Key Points
Question
Do individuals diagnosed with HIV in New Orleans, Louisiana, while living in historically redlined neighborhoods experience a longer time to viral suppression than individuals living in nonredlined neighborhoods?
Findings
This study of 1132 individuals newly diagnosed with HIV with a residential address within a neighborhood historically mapped by the Home Owners’ Loan Corporation showed that persons with HIV diagnosed while living in redlined neighborhoods experienced a longer time to viral suppression than those diagnosed while living in nonredlined neighborhoods. This correlation persists even when additionally accounting for the effect of gentrification.
Meaning
For racial and ethnic minority individuals diagnosed with HIV in historically redlined neighborhoods of New Orleans, the experience during the HIV epidemic is measurably more difficult due, in part, to the concentrated disadvantage shaped by discriminatory housing policies.
Abstract
Importance
Structural racism in the US is evidenced in the discriminatory practice of historical racial redlining when neighborhoods were valued, in part, based on the community’s racial and ethnic compositions. However, the influence of these systemic practices in the context of the HIV epidemic is not well understood.
Objective
To assess the effect of redlining on time to viral suppression among people newly diagnosed with HIV.
Design, Setting, and Participants
Observational study that included individuals diagnosed with HIV from January 1, 2011, to December 31, 2019, in New Orleans, Louisiana. At the time of their HIV diagnosis, these individuals lived in neighborhoods historically mapped by the Home Owners’ Loan Corporation (HOLC). The HOLC lending risk maps classified neighborhoods into 1 of 4 color-coded grades: A (best), B (still desirable), C (definitely declining), and D (hazardous).
Main Outcome and Measures
The primary outcome of interest was time to viral suppression (estimated as the time from the diagnosis date to the date of the first recorded viral load that was <200 copies/mL). Individual-level demographic factors were used to evaluate time to viral suppression along with a neighborhood measure of gentrification (based on US census tract–level characteristics for educational attainment, housing development and value, and household income) and a Cox gamma frailty model with census tract used as the frailty term.
Results
Of 1132 individuals newly diagnosed with HIV, 871 (76.9%) were men and 620 (54.8%) were 25 to 44 years of age. Of the 697 individuals living in historically redlined neighborhoods (HOLC grade D), 100 (14.6%) were living in neighborhoods that were gentrifying. The median time to viral suppression was 193 days (95% CI, 167-223 days) for persons with HIV living in redlined neighborhoods compared with 164 days (95% CI, 143-185 days) for the 435 persons with HIV living in HOLC grade A, B, or C (nonredlined) neighborhoods. Among persons with HIV living in gentrifying neighborhoods, those living in redlined neighborhoods had a longer time to viral suppression compared with persons living in nonredlined neighborhoods (hazard ratio, 0.54 [95% CI, 0.36-0.82]).
Conclusions and Relevance
These findings suggest the enduring effects of systemic racism on present-day health outcomes among persons with HIV. Regardless of their neighborhood’s contemporary level of gentrification, individuals diagnosed with HIV while living in historically redlined neighborhoods may experience a significantly longer time to viral suppression.
This study assesses the effect of redlining on time to viral suppression among people newly diagnosed with HIV in New Orleans, Louisiana.
Introduction
Structural racism is deeply rooted in the laws, policies, and practices that have sustained racial discrimination for hundreds of years in the US.1,2,3 One readily visible and enduring example of these discriminatory practices can be found in the built environment and racially segregated communities.1,4 The US federal housing policy throughout the 20th century created and reinforced segregation.5 As part of the New Deal programs developed in the 1930s to stabilize the housing market in the aftermath of the Great Depression, the Home Owners’ Loan Corporation (HOLC) assessed neighborhood lending risk for government-backed mortgages in more than 200 cities in the US.6,7
Neighborhoods were systematically graded by HOLC examiners in consultation with local bank loan officers, appraisers, and real estate agents on criteria related to housing stock age and condition, household income, employment status of residents, proximity to public transit and other amenities, closeness to unfavorable factors like industrial and commercial areas, and the resident’s racial and ethnic composition.6,8 The HOLC lending risk maps classified neighborhoods into 1 of 4 color-coded grades: A (best), B (still desirable), C (definitely declining), and D (hazardous) (Figure 1).9 The neighborhoods composed of racial and ethnic minority individuals, immigrants, or Jewish residents were almost always graded D (redlined).10 Although the extent of the use of the HOLC lending risk maps is debated, the maps are highly detailed and are evidence of the federal government judging racial and ethnic minority individuals as higher risk and codifying racial segregation.5
Figure 1. Home Owners’ Loan Corporation Map of New Orleans, Louisiana, in 1939.
This is an original map depicting the color-coded grades for the neighborhoods. An “A” grade (green) signified the “best” neighborhoods; a “B” grade (blue) signified “still desirable” neighborhoods; a “C” grade (yellow) signified “definitely declining” neighborhoods; and a “D” grade (red) signified “hazardous” redlined neighborhoods. This map was downloaded from the Digital Scholarship Lab Mapping Inequality Project at the University of Richmond,9 which has a repository of publicly available Home Owners’ Loan Corporation maps.
People living in redlined neighborhoods (HOLC grade D) were routinely denied access to mortgage financing and a pathway to homeownership, one of the most significant means of intergenerational wealth building in the 20th century.9,10 In addition, decades of public and private capital disinvestment in redlined communities has had enduring effects on the health and well-being of the people living in these neighborhoods.11 Research has demonstrated an association between redlining and health-related outcomes, including cardiovascular health, firearm deaths, and preterm birth.10,11,12 Research on health-related outcomes and historical redlining has gained contemporary interest, but the effect of redlining and HIV-related outcomes has not been reported.13
Achieving early and sustained viral suppression after a diagnosis of HIV is an ongoing focus and national priority, especially in Southern states in the US that experience a disproportionate HIV burden and where health disparities are exacerbated.14,15 The importance of achieving rapid viral suppression (within 3 months of HIV diagnosis) has been emphasized as a metric of HIV care success.16 Time to viral suppression (defined as time from HIV diagnosis to initial viral load of <200 copies/mL) is substantially longer in the South than in other regions of the US.17
Identifying neighborhoods most at risk for a longer time to viral suppression allows for a more granular inspection of the HIV epidemic and highlights the role of geographic place on health outcomes, which can better inform the development and deployment of HIV-specific resources for persons with HIV and those at risk for HIV in these communities.18 A deeper understanding of how systemic and neighborhood factors disadvantage persons with HIV can guide the development of evidence-based public health approaches to improve linkage to care and reduce HIV transmission rates by achieving undetectable viral loads sooner.
This observational study seeks to characterize how historical and contemporary housing practices are connected to the experiences of persons with HIV. The current analysis closely followed the conceptual framework outlined by Swope et al10 and investigated the role of gentrification, subsequent of historical redlining, and the intergenerational effects that these practices have had on time to viral suppression among persons newly diagnosed with HIV. Specifically, the focus of this analysis was on the association of historical redlining and contemporary gentrification and individual-level factors of persons with HIV in relation to time to viral suppression among people newly diagnosed with HIV in New Orleans, Louisiana (the most populous city in the state) (eTable and eFigure in Supplement 1).
Methods
Population
Among persons with HIV diagnosed in the State of Louisiana, study inclusion was limited to individuals aged 13 years or older who were diagnosed while living in New Orleans. A person’s address at the time of HIV diagnosis must have been complete (ie, house number, street, city, state, and zip code) to facilitate the geocoding process.
Persons with HIV were excluded if they were diagnosed while living at addresses corresponding to a post office box or a correctional facility. Persons diagnosed with HIV in a correctional setting have a significantly different experience accessing HIV care compared with nonincarcerated individuals. For the primary analysis, individuals were included only if their residence at HIV diagnosis was within a HOLC-graded neighborhood.
The residential address data for persons with HIV who met these criteria were imported into ArcGIS Desktop (Environmental Systems Research Institute) and geocoded.19 Once assigned spatial coordinates, the addresses were spatially joined with both the US census data and the HOLC neighborhood color-coded grade data.
This study received institutional review board approval from the University of Alabama, Birmingham, to analyze the data and publish the results of the analyses. Informed consent was not required due to the observational nature of the study. The Louisiana health department employees were the only ones who had direct access to the data regarding HIV status.
Data Linkage and Variables
HOLC Maps and Area Description Files
The Digital Scholarship Lab Mapping Inequality Project at the University of Richmond has publicly available files of original HOLC maps and digitized area descriptions for cities graded by the HOLC from 1935 to 1940.9 The HOLC residential lending risk maps were created for cities and towns with populations of 40 000 or larger. From this data, the HOLC-graded neighborhoods in New Orleans were identified and mapped.
HOLC Neighborhood Color-Coded Grading System
The HOLC examiners developed a color-coded classification system for grading neighborhoods. On the HOLC lending risk maps, the A grade signified the “best” neighborhoods and these neighborhoods were marked green; the B grade signified “still desirable” neighborhoods and were marked blue; the C grade signified “definitely declining” neighborhoods and were marked yellow; and the D grade signified “hazardous” neighborhoods and were marked red. Hence, the neighborhoods with a D grade were redlined. Using ArcGIS version 10.8.2, the geocoded residence for an individual with a diagnosis of HIV was joined with HOLC-graded shapefiles to match the address with the corresponding HOLC-graded neighborhood.9,19
Enhanced HIV/AIDS Reporting System
Louisiana’s HIV case surveillance records are entered and maintained in a database called the Enhanced HIV/AIDS Reporting System (eHARS) that was developed by the US Centers for Disease Control and Prevention. The eHARS database is maintained by state health departments and includes all case reports, laboratory results, and other documentation related to persons with HIV.20 Other information stored in the eHARS database includes demographics, HIV risk behaviors, vital status, HIV-related laboratory data, and address at time of diagnosis for all persons with HIV or AIDS reported in each state.
HIV Outcome of Time to Viral Suppression
Time to viral suppression was estimated as the time in days between the date of the HIV diagnosis and the date of the first recorded viral load that was less than 200 copies/mL. All censoring was assumed to be independent and individuals were censored if (1) the person with HIV died or moved out of the state, (2) viral suppression was not achieved during the study period, or (3) either the date of HIV diagnosis or the HIV viral load collection date was not available. In addition, all dates were checked for chronological correctness (eg, that the HIV diagnosis date was before and was not the same as the viral load collection date) to ensure all time calculations were feasible.
Individual-Level Factors
The individual-level data were obtained from the eHARS database and were made accessible in this analysis by means of a distributed data model in which all data remained on secure health department computers behind firewalls.21 In addition, the eHARS database contains information on demographics, laboratory tests, and health service use. From these data, 6 individual-level factors were included in the analysis (Box 1).
Box 1. Individual-Level Factors Included in the Analysisa.
Sex at birth (male or female)
Race and ethnicity
Age at HIV diagnosis
Mode of HIV transmission: male-to-male sexual contact (MSM), injecting drug user (IDU), both MSM and IDU, or high-risk heterosexual contact
Year of HIV diagnosis: Louisiana expanded Medicaid in 2016 so to account for this major shift in health care access, a binary variable was created based on the year of HIV diagnosis (2011-2015 or 2016-2019)
Stage 3 HIV (AIDS) at diagnosis, which was defined as having ≥1 of the following within 3 mo of HIV diagnosis date: (1) CD4 cell count <200 cells/μL, (2) CD4 cell percentage <14%, or (3) a diagnosis of an opportunistic infection (eg, pneumocystis carinii pneumonia, candidiasis, Kaposi sarcoma)
Census Data
The 5-year estimates from the American Community Survey (ACS) were matched with geocoded, individual-level data by US census tract. The datasets used for the analysis included the Demographic and Housing Estimates (DP05), the Selected Social Characteristics in the US (DP02), the Selected Economic Characteristics (DP03), and the Selected Housing Characteristics (DP04).22 To allow for comparison across time and by census tract, the ACS data from 2010 and 2019 were used in the derivation of the gentrification measure for the primary analysis, which includes the HIV surveillance data time frame.
Neighborhood Gentrification
Gentrification of a neighborhood is a complex, rapid transformation that is typically characterized by the displacement of lower-income residents by higher-income residents, increased investment in new or rehabilitated housing stock (existing housing units), and the displacement of racial and ethnic minority groups by higher-income White residents.23 The evidence of gentrification for the current analysis was quantified based on criteria outlined by Freeman.24
The modified version of this gentrification measure appears in Box 2 and used data from the ACS to estimate the presence of gentrification from 2010 to 2019 by census tract. The characteristics of gentrification associated with a census tract were defined as being (1) within a definable city (New Orleans), (2) populated by low-income households, (3) with evidence of disinvestment, (4) displacement of poorer residents by more wealthy residents, and (5) an increase in economic investment.24 Based on data availability, these measures were estimated at the census tract level (not per individual) and were quantified using the criteria in Box 2.
Box 2. Criteria for Gentrification at the Census Tract Levela.
Based on the 2010 census, the census tract was within the metropolitan area of New Orleans, Louisiana
-
The median income (for the census tract) was <40th percentile for the respective metropolitan area of New Orleans at the beginning of the intercensal period (2019)
The 5-y estimate of median household income and benefits for all households was adjusted for inflation for 2019 dollars
-
The proportion of housing built from 2010 to 2019 was lower than the proportion found at the 40th percentile for the respective metropolitan area of New Orleans
The 5-y estimate of total housing units built from 2010 to 2019
-
The percentage increase in educational attainment from 2010 to 2019 was greater than the median increase in educational attainment for the respective metropolitan area of New Orleans
The estimated difference in the 5-y estimate of educational attainment (in the population aged ≥25 y) in 2019 and the 5-y estimate of educational attainment (in the population aged ≥25 y) in 2010 that was expressed as a percentage of the 5-y estimate of educational attainment in 2019
-
Experienced an increase in real housing value from 2010 to 2019 for the respective metropolitan area of New Orleans
The estimated difference in the 5-y estimate of the median value of owner-occupied units (dollars) in 2019 and the 5-y estimate of the median value of owner-occupied units in 2010 (adjusted for inflation for 2019 dollars)
By definition, all persons with HIV in the sample lived in New Orleans; therefore, the first criterion for gentrification was met for all individuals. The census tracts that met the remaining criteria (2-5) were considered “gentrifying.” The census tracts that met the first 3 criteria (1-3) or either the fourth or fifth criterion were considered “potentially gentrifying.” All other census tracts were considered “not gentrifying.” If any criteria were not estimable (ie, the ACS estimates were not available for a particular census tract), an estimate for gentrification was not made for that particular census tract.
Neighborhood Choropleth Mapping
For each HOLC-graded neighborhood, the median time to viral suppression was estimated using Kaplan-Meier survival estimates. The median time to viral suppression was mapped for each neighborhood based on the publicly available HOLC-graded shapefiles.9 The median time to viral suppression for HOLC-graded neighborhoods was not mapped if it was not estimable or if a HOLC-graded neighborhood had fewer than 5 persons with HIV diagnosed from 2011 to 2019 (to maintain confidentiality). Importantly, the neighborhood choropleth map does not depict the address of any persons with HIV.
All graphic representations and mapping procedures for median time to viral suppression were generated in R version 4.4.1 (R Foundation for Statistical Computing); in particular, the R packages tidycensus and leaflet were required for the mapping procedure.25,26,27 To facilitate exploration of the choropleth map and the relationships among HOLC-graded neighborhoods, an interactive RShiny web application was created28; this application, supporting documentation, and supplemental R code are freely available online29 (https://jbassler.shinyapps.io/NOLA-TVS-Redlining/).
Statistical Analysis
The preliminary analyses consisted of summarizing available data by HOLC grade using measures of central tendency and dispersion for continuous factors (sample medians and IQRs), and distributional characteristics for categorical factors (frequencies and percentages). Missing data were assumed to be missing at random and were reported in the descriptive summary, but not included in the calculations. For each person with HIV, their exact date of diagnosis and viral suppression were assumed to have occurred at the end of each respective time interval. The difference between these 2 dates is the basis for which time to viral suppression was calculated.
Kaplan-Meier survival estimates and a Cox proportional hazards regression model were used to describe time to viral suppression by HOLC grade. The Cox regression model was controlled for year of HIV diagnosis, age, race and ethnicity, sex at birth by mode of HIV transmission, and stage 3 HIV (AIDS) at diagnosis.
To further examine how the socioeconomic factors affected the association between time to viral suppression and HOLC grade, a Cox gamma frailty model was implemented with census tract used as the frailty term; this model included individual-level factors and a multilevel factor of HOLC grade by gentrification status. Because this model included both individual-level factors and neighborhood characteristics (joined by census tract), census tract was used as the frailty term and was incorporated to account for the possibility of unobserved heterogeneity at the census tract level. The frailty term in this model is equivalent to the random component included in a generalized linear mixed model.
The Efron approximate likelihood method was used to handle potential ties in the data for time to viral suppression. The proportional hazards assumption was assessed and verified by assessment of Schoenfeld residuals for all Cox models. P < .05 was considered statistically significant. All analyses were conducted using SAS version 9.4 (SAS Institute Inc) and implemented using a data distribution model.21,30
Results
From January 1, 2011, to December 31, 2019, 1132 individuals (871 [76.9%] were men and 620 [54.8%] were 25-44 years of age) were newly diagnosed with HIV in New Orleans. These individuals had a residential address at the time of their HIV diagnosis that was within a historically mapped HOLC-graded neighborhood.
Most demographic characteristics were similar for the 697 individuals living in HOLC grade D (redlined) neighborhoods and for the 435 individuals living in HOLC grade A, B, or C (nonredlined) neighborhoods (Table 1). The majority of people (61.6%) lived in redlined neighborhoods. More than half of the HIV diagnoses were among men who have sex with men. In addition, approximately 25% met stage 3 HIV criteria (AIDS) at the time of their HIV diagnosis.
Table 1. Demographics and Characteristics of Persons With HIV by Home Owners’ Loan Corporation (HOLC) Grade for the Neighborhooda.
| Characteristic | HOLC grade for the neighborhood, No. (%)b | Total (N = 1132) | |
|---|---|---|---|
| D (redlined) (n = 697) | A, B, or C (n = 435) | ||
| Achieved viral suppression (viral load <200 copies/mL) by time from HIV diagnosis | |||
| By 3 mo | 182 (26.1) | 132 (30.3) | 314 (27.7) |
| By 6 mo | 332 (47.6) | 232 (53.3) | 564 (49.8) |
| By 12 mo | 434 (62.3) | 299 (68.7) | 733 (64.8) |
| By end of study | 600 (86.1) | 388 (89.2) | 988 (87.3) |
| Never | 97 (13.9) | 47 (10.8) | 144 (12.7) |
| Year range for HIV diagnosis | |||
| 2011-2015 | 430 (61.7) | 284 (65.3) | 714 (63.1) |
| 2016-2019 | 267 (38.3) | 151 (34.7) | 418 (36.9) |
| Age group at time of HIV diagnosis, y | |||
| 13-24 | 160 (23.0) | 94 (21.6) | 254 (22.4) |
| 25-44 | 387 (55.5) | 233 (53.6) | 620 (54.8) |
| ≥44 | 150 (21.5) | 108 (24.8) | 258 (22.8) |
| Sex at birth | |||
| Female | 157 (22.5) | 104 (23.9) | 261 (23.1) |
| Male | 540 (77.5) | 331 (76.1) | 871 (76.9) |
| Race and ethnicity | |||
| Hispanic or Latino | 45 (6.5) | 44 (10.1) | 89 (7.9) |
| Black | 487 (69.9) | 247 (56.8) | 734 (64.8) |
| White | 154 (22.1) | 134 (30.8) | 288 (25.4) |
| Other racec | 11 (1.6) | 10 (2.3) | 21 (1.9) |
| Method of HIV transmission | |||
| Male sex at birth | |||
| MSM | 369 (52.9) | 235 (54.0) | 604 (53.4) |
| Heterosexual contact | 77 (11.0) | 40 (9.2) | 117 (10.3) |
| Unknown or other | 45 (6.5) | 31 (7.1) | 76 (6.7) |
| MSM and IDU or IDU only | 49 (7.0) | 25 (5.7) | 74 (6.5) |
| Female sex at birth | |||
| Heterosexual contact | 121 (17.4) | 78 (17.9) | 199 (17.6) |
| Unknown or other | 20 (2.9) | 18 (4.1) | 38 (3.4) |
| IDU | 16 (2.3) | 8 (1.8) | 24 (2.1) |
| Diagnosed with stage 3 HIV (AIDS) within 3 mo of initial HIV diagnosis | |||
| Yes | 104 (23.9) | 174 (25.0) | 278 (24.6) |
| No | 331 (76.1) | 523 (75.0) | 854 (75.4) |
| Gentrification of neighborhoodd | |||
| No | 420 (61.5) | 364 (88.1) | 784 (71.5) |
| Potentially | 163 (23.9) | 14 (3.4) | 177 (16.1) |
| Yes | 100 (14.6) | 35 (8.5) | 135 (12.3) |
| Missinge | 14 | 22 | 36 |
Abbreviations: IDU, injecting drug user; MSM, men who have sex with men.
The HOLC grade was based on the person’s neighborhood at the time of HIV diagnosis.
An “A” grade signified the “best” neighborhoods; a “B” grade signified “still desirable” neighborhoods; a “C” grade signified “definitely declining” neighborhoods; and a “D” grade signified “hazardous” redlined neighborhoods.
Included the categories of American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, and multiple races.
Defined as being (1) within a definable city (New Orleans), (2) populated by low-income households, (3) with evidence of disinvestment, (4) displacement of poorer residents by more wealthy residents, and (5) an increase in economic investment.23
Data were not included in calculations.
There were more Black individuals (69.9%; 487 of 697) living in redlined neighborhoods compared with Black individuals living in nonredlined neighborhoods (56.8%; 247 of 435). Even though the largest proportion of people in both groups lived in census tracts that were classified as not gentrifying, the redlined neighborhoods had a higher percentage of people living in gentrifying census tracts (14.6%; 100 of 683) compared with people in nonredlined neighborhoods (8.5%; 35 of 413).
The maximum time from HIV diagnosis was 4 years; there were 988 persons (87.3%) who achieved viral suppression. Within 3 months of HIV diagnosis, 182 persons (26.1%) achieved viral suppression who were diagnosed with HIV while living in redlined neighborhoods compared with 132 persons (30.3%) diagnosed in nonredlined neighborhoods. People newly diagnosed with HIV in redlined neighborhoods experienced a longer time to viral suppression than those from nonredlined neighborhoods (Figure 2). The unadjusted median time to viral suppression among persons with HIV in redlined neighborhoods was 193 days (95% CI, 167-223 days) compared with 164 days (95% CI, 143-185 days) among persons with HIV in nonredlined neighborhoods, or 15% fewer days to achieve viral suppression (Table 2).
Figure 2. Cumulative Incidence of Viral Suppression by Home Owners’ Loan Corporation (HOLC) Grade.
Compares persons with HIV who had been living in historically HOLC-graded “D” neighborhoods at the time of HIV diagnosis vs persons with HIV who had been living in historically HOLC-graded “A”, “B”, or “C” neighborhoods at the time of HIV diagnosis. An “A” grade signified the “best” neighborhoods; a “B” grade signified “still desirable” neighborhoods; a “C” grade signified “definitely declining” neighborhoods; and a “D” grade signified “hazardous” redlined neighborhoods.
Table 2. Time to Viral Suppression by Home Owners’ Loan Corporation (HOLC) Grade and Gentrification Status of Neighborhooda.
| Type of analysis | No. of individuals with HIV | Achieved viral suppression, No. (%) |
Hazard ratio (95% CI)b | P value |
|---|---|---|---|---|
| Bivariate analysis | ||||
| Time to viral suppression by HOLC grade | ||||
| D (redlined)c | 697 | 600 (86) | 0.84 (0.74-0.95) | .007e |
| A, B, or Cd | 435 | 388 (89) | 1 [Reference] | |
| Multivariate analysis | ||||
| Time to viral suppression by HOLC gradef | ||||
| D (redlined) | 697 | 600 (86) | 0.86 (0.76-0.91) | .03g |
| A, B, or C | 435 | 388 (89) | 1 [Reference] | |
| Time to viral suppression by HOLC grade and gentrification status of neighborhoodh | ||||
| D (redlined) | .049i | |||
| No, not gentrifying | 420 | 363 (86) | 0.62 (0.43-0.89) | |
| Potentially gentrifying | 163 | 143 (88) | 0.60 (0.41-0.88) | |
| Yes, gentrifying | 100 | 82 (82) | 0.54 (0.36-0.82) | |
| A, B, or C | ||||
| No, not gentrifying | 364 | 322 (88) | 0.68 (0.47-0.98) | |
| Potentially gentrifying | 14 | 14 (100) | 0.75 (0.40-1.43) | |
| Yes, gentrifying | 35 | 33 (94) | 1 [Reference] | |
An “A” grade signified the “best” neighborhoods; a “B” grade signified “still desirable” neighborhoods; a “C” grade signified “definitely declining” neighborhoods; and a “D” grade signified “hazardous” redlined neighborhoods.
A Cox proportional hazards regression model was used.
The time to viral suppression was a median of 193 days (IQR, 167-223 days) and was estimated using Kaplan-Meier survival estimates.
The time to viral suppression was a median of 164 days (IQR, 143-185) days and was estimated using Kaplan-Meier survival estimates.
Calculated using the log-rank test.
The other model covariates included year of HIV diagnosis (categorized), age at HIV diagnosis (years), race and ethnicity, method of HIV transmission by sex at birth, and HIV stage at time of HIV diagnosis.
Calculated using the Wald χ21 test.
A Cox gamma frailty model was implemented with census tract used as the frailty term. The other model covariates included year of HIV diagnosis (categorized), age at HIV diagnosis (years), race and ethnicity, method of transmission by sex at birth, AIDS diagnosis (within 3 months of HIV diagnosis).
Calculated using the Wald χ25 test with adjusted degrees of freedom.
Of 697 persons with HIV in redlined neighborhoods, 600 achieved viral suppression vs 388 of 435 persons in nonredlined neighborhoods (hazard ratio [HR], 0.84 [95% CI, 0.74-0.95]). After adjusting for year of HIV diagnosis, race and ethnicity, mode of transmission, and stage 3 HIV status (AIDS) at diagnosis, the association between living in a redlined neighborhood when diagnosed with HIV and time to viral suppression was slightly attenuated (Table 2), but remained statistically significant (HR, 0.86 [95% CI, 0.76-0.91], P = .03).
The differences in unadjusted time to viral suppression by redlined neighborhood appear in Figure 3. Of the 134 HOLC-graded neighborhoods in New Orleans, 104 (77.6%) included persons with HIV diagnosed from 2011 to 2019 and 57 (54.8%) met the eligibility criteria to be represented in the publicly available choropleth map. Among the mapped HOLC-graded neighborhoods, the majority (30; 52.6%) represent HOLC grade D (redlined) neighborhoods. Of the 13 neighborhoods with the longest time (≥270 days) to viral suppression, 7 (53.8%) were in redlined neighborhoods.
Figure 3. Median Time to Viral Suppression (VS) in Redlined Neighborhoods.
This Figure does not represent the incidence of HIV for each neighborhood by grade from the Home Owners’ Loan Corporation (HOLC). The darker colors indicate the neighborhoods with longer times to viral suppression. To facilitate exploration of the choropleth map and the relationships among HOLC-graded neighborhoods, an interactive RShiny web application was created28; this application, supporting documentation, and supplemental R code are freely available online29 (https://jbassler.shinyapps.io/NOLA-TVS-Redlining/).
For the multivariate Cox model for frailty (Table 2), 1096 (96.8%) people lived in census tracts with ACS estimates for all required gentrification criteria (Box 2) and were included in the analysis. Among persons with HIV living in gentrifying neighborhoods, those living in HOLC grade D (redlined) neighborhoods had a longer time to viral suppression compared with persons living in HOLC grade A, B, or C (nonredlined) neighborhoods (hazard ratio, 0.54 [95% CI, 0.36-0.82]; Table 2). Among persons with HIV living in neighborhoods not gentrifying, those living in redlined neighborhoods were less likely to achieve viral suppression at any given point vs persons with HIV in nonredlined neighborhoods gentrifying (HR, 0.62 [95% CI, 0.43-0.89]).
Discussion
Among people diagnosed with HIV in New Orleans from 2011 to 2019, persons who were diagnosed while living in historically redlined neighborhoods experienced a longer time to viral suppression compared with persons with HIV diagnosed while living in nonredlined neighborhoods. This association remained after adjusting for individual-level factors (year of HIV diagnosis, age at HIV diagnosis, race and ethnicity, mode of transmission by sex at birth, and having stage 3 HIV [AIDS] at diagnosis) and contemporary neighborhood characteristics (gentrification status). These findings mirror studies for other health conditions,31,32 suggesting deleterious health effects of historical racism on contemporary outcomes, and further demonstrating the role of contemporary gentrification in prolonging time to viral suppression among newly diagnosed persons with HIV.
The research from the current study shows that persons with HIV diagnosed in neighborhoods that historically had been given preferable HOLC grades (A, B, or C) and recently experienced reinvestment in their community associated with gentrification achieved viral suppression significantly faster than persons in redlined neighborhoods (grade D). This relationship is consistent, regardless of the gentrification status of the neighborhood. In addition, there is evidence of the combined effect of redlined status and gentrification because persons with HIV diagnosed in nonredlined neighborhoods that were not gentrifying also have a longer time to viral suppression vs persons in nonredlined neighborhoods that were gentrifying. This suggests that, even in redlined neighborhoods that experience an influx of reinvestment, persons with HIV in these communities, who are predominantly racial and ethnic minority individuals, are not benefiting.
The influence of place on a person’s health and well-being is profound, and is illustrated in Figure 3 for median time to viral suppression by redlined neighborhoods. In particular, this analysis sheds light on the detrimental effects of one particular facet of systemic racism in US society and its self-perpetuating legacy. The legacy of structural racism persists, and it cannot be ignored because these same communities endure contemporary forms of discrimination through gentrification. This report highlights the need for these communities to be a focal point of ongoing efforts of the Ending the HIV Epidemic in the US initiative.
Although an earlier study did not find evidence of an association between redlining and HIV care outcomes independent of Medicaid and Ryan White HIV/AIDS Program use,13 the current research was more focused in its scope. In addition, a key difference in the approach taken by Logan et al13 is the differing mechanism for indicating redlined neighborhoods. Logan et al13 used data from the Home Mortgage Disclosure Act to derive a contemporary racial mortgage redlining index at the census tract level, whereas the current analysis used individual-level address data to geocode persons within neighborhoods historically redlined by the HOLC.
Racism is a public health crisis, and for racial and ethnic minority individuals diagnosed with HIV in historically redlined neighborhoods of New Orleans, their experience in the HIV epidemic is shown to be measurably more difficult, partially owing to the concentrated disadvantage shaped by discriminatory housing policies. Understanding the complex role that place has on the experience of persons with HIV and applying strategies to address the needs of persons with HIV in these communities can mitigate the compounding effect of generational and modern systemic inequities on time to viral suppression and other HIV-related outcomes.
Limitations
There are limitations to this study. First, individual-level data are dependent on the consistency and quality of the eHARS data used for analysis and the geocoding process. Incomplete addresses or missing address information at time of HIV diagnosis, even in small numbers, could bias time to viral suppression in neighborhoods with low HIV morbidity. Second, this analysis focused on persons by residency at time of HIV diagnosis for study inclusion and we were not able to account for possible housing changes within the city for the duration of the study, including persons who may have become unstably housed.
Third, census measures were applied uniformly across each neighborhood even though we know that there is variability within the census tract. Also, it is reasonable to assume that neighborhoods that were originally graded as A, B, or C would be less susceptible to mechanisms associated with gentrification. There were 49 persons (11.9%) with HIV diagnosed while living in HOLC-graded A, B, or C neighborhoods who were found to be living in neighborhoods that were gentrifying or potentially gentrifying.
Fourth, time to viral suppression from HIV diagnosis was assessed by comparison of median number of days over time. For this interval, it is assumed that both HIV diagnosis and viral suppression occurred at the end of each respective time interval. Specifically, there are likely many factors contributing to time to viral suppression at the community level, and this is one of the reasons that the differences in time to viral suppression between redlined and nonredlined neighborhoods were modest.
Conclusions
These findings suggest the enduring effects of systemic racism on present-day health outcomes among persons with HIV. Regardless of their neighborhood’s contemporary level of gentrification, individuals diagnosed with HIV while living in historically redlined neighborhoods may experience a significantly longer time to viral suppression.
eTable. Time to Viral Suppression by HOLC Grade, Univariable and Multivariable Analysis, Full Model
eFigure. Flow Diagram of PWH Diagnosed With HIV in New Orleans From 2011-2019 and Study Inclusion/Exclusion Criteria
Data sharing statement
Footnotes
The data for these individual-level factors were obtained from the Enhanced HIV/AIDS Reporting System.
American Community Survey data were used to estimate the presence of gentrification from 2010 to 2019.
References
- 1.Braveman P, Arkin E, Proctor D, Kauh T, Holm N. Systemic Racism and Health Equity. Robert Wood Johnson Foundation; 2022. [Google Scholar]
- 2.Bailey ZD, Feldman JM, Bassett MT. How structural racism works—racist policies as a root cause of US racial health inequities. N Engl J Med. 2021;384(8):768-773. doi: 10.1056/NEJMms2025396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Schindler S. Architectural exclusion: discrimination and segregation through physical design of the built environment. Yale Law J. 2015;124(6):1934-2024. https://www.yalelawjournal.org/article/architectural-exclusion [Google Scholar]
- 4.Bailey ZD, Krieger N, Agénor M, Graves J, Linos N, Bassett MT. Structural racism and health inequities in the USA: evidence and interventions. Lancet. 2017;389(10077):1453-1463. doi: 10.1016/S0140-6736(17)30569-X [DOI] [PubMed] [Google Scholar]
- 5.Rothstein R. The Color of Law: A Forgotten History of How our Government Segregated America. Liveright Publishing Corporation; 2017. [Google Scholar]
- 6.Aaronson D, Hartley D, Mazumder B. The effects of the 1930s HOLC “redlining” maps. Am Econ J Econ Policy. 2021;13(4):355-392. doi: 10.1257/pol.20190414 [DOI] [Google Scholar]
- 7.Wheelock DC. The federal response to home mortgage distress: lessons from the Great Depression. Fed Reserve Bank of St Louis Rev. 2008;90(3):133-148. doi: 10.20955/r.90.133-148 [DOI] [Google Scholar]
- 8.Mitchell B. HOLC “redlining” maps: the persistent structure of segregation and economic inequality. Published March 20, 2018. Accessed August 26, 2024. https://ncrc.org/holc/
- 9.Nelson RK, Winling L, et al. ; Digital Scholarship Lab Mapping Inequality Project at the University of Richmond . Mapping inequality: redlining in New Deal America. Published 2023. Accessed August 26, 2024. https://dsl.richmond.edu/panorama/redlining
- 10.Swope CB, Hernández D, Cushing LJ. The relationship of historical redlining with present-day neighborhood environmental and health outcomes: a scoping review and conceptual model. J Urban Health. 2022;99(6):959-983. doi: 10.1007/s11524-022-00665-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Mujahid MS, Gao X, Tabb LP, Morris C, Lewis TT. Historical redlining and cardiovascular health: the Multi-Ethnic Study of Atherosclerosis. Proc Natl Acad Sci U S A. 2021;118(51):e2110986118. doi: 10.1073/pnas.2110986118 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Benns M, Ruther M, Nash N, Bozeman M, Harbrecht B, Miller K. The impact of historical racism on modern gun violence: redlining in the city of Louisville, KY. Injury. 2020;51(10):2192-2198. doi: 10.1016/j.injury.2020.06.042 [DOI] [PubMed] [Google Scholar]
- 13.Logan J, Crepaz N, Luo F, et al. HIV care outcomes in relation to racial redlining and structural factors affecting medical care access among Black and White persons with diagnosed HIV—United States, 2017. AIDS Behav. 2022;26(9):2941-2953. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cohen MS, Chen YQ, McCauley M, et al. ; HPTN 052 Study Team . Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365(6):493-505. doi: 10.1056/NEJMoa1105243 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lundgren JD, Babiker AG, Gordin F, et al. ; INSIGHT START Study Group . Initiation of antiretroviral therapy in early asymptomatic HIV infection. N Engl J Med. 2015;373(9):795-807. doi: 10.1056/NEJMoa1506816 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Dombrowski JC, Baeten JM. It’s time to make the time to viral suppression after HIV diagnosis a metric of HIV care success. J Infect Dis. 2019;219(6):845-847. doi: 10.1093/infdis/jiy539 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention Division of HIV/AIDS Prevention . Diagnoses of HIV infection in the United States and dependent areas, 2016. Published November 2017. Accessed August 30, 2024. https://stacks.cdc.gov/view/cdc/50233
- 18.Batey DS. Exploring individual- and community-level predictors and mediators of suboptimal HIV primary care appointment adherence: the importance of place. Accessed August 26, 2024. https://ir.ua.edu/handle/123456789/1929
- 19.Environmental Systems Research Institute . ArcGIS Desktop: Release 10.8.2. Environmental Systems Research Institute; 2023. [Google Scholar]
- 20.Cohen SM, Gray KM, Ocfemia MC, Johnson AS, Hall HI. The status of the national HIV surveillance system, United States, 2013. Public Health Rep. 2014;129(4):335-341. doi: 10.1177/003335491412900408 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bassler JR, Cagle I, Crear D, et al. Development and implementation of a distributed data network between an academic institution and state health departments to investigate variation in time to HIV viral suppression in the Deep South. BMC Public Health. 2023;23(1):937. doi: 10.1186/s12889-023-15924-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.US Census Bureau . American Community Survey data. Published May 25, 2023. Accessed June 12, 2023. https://www.census.gov/programs-surveys/acs/data.html
- 23.University of Minnesota Law School . American neighborhood change in the 21st century: gentrification and decline. Published 2024. Accessed February 29, 2024. https://law.umn.edu/institute-metropolitan-opportunity/studies/housing-and-planning/american-neighborhood-change
- 24.Freeman L. Displacement or succession? residential mobility in gentrifying neighborhoods. Urban Aff Rev. 2005;40(4):463-491. doi: 10.1177/1078087404273341 [DOI] [Google Scholar]
- 25.R Foundation for Statistical Computing . R: a language and environment for statistical computing. Accessed August 26, 2024. http://www.R-project.org/
- 26.Cheng J, Karambelkar B, Xie Y. Leaflet: create interactive web maps with the JavaScript Leaflet library. Accessed August 26, 2024. https://rstudio.github.io/leaflet/
- 27.Walker K, Herman M. Tidycensus: load US Census boundary and attribute data as ‘tidyverse’ and ‘sf’—ready data frames: R version 1.6.1. Accessed August 26, 2024. https://walker-data.com/tidycensus/
- 28.Chang W, Cheng J, Allaire J, et al. Shiny: web application framework for R version 1.8.0.9000. Accessed August 26, 2024. https://github.com/rstudio/shiny
- 29.Bassler J, Ostenga L. Map supplement: redlining and time to viral suppression among persons with HIV. Accessed September 3, 2024. 10.17605/OSF.IO/YJKUN [DOI] [PMC free article] [PubMed]
- 30.SAS Institute Inc . TS1M7 version 9.4. Accessed August 26, 2024. https://support.sas.com
- 31.Hollenbach SJ, Thornburg LL, Glantz JC, Hill E. Associations between historically redlined districts and racial disparities in current obstetric outcomes. JAMA Netw Open. 2021;4(9):e2126707. doi: 10.1001/jamanetworkopen.2021.26707 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Hussaini SMQ, Fan Q, Barrow LCJ, Yabroff KR, Pollack CE, Nogueira LM. Association of historical housing discrimination and colon cancer treatment and outcomes in the United States. JCO Oncol Pract. 2024;20(5):678-687. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable. Time to Viral Suppression by HOLC Grade, Univariable and Multivariable Analysis, Full Model
eFigure. Flow Diagram of PWH Diagnosed With HIV in New Orleans From 2011-2019 and Study Inclusion/Exclusion Criteria
Data sharing statement