, Other non-Hispanic (including multiracial participants), and Hispanic. Participants were asked about their income in 2019, (<$24,999, $25,000-$49,999, $50,000-$74,999, $75,000-$99,999, $100,000-$149,999, $150,000-$199,999, $200,000+) and current employment status (yes or no).

2.3. Primary Exposures

We captured ecological data for each super neighborhood after Harvey but before COVID-19 – by downloading the 5-year American Community Survey (ACS) (Bureau, n.d.) tables for 2016–2020 at the block group level, using the R v4.2.1 and the packages data.table (Dowle and Srinivasan, 2021), readxl (Wickham and Bryan, 2022), tidyverse (Wickham et al., 2019), tidycensus (Walker and Herman, 2023) and here (Müller, 2020).

Block groups do not align perfectly with super neighborhoods, so we created a two-step process in QGIS v3.22.4 to assign each block group to a super neighborhood. First, we removed block groups with less than 50% overlap with Houston proper. Second, we assigned each remaining block group to the super neighborhood intersecting with the largest proportional area of the block group. We subsequently aggregated block groups to the super neighborhood level.

Median Household Income.

Since the super neighborhood is not a unit of analysis published by the U.S. Census Bureau, we estimated the median income using block group-level data. Dividing the total number of households for each super neighborhood by two, we determined the income bracket below which half of households fell. More specifically, we downloaded block group-level ACS table B19001 for the period of 2016–2020, assigned block groups to each super neighborhood and aggregated the number of households by income range and super neighborhood. We determined the income range holding the mid-point income for each super neighborhood and found the proportion of the number of households in that income range needed to reach the mid-point. We multiplied this proportion by the width of the income range and added this to the lower end of the range to estimate the median household income for each super neighborhood.

Unemployment Rate.

We downloaded ACS table B23025 to calculate the unemployment rate over the period of 2016–2020. We divided the unemployed civilian labor force by the total civilian labor force and expressed this rate as a percentage for each super neighborhood.

Income Inequality.

We downloaded ACS table B19001 and the R package “DescTools” to calculate the Gini coefficient for each super neighborhood over the period 2016–2020. The Gini coefficient is calculated from the Lorenz curve, which plots the distribution of income cumulatively by percentage of population. The Gini coefficient is calculated by taking the difference between a given Lorenz Curve and a perfectly equal distribution of income (Bureau, 2021; De Maio, 2007). The Gini coefficient is expressed as a number between 0 and 1, with 0 representing perfect equality and 1 representing perfect inequality.

Harvey Stressors and Traumas.

To measure stressors related to Hurricane Harvey, we asked participants whether they experienced 10 storm-related events listed in table S1, including for example experiencing strong winds and heavy rain, damage to one’s home, and losing a job or getting laid off. In order to measure traumas related to Hurricane Harvey, we asked participants whether they experienced 5 storm-related traumas listed in table S2, and including being seriously injured, or having a friend or family member be killed as a result of the storm. Harvey stressors and traumas were included in our analyses as separate counts of stressful and traumatic events, respectively.

COVID-19 Stressors and Traumas.

We measured COVID-19 related stressors by asking whether participants had experienced 5 COVID-19 related events listed in table S3, including losing one’s job, having to move, and having trouble paying bills. COVID-19 related traumas, listed in table S4, include 6 traumas, including becoming seriously ill from COVID-19 or having a family or friend become seriously ill or dying from COVID-19. COVID-19 stressors and traumas were included in our analyses as separate counts of stressful and traumatic events, respectively.

2.4. Primary Outcomes

Depression Symptoms.

The PHQ-9 (Kroenke et al., 2010, 2001) was used to assess depression in the past 2 weeks among study participants at the individual-level. Each of the nine questions was scored as 0 (not at all), 1 (several days), 2 (more than half the days) or 3 (nearly every day), with total scores ranging from 0 to 27. Cronbach’s alpha is a measure of internal consistency and is strong for the PHQ-9, with values of between 0.89–0.86 in a major validation study (Kroenke et al., 2001), and a value of 0.91 in our study.

PTSD Symptoms.

We used the PTSD Checklist for DSM-5 (PCL-5) (Bovin et al., 2016) to evaluate past-month PTSD symptoms (APA, 2013) at the individual-level, which were not queried in reference to a specific event. Each of the 20 symptom questions could be rated as 0 (not at all), 1 (a little bit), 2 (moderately), 3 (quite a bit), or 4 (extremely). Taken together, total scores could fall between 0 and 80. According to validation studies, Cronbach’s alpha for the PCL-5 is very strong, with values of 0.94 (Morrison et al., 2021) and 0.96 (Bovin et al., 2016) in separate validation studies, and a value of 0.94 in our study.

2.5. Analytic Approach

We first calculated summary statistics and estimated bivariate associations between individual-level demographic characteristics using individual-level weights and neighborhood-level economic factors and current depression and PTSD using neighborhood-level weights (Table 1). Second, we examined the adjusted associations between neighborhood-level median household income (over the period of 2016–2020) and current depression and PTSD (Tables 2a, 2b). Third, we calculated the adjusted associations between neighborhood-level unemployment rate (over the period of 2016–2020) and current depression and PTSD (Tables 3a, 3b). Fourth, we estimated the adjusted associations between neighborhood-level income inequality (over the period of 2016–2020) and current depression and PTSD (Tables 4a, 4b). All adjusted associations were analyzed using multilevel linear regression models that applied neighborhood-level weights, with the intercept, demographic characteristics, and event-specific stressors and traumas treated as random effects. All analyses were conducted with SAS 9.4 (Institute, 2023), and p-values less than 0.05 were considered statistically significant. Fifth, we reported the distribution of individual-level income within each quartile of neighborhood-level median household income, and income inequality, respectively (Supplementary Tables 5 and 6), as a check on the degree of variation in individual-level income within neighborhood-level exposure levels. For example, we sought to demonstrate that there was a range of individual-level income across quartiles of median neighborhood-level income.

Table 1:

Bivariable associations between individual- and neighborhood-level factors and individual-level current depression or current PTSD among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category N (%) / Mean (SD) Current Depression Symptom Score (Range: 0–27) Current PTSD Symptom Score (Range: 0–80)
Estimate SE P Estimate SE P
Individual-level
Age 18–99 49.5 (17.05) −0.06 0.01 <.001 −0.11 0.03 0.002
Sex Male 315 (50.5) Ref. - - Ref. - -
Female 549 (47.8) 1.39 0.60 0.021 3.87 1.58 0.014
Other 8 (1.7) 9.46 2.19 <.001 13.05 6.35 0.043
Race/Ethnicity White non-Hispanic 327 (38.3) Ref. - - Ref. - -
Black non-Hispanic 225 (12.7) 3.02 0.98 0.002 10.89 2.62 <.001
Other non-Hispanic 70 (5.7) 0.02 1.67 0.989 2.89 4.04 0.479
Hispanic 250 (43.3) 2.05 0.75 0.007 5.39 2.08 0.010
Education High school or less 234 (49.8) 1.01 0.57 0.077 3.51 1.74 0.045
Some college or higher 638 (50.3) Ref. - - Ref. - -
Marital Status Divorced / Separated 263 (15.1) 0.16 0.74 0.831 2.53 2.10 0.228
Never Married 208 (29.5) 3.33 0.61 <.001 6.21 1.78 <.001
Married 402 (55.3) Ref. - - Ref. - -
Income (2019) <$25,000 183 (22.3) 3.43 0.98 0.001 10.54 2.59 <.001
$25,000–$49,999 194 (21.7) 1.94 1.04 0.063 5.65 2.69 0.036
$50,000–$74,999 136 (15.3) 0.57 0.99 0.562 1.98 2.61 0.448
$75,000–$99,999 80 (8.6) 1.70 1.15 0.140 2.50 3.02 0.409
$100,000–$149,999 86 (9.5) 0.38 1.12 0.735 1.24 2.92 0.672
$150,000–$199,999 79 (8.9) 0.87 1.33 0.514 2.78 3.72 0.458
$200,000+ 114 (13.6) Ref. - - Ref. - -
Employment Status Yes 434 (54.2) Ref. - - Ref. - -
No 438 (45.8) 1.11 0.59 0.062 2.96 1.51 0.051
Sum of Harvey Stressors 0–10 2.8 (0.11) 0.75 0.12 <.001 2.66 0.34 <.001
Sum of Harvey Traumas 0–5 1.5 (0.05) 1.31 0.26 <.001 3.68 0.70 <.001
Sum of COVID Stressors 0–5 0.7 (0.05) 2.08 0.26 <.001 5.00 0.74 <.001
Sum of COVID Traumas 0–6 0.8 (0.06) 0.99 0.27 <.001 3.22 0.69 <.001
Neighborhood-level
Median Household Income (2016–2020) Q1 29209.0 (3196.31) 1.35 0.71 0.055 6.05 1.54 <.001
Q2 39711.5 (3611.54) 1.62 0.61 0.008 4.36 1.16 <.001
Q3 53320.2 (3468.90) −0.01 0.58 0.991 2.12 1.13 0.060
Q4 75285.2 (11647.84) Ref. - - Ref. - -
Unemployment rate (2016–2020) Q1 3.5 (0.80) Ref. - - Ref. - -
Q2 5.5 (0.59) 0.03 0.58 0.957 0.40 1.17 0.733
Q3 7.5 (0.75) 0.26 0.63 0.683 2.81 1.32 0.034
Q4 12.2 (2.02) 1.51 0.84 0.074 6.02 2.04 0.003
Gini Coefficient (2016–2020) Q1 0.36 (0.02) Ref. - - Ref. - -
Q2 0.40 (0.01) 1.61 0.59 0.006 2.36 1.19 0.048
Q3 0.43 (0.01) 1.30 0.51 0.012 2.87 1.05 0.006
Q4 0.48 (0.02) 2.54 0.69 <.001 7.69 1.47 <.001
Open in a new tab
*

Frequencies are unweighted; percentages, means and bivariable models are weighted

Table 2a:

Multivariable associations between neighborhood-level household income and individual-level current depression symptom scores among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category Current Depression Symptom Score
Estimate SE P Estimate SE P Estimate SE P Estimate SE P
Neighborhood-level Median Household Income (2016–2020) Model 1 Model 2 Model 3 Model 4
Q1 0.97 0.93 0.295 −1.33 1.21 0.272 0.11 0.92 0.901 −1.40 1.14 0.220
Q2 0.99 0.85 0.245 −1.14 1.12 0.310 0.55 0.82 0.504 −0.65 1.03 0.526
Q3 −0.54 0.80 0.495 −1.48 1.07 0.168 −0.91 0.77 0.238 −1.64 0.97 0.091
Q4 Ref. - - Ref. - - Ref. - - Ref. - -
Sum of Harvey Stressors 0–10 0.44 0.14 0.002 0.32 0.13 0.014
Sum of Harvey Traumas 0–5 0.92 0.22 <.001 0.55 0.19 0.005
Sum of COVID Stressors 0–5 1.76 0.31 <.001 1.50 0.30 <.001
Sum of COVID Traumas 0–6 0.78 0.22 <.001 0.71 0.21 <.001
Open in a new tab
*

Multivariable estimates adjusted for individual-level age, sex, race/ethnicity, education, marital status, and income

Table 2b:

Multivariable associations between neighborhood-level household income and individual-level current PTSD symptom scores among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category Current PTSD Symptom Score
Estimate SE P Estimate SE P Estimate SE P Estimate SE P
Neighborhood-level Median Household Income (2016–2020) Model 1 Model 2 Model 3 Model 4
Q1 5.72 2.57 0.026 1.72 3.25 0.597 3.12 2.43 0.199 0.15 2.81 0.958
Q2 3.85 2.42 0.112 −0.15 3.14 0.961 3.13 2.20 0.155 1.09 2.59 0.674
Q3 0.02 2.30 0.993 −3.37 3.04 0.268 −0.30 2.12 0.888 −2.83 2.55 0.267
Q4 Ref. - - Ref. - - Ref. - - Ref. - -
Sum of Harvey Stressors 0–10 1.84 0.34 <.001 1.55 0.31 <.001
Sum of Harvey Traumas 0–5 2.19 0.55 <.001 0.85 0.52 0.101
Sum of COVID Stressors 0–5 4.18 0.90 <.001 3.32 0.87 <.001
Sum of COVID Traumas 0–6 2.36 0.59 <.001 2.45 0.55 <.001
Open in a new tab
*

Multivariable estimates adjusted for individual-level age, sex, race/ethnicity, education, marital status, and income

Table 3a:

Multivariable associations between neighborhood-level unemployment rate and individual-level current depression symptom scores among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category Current Depression Symptom Score
Estimate SE P Estimate SE P Estimate SE P Estimate SE P
Neighborhood-level Unemployment Rate (2016–2020) Model 1 Model 2 Model 3 Model 4
Q1 Ref. - - Ref. - - Ref. - - Ref. - -
Q2 0.17 0.71 0.807 −0.37 0.87 0.668 0.13 0.76 0.867 −0.18 0.92 0.847
Q3 −0.12 0.75 0.868 −0.78 0.89 0.384 −0.40 0.80 0.616 −0.67 0.94 0.478
Q4 0.37 0.98 0.703 −1.00 1.14 0.380 −0.16 1.01 0.876 −0.79 1.15 0.490
Sum of Harvey Stressors 0–10 0.41 0.13 0.002 0.32 0.12 0.009
Sum of Harvey Traumas 0–5 0.87 0.22 <.001 0.38 0.20 0.063
Sum of COVID Stressors 0–5 1.91 0.33 <.001 1.71 0.34 <.001
Sum of COVID Traumas 0–6 0.82 0.22 <.001 0.69 0.21 0.001
Open in a new tab
*

Multivariable estimates adjusted for individual-level age, sex, race/ethnicity, education, marital status, and employment status

Table 3b:

Multivariable associations between neighborhood-level unemployment rate and individual-level current PTSD symptom scores among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category Current PTSD Symptom Score
Estimate SE P Estimate SE P Estimate SE P Estimate SE P
Neighborhood-level Unemployment Rate (2016–2020) Model 1 Model 2 Model 3 Model 4
Q1 Ref. - - Ref. - - Ref. - - Ref. - -
Q2 −0.28 1.98 0.889 −3.14 2.37 0.186 0.48 1.93 0.802 −0.56 2.11 0.791
Q3 1.61 2.07 0.435 −0.80 2.38 0.739 0.83 2.03 0.681 0.09 2.20 0.966
Q4 1.97 2.67 0.460 −1.78 3.04 0.558 1.85 2.58 0.475 0.30 2.72 0.911
Sum of Harvey Stressors 0–10 1.67 0.33 <.001 1.31 0.28 <.001
Sum of Harvey Traumas 0–5 2.36 0.65 <.001 0.85 0.52 0.101
Sum of COVID Stressors 0–5 4.67 0.91 <.001 3.94 0.91 <.001
Sum of COVID Traumas 0–6 2.67 0.61 <.001 2.44 0.57 <.001
Open in a new tab
*

Multivariable estimates adjusted for individual-level age, sex, race/ethnicity, education, marital status, and employment status

Table 4a:

Multivariable associations between neighborhood-level Gini Coefficient and individual-level current depression symptom scores among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category Current Depression Symptom Score
Estimate SE P Estimate SE P Estimate SE P Estimate SE P
Neighborhood-level Gini Coefficient (2016–2020) Model 1 Model 2 Model 3 Model 4
Q1 Ref. - - Ref. - - Ref. - - Ref. - -
Q2 1.38 0.74 0.062 0.74 0.96 0.445 0.97 0.83 0.245 0.91 1.02 0.374
Q3 1.25 0.66 0.057 0.71 0.89 0.422 1.07 0.77 0.166 0.70 0.97 0.472
Q4 1.76 0.83 0.035 0.17 1.05 0.870 0.79 0.93 0.393 0.24 1.11 0.829
Sum of Harvey Stressors 0–10 0.43 0.13 0.001 0.31 0.12 0.010
Sum of Harvey Traumas 0–5 0.83 0.22 <.001 0.36 0.21 0.083
Sum of COVID Stressors 0–5 1.96 0.33 <.001 1.76 0.34 <.001
Sum of COVID Traumas 0–6 0.82 0.21 <.001 0.69 0.21 0.001
Open in a new tab
*

Multivariable estimates adjusted for individual-level age, sex, race/ethnicity, education, and marital status

Table 4b:

Multivariable associations between neighborhood-level Gini Coefficient and individual-level current PTSD symptom scores among Houston adults who lived at same address following Hurricane Harvey (n=872) *

Exposure Category Current PTSD Symptom Score
Estimate SE P Estimate SE P Estimate SE P Estimate SE P
Neighborhood-level Gini Coefficient (2016–2020) Model 1 Model 2 Model 3 Model 4
Q1 Ref. - - Ref. - - Ref. - - Ref. - -
Q2 2.37 1.91 0.215 2.33 2.42 0.335 1.14 2.01 0.570 2.05 2.23 0.357
Q3 3.40 1.70 0.046 1.80 2.23 0.419 3.52 1.87 0.059 2.63 2.10 0.209
Q4 6.29 2.14 0.003 4.24 2.67 0.112 3.83 2.25 0.089 3.97 2.44 0.104
Sum of Harvey Stressors 0–10 1.66 0.33 <.001 1.30 0.28 <.001
Sum of Harvey Traumas 0–5 2.27 0.63 <.001 0.89 0.51 0.081
Sum of COVID Stressors 0–5 4.58 0.91 <.001 3.85 0.90 <.001
Sum of COVID Traumas 0–6 2.67 0.61 <.001 2.44 0.57 <.001
Open in a new tab
*

Multivariable estimates adjusted for individual-level age, sex, race/ethnicity, education, and marital status

3. RESULTS

Sample Characteristics.

Table 1 shows that the mean age of our study participants was 49.5 years, and that participants were nearly evenly split between male (50.5%) and female sex (47.8%), with 1.7% identifying as other sex. Our sample was predominately of Hispanic ethnicity (43.3%), followed by White non-Hispanic (38.3%), with the smallest percentage of participants identifying as Black non-Hispanic (12.7%), and other non-Hispanic (5.7%). Educational attainment was about evenly split between those with high school or less (49.8%) and some college or higher (50.3%). Participants were predominately married (55.3%), followed by never married (29.5%), and divorced/separated (15.1%). A preponderance of participants reported income below $50,000 (44.1%), while 15.3% reported earning $50,000-$74,999, and 13.6% reported earning $200,000 or more. Most participants reported being employed (54.2%). The mean number of Harvey stressors was 2.8, while the mean number of Harvey traumas was 1.5; mean number of COVID-19 stressors was 0.7, while the mean number of COVID-19 traumas was 0.8.

Neighborhood Exposure Characteristics.

Table 1 shows that neighborhood-level median household income ranged from $29,209 (Q1) to $75,285 (Q4); the neighborhood-level unemployment rate ranged from 3.5% (Q1) to 12.2% (Q4); and the neighborhood-level Gini Coefficient ranged from 0.36 (Q1) to 0.48 (Q4).

Bivariable Associations Between Individual-Level Demographics and Depression Score.

Table 1 shows that each unit increase in year of age was associated with a 0.06 point decrease (p<0.001) on the PHQ-9. Compared to male gender, female gender was associated with 1.39 additional points (p=0.021) on the PHQ-9, while other gender was associated with 9.46 additional points (p<0.001) on the PHQ-9. Compared to White non-Hispanic race-ethnicity, being of Black non-Hispanic race-ethnicity was associated with 3.02 additional points (p=0.002) on the PHQ-9 and being of Hispanic ethnicity was associated with 2.05 additional points (p=0.007). Compared to those who were married, being never married was associated with 3.33 additional points (p<0.001) on the PHQ-9. Compared to annual income of $200,000+, annual income of <$25,000 was associated with 3.4 additional PHQ-9 points (p=0.001), while annual income of $25,000-$49,999 was associated with an additional 1.94 PHQ-9 points (p=0.063). Regarding employment status, compared to those who were employed, being unemployed was associated with an additional 1.11 points on the PHQ-9 (p=0.062). Each additional Harvey stressor was associated with a 0.75 point increase on the PHQ-9 (p<0.001), while each Harvey trauma was associated with a 1.31 point increase on the PHQ-9 (p<0.001). Each COVID-19 stressor was associated with a 2.08 point increase on the PHQ-9 (p<0.001), while each COVID-19 trauma was associated with a 0.99 point increase on the PHQ-9 (p<0.001).

Bivariable Associations Between Neighborhood-Level Exposures and Depression Score.

Table 1 shows that for median household income, compared to quartile 4, those in quartile 2 had a PHQ-9 score 1.62 points higher (p=0.008), while those in quartile 1 had a score 1.35 points higher (p=0.055). Unemployment rate was not associated with depression symptoms. For Gini coefficient, compared to those in quartile 1, those in quartile 2 had an additional 1.61 points (p=0.006) on the PHQ-9, while those in quartile 3 had an additional 1.30 points (p=0.012), and those in quartile 4 had an additional 2.54 points (p<0.001).

Bivariable Associations Between Individual-Level Demographics and PTSD Score.

Table 1 shows that each unit increase in age was associated with 0.11 points decrease (p=0.002) on the PCL-5. Compared to male gender, female gender was associated with an increase in PCL-5 score of 3.87 points (p=0.014), while other gender was associated with an increase in PCL-5 score of 13.05 points (p=0.043). Examining race-ethnicity, compared to White non-Hispanic ethnicity, being of Black non-Hispanic ethnicity was associated with 10.89 (p<0.001) additional PCL-5 score points, while being of Hispanic ethnicity was associated with 5.39 additional PCL-5 score points (p=0.010). Compared to those with some college or higher, having an educational attainment of high school or less was associated with 3.51 additional PCL-5 score points (p=0.045). Finally, compared to those who were married, being never married was associated with 6.21 additional PCL-5 Score points (p<0.001). Compared to annual income of $200,000+, annual income of <$25,000 was associated with 10.54 additional PCL-5 points (p<0.001), while annual income of $25,000-$49,999 was associated with an additional 5.65 PCL-5 points (p=0.036). Regarding employment status, compared to those who were employed, being unemployed was associated with an additional 2.96 points on the PCL-5 (p=0.051). Each additional Harvey stressor was associated with a 2.66 point increase on the PCL-5 (p<0.001), while each Harvey trauma was associated with a 3.68 point increase on the PCL-5 (p<0.001). Each COVID-19 stressor was associated with a 5 point increase on the PCL-5 (p<0.001), while each COVID-19 trauma was associated with a 3.22 point increase on the PCL-5 (p<0.001).

Bivariable Associations Between Neighborhood-Level Exposures and PTSD Score.

For median household income, compared to quartile 1, those in quartile 3 had a PCL-5 score 3.93 points lower (p=0.013), while those in quartile 4 had a score 6.05 points lower (p<0.001). Table 1 shows that for unemployment rate, compared to quartile 1, those in quartile 3 had a PCL-5 score 2.81 points higher (p=0.034), and those in quartile 4 had a score 6.02 points higher (p=0.003). For Gini coefficient, compared to quartile 1, those in quartile 2 had an additional 2.36 PCL-5 points (p=0.048), those in quartile 3 had an additional 2.87 points (p=0.006), and those in quartile 4 had an additional 7.69 points (p<0.001).

Multivariable Associations Between Neighborhood-Level Exposures and Depression Score.

Table 2a shows adjusted associations between median household income and depression score. Models 1–4 show that there were not any statistically significant or meaningful associations between median household income and depression score. Table 3a shows that there were no statistically significant associations between unemployment rate and depression score (Models 1–4). Table 4a shows that for the Gini coefficient that compared to those in quartile 1, those in quartile 2 had a PHQ-9 score that was 1.38 points greater (p=0.062), while those in quartile 3 had a score that was 1.25 points greater (p=0.057), and those in quartile 4 had a score that was 1.76 points greater (p=0.035). After adjusting for Harvey stressors and traumas (Model 2), COVID-19 stressors and traumas (Model 3), and Harvey and COVID-19 stressors and traumas (Model 4), there was no longer a statistically significant or meaningful association between Gini coefficient and depression symptoms.

Multivariable Associations Between Neighborhood-Level Exposures and PTSD Score.

Table 2b shows associations between median household income and PTSD symptom score. Model 1 demonstrates that compared to those in quartile 4, those in quartile 1 had 5.72 more PTSD symptom score points (p=0.026). After adjusting for Harvey stressors and traumas (Model 2), COVID-19 stressors and traumas (Model 3), and Harvey and COVID-19 stressors and traumas (Model 4), there was no longer a statistically significant or meaningful association between median household income and PTSD symptoms. Table 3b, Models 1–4 show that there were not any statistically significant or meaningful associations between unemployment rate and PTSD symptom score after multivariable adjustment. Table 4b, Model 1 shows for Gini coefficient, that compared to those in quartile 1, those in quartile 3 had 3.4 additional PTSD symptom score points (p=0.046), while those in quartile 4 had 6.29 additional points (p=0.003). After adjusting for Harvey stressors and traumas (Model 2), Covid stressors and traumas (Model 3), and Harvey and COVID-19 stressors and traumas (Model 4), there was no longer a statistically significant association between Gini coefficient and PTSD symptoms.

Supplementary Analyses.

Supplementary Table 5 shows the distribution of individual-level income within each quartile of neighborhood-level median income, demonstrating variation in individual income within each quartile. Supplementary Table 6 shows the distribution of individual-level income within each quartile of Gini coefficient.

4. DISCUSSION

This study of Houston residents who experienced Hurricane Harvey and COVID-19 examined associations between neighborhood-level median household income, unemployment, and income inequality and symptoms of depression and PTSD at the individual-level. Partially adjusted models showed that lower neighborhood-level income was associated with greater individual-level PTSD symptom scores, after adjusting for individual-level income, while greater neighborhood-level Gini coefficient was associated with both higher depression and PTSD symptom scores. After adjusting for event-specific individual-level stressors and traumas, these associations were no longer statistically significant.

Our findings on neighborhood-level income and depression are consistent with a study of New York City residents who experienced Hurricane Sandy (Lowe et al., 2014), while our null finding on the association between neighborhood-level unemployment and depression symptom scores is consistent with a study of survivors of the Deepwater Horizon oil spill (Gaston et al., 2017). In contrast, our null results on the income inequality and depression are inconsistent with prior work showing that neighborhood-level income inequality is positively associated with depression among survivors of the September 11th terrorist attacks in New York City (Ahern and Galea, 2006). A prior review and meta-analysis yielded mixed findings of the relationship between neighborhood socioeconomic exposures and depression (Richardson et al., 2015), noting that study results varied by follow-up time, with studies that had a longer follow-up time showing no significant associations. As our study interviewed survivors of Hurricane Harvey over three years after the event, follow-up time may play a role. There may, for example, have been more variance in depression symptoms by neighborhood socioeconomic exposures in the immediate aftermath of the hurricane.

Our findings on neighborhood-level income and PTSD are inconsistent with results from a study of NYC residents following September 11, 2001 (Ahern et al., 2004), and a study on persistence of PTSD following disasters (Neria et al., 2008), while the lack of association between neighborhood-level unemployment and PTSD symptom scores, is inconsistent with a Hurricane Sandy study (Lowe et al., 2016). Finally, our finding that neighborhood income-inequality is not associated with PTSD symptoms is inconsistent with a prior neighborhood study (Ahern, 2004) and state-level work showing that income inequality was positively associated with PTSD (Pabayo et al., 2017). Like our findings for depression, it appears that partially adjusted associations between neighborhood income, Gini coefficient, and PTSD are better explained by event-specific stressful events and traumas.

As posited by McLeod and Kessler (1990), neighborhood-level socioeconomic exposures may structure exposure to stressors and traumatic events, which are signature features of disasters and are linked to the development of both depression and PTSD. Overall, our findings demonstrate that the influences of neighborhood-level economic indicators on depression (Gini coefficient) and PTSD symptoms (median income, Gini coefficient) are indeed better accounted for in this study by Hurricane Harvey and COVID-19 related stressors and traumas. Future studies may fruitfully test whether event-specific stressors and traumas mediate the relationship between neighborhood-level socioeconomic exposures and depression and PTSD symptoms. Observed inconsistencies between prior studies and ours may also be explained by the size, nature, and characteristics of Houston super neighborhoods, compared to other neighborhood contexts previously studied, such as those in New York City. Finally, another reason for these inconsistencies may be the extensive COVID-19 related benefits (Skinner et al., 2022) provided by the state and federal government in between exposure and outcome that may have alleviated some of the exposure effects of neighborhood median income, neighborhood unemployment and neighborhood income inequality on depression and PTSD symptoms.

This study has three key limitations. First, the study had a relatively low response rate of 11.4%. Nonetheless, the study used an address-based random sampling approach, and we applied complex survey weights to account for neighborhood distributions of our target population. In addition, the Household Pulse Survey, which was also fielded during the pandemic had similarly low response rates (Peterson et al., 2021). Second, we used screeners for depression and PTSD symptoms rather than clinical interviews. Nonetheless, the instruments we used to measure depression symptoms (Kroenke et al., 2010, 2001) and PTSD symptoms (Bovin et al., 2016; Morrison et al., 2021) are well validated. Third, while we examined the relation of neighborhood-level economic factors on current symptom scores, measured between 2020 and 2021, we were unable to include symptom scores measured closer to Hurricane Harvey, which happened in 2017. Still, the purpose of this study was to document the associations between prior exposure to neighborhood-level economic factors and current mental health symptoms among a sample of Houstonians exposed to multiple disasters, including the COVID-19 pandemic.

5. CONCLUSION

Notwithstanding noted limitations, this study shows that for residents of Houston who were exposed to Hurricane Harvey and experienced COVID-19, higher neighborhood-level median income was associated with lower current PTSD symptoms and greater neighborhood-level income inequality (i.e., Gini coefficient) was associated with higher current depression and PTSD symptoms. However, in fully adjusted models, these associations appear to be better explained by Hurricane Harvey and COVID-19 stressors and traumas. This suggests that in the context of multiple large scale traumatic events, neighborhood socioeconomic context plays a role in determining risk of individual-level poor mental health through the structuring of exposure to stressful events and traumas.

Supplementary Material

1

Highlights.

  • We studied 872 Houston residents who experienced Hurricane Harvey and COVID-19.

  • Lower neighborhood median income was linked to greater symptoms of PTSD.

  • Higher neighborhood income inequality was linked to greater symptoms of depression.

  • Higher neighborhood income inequality was linked to greater symptoms of PTSD.

  • Adjusting for event-specific stressors and traumas rendered these associations null.

Footnotes

Declaration of Competing Interests

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Sandro Galea reports financial support was provided by National Institute of Mental Health. Kenneth J Ruggiero reports was provided by National Institute of Mental Health. Sarah Lowe reports financial support was provided by National Institute of Mental Health. Howard Cabral reports financial support was provided by National Institute of Mental Health. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Publisher's Disclaimer: 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.

References

  1. Abdalla SM, Ettman CK, Cohen GH, Galea S, 2021. Mental health consequences of COVID-19: a nationally representative cross-sectional study of pandemic-related stressors and anxiety disorders in the USA. BMJ Open 11, e044125. 10.1136/bmjopen-2020-044125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ahern J, Galea S, 2006. Social context and depression after a disaster: the role of income inequality. Journal of Epidemiology & Community Health 60, 766–770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Ahern J, Galea S, Tracy M, Vlahov D, 2004. Neighborhood characteristics, trauma, and the risk of post-traumatic stress disorder in New York City. Annals of Epidemiology 14, 623. 10.1016/j.annepidem.2004.07.085 [DOI] [Google Scholar]
  4. Amiri S, 2022. Unemployment associated with major depression disorder and depressive symptoms: a systematic review and meta-analysis. International Journal of Occupational Safety and Ergonomics 28, 2080–2092. 10.1080/10803548.2021.1954793 [DOI] [PubMed] [Google Scholar]
  5. Bevilacqua K, Rasul R, Schneider S, Guzman M, Nepal V, Banerjee D, Schulte J, Schwartz RM, 2020. Understanding associations between Hurricane Harvey exposure and mental health symptoms among greater Houston-area residents. Disaster medicine and public health preparedness 14, 103–110. [DOI] [PubMed] [Google Scholar]
  6. Blake ES, Zelinsky DA, 2017. National Hurricane Center Tropical Cyclone Report: Hurricane Harvey. [Google Scholar]
  7. Bovin MJ, Marx BP, Weathers FW, Gallagher MW, Rodriguez P, Schnurr PP, Keane TM, 2016. Psychometric properties of the PTSD checklist for diagnostic and statistical manual of mental disorders–fifth edition (PCL-5) in veterans. Psychological assessment 28, 1379. [DOI] [PubMed] [Google Scholar]
  8. Bureau, U.C., n.d. American Community Survey Information Guide. [Google Scholar]
  9. Bureau, U.S.C., 2021. Gini Index [WWW Document]. [Google Scholar]
  10. Case Data | Harris County COVID-19 Data Hub prod2 [WWW Document], n.d [Google Scholar]
  11. City of Houston, n.d. Super Neighborhoods [WWW Document]. URL http://www.houstontx.gov/superneighborhoods/ (accessed 8.7.22).
  12. Costello EJ, Compton SN, Keeler G, Angold A, 2003. Relationships between poverty and psychopathology: A natural experiment. Jama 290, 2023–2029. [DOI] [PubMed] [Google Scholar]
  13. Cummings JR, Allen L, Clennon J, Ji X, Druss BG, 2017. Geographic Access to Specialty Mental Health Care Across High- and Low-Income US Communities. JAMA Psychiatry 74, 476–484. 10.1001/jamapsychiatry.2017.0303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. De Maio FG, 2007. Income inequality measures. Journal of Epidemiology and Community Health 61, 849 LP–852. 10.1136/jech.2006.052969 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Deville J-C, Särndal C-E, Sautory O, 1993. Generalized raking procedures in survey sampling. Journal of the American statistical Association 88, 1013–1020. [Google Scholar]
  16. Dowle M, Srinivasan A, 2021. data.table: Extension of `data.frame`. [Google Scholar]
  17. Ettman CK, Abdalla SM, Cohen GH, Sampson L, Vivier PM, Galea S, 2020. Prevalence of depression symptoms in US adults before and during the COVID-19 pandemic. JAMA network open 3, e2019686–e2019686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. FEMA, 2018. 2017 Hurricane Season FEMA After-Action Report. [Google Scholar]
  19. Fitzpatrick KM, 2021. Post-traumatic stress symptomatology and displacement among Hurricane Harvey survivors. Social Science & Medicine 270, 113634. [DOI] [PubMed] [Google Scholar]
  20. Gaston SA, Volaufova J, Peters ES, Ferguson TF, Robinson WT, Nugent N, Trapido EJ, Rung AL, 2017. Individual-level exposure to disaster, neighborhood environmental characteristics, and their independent and combined associations with depressive symptoms in women. Social psychiatry and psychiatric epidemiology, 52, 1183–1194 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. GBD 2019 Mental Disorders Collaborators, 2022. Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. The lancet. Psychiatry 9, 137–150. 10.1016/S2215-0366(21)00395-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Goldmann E, Galea S, 2014. Mental Health Consequences of Disasters. 10.1146/annurev-publhealth-032013-182435 35, 169–183. [DOI] [PubMed] [Google Scholar]
  23. Hikichi H, Aida J, Tsuboya T, Kondo K, Kawachi I, 2016. Can community social cohesion prevent posttraumatic stress disorder in the aftermath of a disaster? A natural experiment from the 2011 Tohoku earthquake and tsunami. American journal of epidemiology 183, 902–910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kawachi I, Kennedy BP, 1999. Income inequality and health: pathways and mechanisms. Health services research 34, 215. [PMC free article] [PubMed] [Google Scholar]
  25. Kroenke K, Spitzer RL, Williams JBW, 2001. The PHQ-9: validity of a brief depression severity measure. Journal of general internal medicine 16, 606–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kroenke K, Spitzer RL, Williams JBW, Löwe B, 2010. The Patient Health Questionnaire Somatic, Anxiety, and Depressive Symptom Scales: a systematic review. General Hospital Psychiatry 32, 345–359. 10.1016/j.genhosppsych.2010.03.006 [DOI] [PubMed] [Google Scholar]
  27. Lê F, Tracy M, Norris FH, Galea S, 2013. Displacement, county social cohesion, and depression after a large-scale traumatic event. Social psychiatry and psychiatric epidemiology 48, 1729–1741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lowe SR, Sampson L, Gruebner O, Galea S, 2015. Psychological resilience after Hurricane Sandy: the influence of individual-and community-level factors on mental health after a large-scale natural disaster. PloS one 10, e0125761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lowe SR, Sampson L, Gruebner O, Galea S, 2016. Community unemployment and disaster-related stressors shape risk for posttraumatic stress in the longer-term aftermath of Hurricane Sandy. Journal of Traumatic Stress, 29(5), 440–447. [DOI] [PubMed] [Google Scholar]
  30. Margerison-Zilko C, Goldman-Mellor S, Falconi A, Downing J, 2016. Health impacts of the great recession: a critical review. Current epidemiology reports 3, 81–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. McLeod JD, Kessler RC, 1990. Socioeconomic status differences in vulnerability to undesirable life events. Journal of health and social behavior 162–172. [PubMed] [Google Scholar]
  32. Modrek S, Hamad R, Cullen MR, 2015. Psychological well-being during the great recession: changes in mental health care utilization in an occupational cohort. American Journal of Public Health 105, 304–310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Morrison K, Su S, Keck M, Beidel DC, 2021. Psychometric properties of the PCL-5 in a sample of first responders. Journal of Anxiety Disorders 77, 102339. [DOI] [PubMed] [Google Scholar]
  34. Müller K, 2020. here: A Simpler Way to Find Your Files. [Google Scholar]
  35. Nandi A, Galea S, Tracy M, Ahern J, Resnick H, Gershon R, Vlahov D, 2004. Job loss, unemployment, work stress, job satisfaction, and the persistence of posttraumatic stress disorder one year after the September 11 attacks. Journal of Occupational and Environmental Medicine 1057–1064. [DOI] [PubMed] [Google Scholar]
  36. Neria Y, Nandi A, Galea S, 2008. Post-traumatic stress disorder following disasters: a systematic review. Psychological Medicine 38, 467–480. 10.1017/S0033291707001353 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Pabayo R, Fuller D, Goldstein RB, Kawachi I, Gilman SE, 2017. Income inequality among American states and the conditional risk of post-traumatic stress disorder. Social psychiatry and psychiatric epidemiology 52, 1195–1204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Patel V, Burns JK, Dhingra M, Tarver L, Kohrt BA, Lund C, 2018. Income inequality and depression: a systematic review and meta-analysis of the association and a scoping review of mechanisms. World Psychiatry 17, 76–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Peterson S, Toribio N, Farber J, Hornick D, 2021. Nonresponse bias report for the 2020 Household Pulse Survey. Washington, DC: United States Census Bureau. [Google Scholar]
  40. Richardson R, Westley T, Gariépy G, Austin N, Nandi A, 2015. Neighborhood socioeconomic conditions and depression: a systematic review and meta-analysis. Social psychiatry and psychiatric epidemiology 50, 1641–1656. [DOI] [PubMed] [Google Scholar]
  41. Rubin DB, 2004. Multiple imputation for nonresponse in surveys. John Wiley & Sons. [Google Scholar]
  42. Schwartz RM, Tuminello S, Kerath SM, Rios J, Lieberman-Cribbin W, Taioli E, 2018. Preliminary assessment of Hurricane Harvey exposures and mental health impact. International journal of environmental research and public health 15, 974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Skinner A, Flannery K, Nocka K, Bor J, Dean LT, Jay J, Lipson SK, Cole MB, Benfer EA, Scheckman R, 2022. A database of US state policies to mitigate COVID-19 and its economic consequences. BMC public health 22, 1124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. U.S. Census Bureau, n.d. American Community Survey Information Guide. [Google Scholar]
  45. U.S. Census Bureau, 2021. Gini Index [WWW Document]. [Google Scholar]
  46. Walker K, Herman M, 2023. tidycensus: Load US Census Boundary and Attribute Data as “tidyverse” and ‘sf’-Ready Data Frames. [Google Scholar]
  47. Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Robinson D, Seidel DP, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H, 2019. Welcome to the tidyverse. Journal of Open Source Software 4, 1686. 10.21105/joss.01686 [DOI] [Google Scholar]
  48. Wickham H, Bryan J, 2022. readxl: Read Excel Files. [Google Scholar]
  49. Zimmerman FJ, Katon W, 2005. Socioeconomic status, depression disparities, and financial strain: what lies behind the income-depression relationship? Health economics 14, 1197–1215. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1968122-supplement-1.docx (36.6KB, docx)

RESOURCES