Abstract
Racial/ethnic minorities are disproportionately affected by poor HIV care outcomes. Studies have also examined the effects of neighborhood-level factor on an individual's health outcomes. Thus, the objective of this study was to assess the effects of neighborhood factors on the association between race/ethnicity and sustained viral suppression (all viral load tests <200 copies/mL per year). Data for 6491 people with HIV in the 2017 Miami-Dade County Ryan White Program and neighborhood-level data by ZIP code tabulated areas from the American Community Survey were utilized. Multi-level logistic regression models were used to assess the role of neighborhood factors on the association between race/ethnicity and sustained viral suppression. Results show that non-Hispanic Blacks had lower odds of sustained viral suppression in low socioeconomic disadvantage [adjusted odds ratio (aOR): 0.39; 95% confidence interval (CI): 0.20–0.74], moderate residential instability (aOR: 0.31; 95% CI: 0.15–0.65), and low and high racial/language homogeneity neighborhoods (aOR: 0.38; 95% CI: 0.16–0.88) and (aOR: 0.38; 95% CI: 0.19–0.75), respectively, when compared to non-Hispanic Whites (NHWs). Haitians also exhibited poor outcomes in neighborhoods characterized by moderate residential instability (aOR: 0.42; 95% CI: 0.18–0.97) and high racial/language homogeneity (aOR: 0.49; 95% CI: 0.26–0.93), when compared to NHWs. In conclusion, disparities in rates of sustained viral suppression were observed for racial/ethnic minorities within various neighborhood-level factors. These findings indicate the importance of addressing neighborhood characteristics to achieve optimal care for minorities.
Keywords: sustained viral suppression, HIV/AIDS, neighborhood factors, racial/ethnic disparities
Introduction
Neighborhood-level factors such as physical and social environments explain the endogenous and contextual determinants of health that can wield considerable influence on health outcomes.1–4 A neighborhood's social environment can be described using theories such as social disorganization, neighborhood disadvantage, and disorder, which are largely intertwined. The three dimensions of social disorganization (concentrated disadvantage, residential instability, and ethnic heterogeneity) have been used to critically examine the role of neighborhoods on health behavior and crime.5
Concentrated disadvantage is characterized by lack of adequate resources for basic needs, residential instability disturbs social networks and deteriorates community ties due to high mobility, and ethnic heterogeneity hinders shared commonalities and communications, which weakens social cohesion and collective efficacy.5,6 The lack of preservation of social ties and economic resources in neighborhoods promotes social disorganization and disorder.5,7,8 Moreover, neighborhoods with high degree of social and physical disorder create a hostile living environment that has deleterious effects on health2,3,9 and contributes to racial/ethnic differences in life expectancy.10
Furthermore, residential segregation is a major driver for racial and socioeconomic health disparities.4,11 Residential segregation and inequalities in resource distribution affect a neighborhood's environment, which then influences an individual's behavior and stress response, and coupled with psychosocial factors, ultimately affects health outcomes.1 In particular, African Americans often reside in racially segregated, and economically and socially deprived neighborhoods with limited access to resources.9
People with HIV (PWH) who have a suppressed viral load greatly reduce their risk of disease progression to AIDS and transmitting the disease to others.12,13 Once an individual is virally suppressed, it is imperative that viral suppression is sustained long term. Sustained viral suppression, defined as having all viral load tests within a year as <200 copies/mL, is critically important as breaks in viral suppression within a year might increase the possibility of transmitting HIV to others.12,14
Studies have shown that disparities in sustained viral suppression exist by race/ethnicity.13,15 Despite the increased use of antiretroviral medication over the years, African Americans have lower rates of sustained viral suppression than Hispanics and non-Hispanic Whites (NHWs), and spend more than half of the year with viral loads >1500 copies/mL.15,16 Elevated disease burden, health disparities, and inequities are, in part, attributed to neighborhood-level factors,3 including HIV health outcomes.17
Poor HIV care outcomes, including lack of medication adherence and unsuppressed viral load, have been associated with neighborhoods characterized by economic deprivation and residential segregation.3,18–20 Increased neighborhood disorder has also been associated with poor HIV medication adherence, which is necessary to achieve and maintain viral suppression.3,17 However, little is known about the role of neighborhood factors on achievement of sustained viral suppression. Therefore, the objective of this study was to examine the contribution of neighborhood-level factors on sustained viral suppression and to assess how the effects of neighborhood context vary by race/ethnicity. This study hypothesizes that racial/ethnic minority disparities in sustained viral suppression will be larger in areas with poor neighborhood characteristics.
Methods
Datasets and study design
A cross-sectional study was conducted using 2017 client data from the Miami-Dade County, Ryan White Program (RWP).21 The Miami-Dade RWP provides care to nearly 10,000 low-income uninsured or underinsured PWH every year. Clients in this study were adults ≥18 years of age, who had received medical case management services, medical care, and viral load laboratory test throughout 2017. Clients whose cases were closed (due to financial ineligibility, relocation, or mortality), those with missing client assessment or viral load data, those with missing residential ZIP codes, or those who only received ancillary services from the RWP were excluded from this analysis.
Neighborhood-level data were collected from the American Community Survey (ACS)22 and Simply Analytics (an online data visualization platform that pools demographic, economic, and health data from various data partners on various health measures).23 The 2013–2017 5-year estimates were obtained by ZIP code tabulated areas (ZCTA), which are geographic units used to approximate ZIP codes. In addition, homicide data were collected by ZIP codes from US Department of Justice, Uniform Crime Reporting Statistics, which was downloaded from Simply Analytics. A total of 77 Miami-Dade ZCTA were used in this analysis. ZCTA that were commercial zones were excluded from the analysis and one-to-one matching of ZIP codes to ZCTA was conducted.
Predictor variables: individual level
Individual-level client characteristics assessed as categorical variables include age, gender, US nativity, mode of HIV exposure, history of AIDS diagnosis, recipient of RWP subsidized Affordable Care Act (ACA) health care, household income in federal poverty level (FPL), and HIV provider's RWP client load.
A total of 16 vulnerable and enabling related variables were selected, based on the Andersen Behavioral Model for HIV Health Care Utilization24,25 and variables available in the RWP dataset, to create the indices for psychosocial variables using reliability analysis followed by exploratory and confirmatory factor analysis (Supplementary Table S1). Andersen's model assesses health care environment and vulnerable, enabling, and need factors to determine health service-seeking behaviors. Reliability analysis was first conducted by standardizing variables to account for varying measuring scales and to improve convergence. Internal consistency was assessed by computing Cronbach's alpha of 0.68 after the removal of six variables. Next, exploratory factor analysis with and without an orthogonal varimax rotation removed two additional variables with factor loadings <0.4. Based on an eigenvalue >1, three factor indices, mental health index (factor loadings: 0.57–0.58), substance use index (factor loadings: 0.48–0.55), and income/socioeconomic status (SES) index (factor loadings: 0.47–0.48), were selected. Finally, confirmatory factor analysis yielded the continuous composite scores for each index.
Predictor variables: neighborhood-level variables
Neighborhood-level variables were selected based on social disorganization theory, while average number of homicides during 2013–2017 was used to assess neighborhood disorder. Using similar methods as the individual-level psychosocial indices, neighborhood indices were developed for 24 variables that were collected from ACS and were used as a proxy for social characteristics in prior literature (Supplementary Table S1).1,26 This created three indices (SES disadvantage index, residential instability index, and racial/language homogeneity index), which were categorized into tertiles (low, moderate, and high) with a Cronbach's alpha of 0.94 after deleting eight variables.
The SES disadvantage index (factor loadings: 0.54–0.89) was composed of 12 variables, including public assistance, vehicle ownership, crowding, income disparity, education, occupation, and employment. The residential instability index consisted of rented housing and mobility (factor loadings: 0.65–0.75). Finally, racial/language homogeneity index consisted of percent non-Hispanic Black (NHB) and English language proficiency (factor loadings: 0.75–0.78).
Outcome variable
Sustained viral suppression was the outcome of interest and was defined as having <200 copies/mL in all viral load laboratory tests during 2017.15 Clients were required to have more than one viral load test within a year and those with any unsuppressed viral load test were not considered sustained. If an individual had only one suppressed viral load test or had multiple suppressed viral load tests <3 months apart (which could be duplicate test results), the last viral load test in 2016 was assessed to ensure sustained viral suppression. A total of 436 clients with only one viral load test from 2016 to 2017 were removed.
Statistical analysis
Collinearity between all variables was first assessed to circumvent redundancy. High correlation was observed between variables US born and preferred language (0.71) and between homicide and residential instability index (0.70). Homicide was also removed from the final model, as residential instability is a better measure of neighborhood dynamics. A sensitivity analysis was conducted by replacing residential instability with homicide. Bivariate analysis was conducted using chi-squared test for categorical variables and Wilcoxon signed-ranked tests for continuous variables since the data are not normally distributed.
Multi-level logistic regression models were used to examine the association between sustained viral suppression and individual- and neighborhood-level factors and examine the moderating role of neighborhood indices. Four models were fitted for this analysis to evaluate estimates of the multi-level model independently. To assess intraclass correlation (ICC), the first model was an unconditional random intercept model with ZIP code random effects. ICC is a measure of clustering effect used to determine if a multi-level model is necessary.
The second model included the individual-level variables (demographic and psychosocial indices). The third model included variables from the second model and neighborhood-level indices. Finally, the fourth model included all variables from the third model as well as the interaction terms between race/ethnicity and neighborhood indices. All statistical analyses were performed using SAS 9.4. This study was approved by the Florida International University Institutional Review Board.
Results
The study sample comprised 6491 people, of whom 58% were Hispanic, 24% NHB, 11% Haitian, and 7% NHW. NHBs were less likely to achieve sustained viral suppression (70.4%) than Hispanics (85.7%), Haitians (74.1%), and NHW (84.5%), (p-value <0.0001). Table 1 shows the results of the bivariate analysis of each variable by sustained viral suppression. The largest percentage of population was ≥50 years old, males, was US born, had a household income of <100 FPL, was MSM, was without an AIDS diagnosis, had a provider with ≥200 RWP clients, was not ACA recipient, and resided in neighborhoods of high SES disadvantage (35.8%), high residential instability (35.3%), and high racial/language homogeneity (33.9%).
Table 1.
Descriptive Characteristics of Sustained Viral Suppression for Ryan White HIV/AIDS Program Clients by Race/Ethnicity, in Miami, Florida, 2017
| Total | Total, n (%) | NHB |
Hispanic |
Haitian |
NHW |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Not sustained, n (%) |
Sustained, n (%) |
p
|
Not sustained, n (%) |
Sustained, n (%) |
p
|
Not sustained, n (%) |
Sustained, n (%) |
p
|
Not sustained, n (%) |
Sustained, n (%) |
p
|
||
| n = 1579 | n = 3771 | n = 710 | n = 431 | ||||||||||
| Age | <0.0001 | 0.0005 | <0.0001 | 0.0021 | |||||||||
| 18–34 | 1423 (21.9) | 162 (41.4) | 229 (58.6) | 154 (17.9) | 708 (82.1) | 34 (39.1) | 53 (60.9) | 17 (20.5) | 66 (79.5) | ||||
| 35–49 | 2504 (38.6) | 147 (31.5) | 320 (68.5) | 233 (14.3) | 1404 (85.8) | 75 (31.4) | 164 (68.6) | 34 (21.2) | 127 (78.8) | ||||
| ≥50 | 2564 (39.5) | 158 (21.9) | 563 (78.1) | 151 (11.8) | 1121 (88.1) | 75 (19.5) | 309 (80.5) | 16 (8.6) | 171 (91.4) | ||||
| Gender | 0.1276 | 0.0966 | 0.6429 | 0.0938 | |||||||||
| Males | 4988 (76.8) | 298 (31.0) | 664 (69.0) | 461 (13.9) | 2852 (86.1) | 89 (26.7) | 244 (73.3) | 55 (14.5) | 325 (85.5) | ||||
| Female | 1503 (23.2) | 169 (27.4) | 448 (72.6) | 77 (16.8) | 381 (83.2) | 95 (25.2) | 282 (74.8) | 12 (23.5) | 39 (76.5) | ||||
| US born | 0.0026 | 0.0037 | 0.0914 | 0.6428 | |||||||||
| No | 4405 (67.9) | 42 (20.6) | 162 (79.4) | 469 (13.7) | 2946 (86.3) | 166 (25.2) | 494 (74.9) | 18 (14.3) | 108 (85.7) | ||||
| Yes | 2086 (32.1) | 425 (30.9) | 950 (69.1) | 69 (19.4) | 287 (80.6) | 18 (36.0) | 32 (64.0) | 49 (16.1) | 256 (83.9) | ||||
| Household income (FPL) | <0.0001 | <0.0001 | 0.0694 | <0.0001 | |||||||||
| ≥200% | 1516 (23.4) | 41 (17.6) | 192 (82.4) | 90 (8.7) | 942 (91.3) | 25 (21.6) | 91 (78.4) | 7 (5.2) | 128 (94.8) | ||||
| 100%–199% | 2303 (35.5) | 120 (23.6) | 389 (76.4) | 160 (11.5) | 1230 (88.5) | 58 (22.7) | 198 (77.3) | 21 (14.2) | 127 (85.8) | ||||
| <100% | 2672 (41.2) | 306 (36.6) | 531 (63.4) | 288 (21.4) | 1061 (78.7) | 101 (29.9) | 237 (70.1) | 39 (26.4) | 109 (73.7) | ||||
| Mode of transmission | 0.5573 | 0.0005 | 0.0091 | 0.1826 | |||||||||
| Hetero-sexual | 2887 (44.5) | 320 (28.7) | 792 (71.2) | 169 (16.1) | 879 (83.9) | 160 (24.6) | 491 (75.4) | 15 (19.7) | 61 (80.3) | ||||
| MSM | 3359 (51.8) | 122 (31.6) | 264 (68.4) | 341 (13.1) | 2266 (86.9) | 14 (35.0) | 26 (65.0) | 45 (13.8) | 281 (86.2) | ||||
| IDU+other | 245 (3.8) | 25 (30.9) | 56 (69.1) | 28 (24.1) | 88 (75.9) | 10 (52.6) | 9 (47.4) | 7 (24.1) | 22 (75.8) | ||||
| AIDS diagnosis | 0.0005 | <0.0001 | 0.0191 | 0.2311 | |||||||||
| No | 3828 (59.0) | 206 (25.7) | 597 (74.4) | 290 (12.0) | 2131 (88.0) | 66 (21.5) | 241 (78.5) | 42 (14.1) | 255 (85.9) | ||||
| Yes | 2663 (41.0) | 261 (33.6) | 515 (66.4) | 248 (18.4) | 1102 (81.6) | 118 (29.3) | 285 (70.7) | 25 (18.7) | 109 (81.3) | ||||
| Provider's RWP client load | 0.0004 | 0.0015 | 0.1065 | 0.0196 | |||||||||
| 1–99 | 1690 (26.0) | 118 (25.4) | 346 (74.6) | 136 (17.7) | 634 (82.3) | 90 (28.4) | 227 (71.6) | 11 (7.9) | 128 (92.0) | ||||
| 100–199 | 1939 (29.9) | 199 (32.1) | 422 (67.9) | 125 (13.4) | 817 (86.7) | 63 (23.2) | 209 (76.8) | 18 (17.3) | 86 (82.7) | ||||
| ≥200 | 2510 (38.7) | 106 (26.8) | 290 (73.2) | 239 (12.8) | 1624 (87.2) | 21 (21.7) | 76 (78.4) | 30 (19.5) | 124 (80.5) | ||||
| Unknown | 352 (5.4) | 44 (44.9) | 54 (55.1) | 38 (19.4) | 158 (80.6) | 10 (41.7) | 14 (58.3) | 8 (23.5) | 26 (76.5) | ||||
| ACA | 0.0031 | <0.0001 | 0.0939 | 0.2488 | |||||||||
| No | 5521 (85.1) | 445 (30.6) | 1011 (69.4) | 484 (15.8) | 2587 (84.2) | 169 (26.9) | 459 (73.1) | 60 (16.4) | 306 (83.6) | ||||
| Yes | 970 (14.9) | 22 (17.9) | 101 (82.1) | 54 (7.7) | 646 (92.3) | 15 (18.3) | 67 (81.7) | 7 (10.8) | 58 (89.2) | ||||
| Psychosocial indices: mean and SD | |||||||||||||
| Mental health index | <0.0001 | <0.0001 | 0.6446 | 0.0001 | |||||||||
| Mean | 0.46 | 0.11 | 0.22 | −0.12 | −0.28 | −0.35 | 0.78 | 0.06 | |||||
| SD | 1.26 | 1.09 | 1.15 | 0.87 | 0.75 | 0.51 | 1.58 | 1.05 | |||||
| Alcohol/drug use index | <0.0001 | <0.0001 | 0.3804 | 0.0015 | |||||||||
| Mean | 0.50 | 0.08 | 0.18 | −0.14 | −0.26 | −0.22 | 0.51 | 0.05 | |||||
| SD | 1.57 | 1.11 | 1.28 | 0.70 | 0.12 | 0.48 | 1.61 | 1.07 | |||||
| Income/SES index | <0.0001 | <0.0001 | 0.2514 | <0.0001 | |||||||||
| Mean | 0.66 | 0.22 | 0.11 | −0.25 | 0.07 | −0.06 | 0.55 | −0.7 | |||||
| SD | 1.35 | 1.01 | 1.09 | 0.77 | 1.04 | 0.84 | 1.38 | 0.95 | |||||
| Neighborhood indices | |||||||||||||
| SES disadvantage index | 0.6888 | 0.0006 | 0.2810 | 0.1237 | |||||||||
| Low | 2207 (34.0) | 72 (29.3) | 174 (70.7) | 206 (12.6) | 1434 (87.4) | 15 (18.7) | 65 (81.3) | 31 (12.9) | 210 (87.1) | ||||
| Moderate | 1960 (30.2) | 119 (28.1) | 305 (71.9) | 151 (13.5) | 964 (86.5) | 87 (27.4) | 230 (72.6) | 17 (16.4) | 87 (83.7) | ||||
| High | 2324 (35.8) | 276 (30.4) | 633 (69.4) | 181 (17.8) | 835 (82.2) | 82 (26.2) | 231 (73.8) | 19 (22.1) | 67 (77.9) | ||||
| Residential instability index | 0.6184 | 0.0107 | 0.9060 | 0.2398 | |||||||||
| Low | 2111 (32.5) | 106 (27.7) | 277 (72.3) | 165 (12.0) | 1207 (88.0) | 69 (25.0) | 207 (75.0) | 11 (13.8) | 69 (86.3) | ||||
| Moderate | 2086 (32.1) | 200 (30.5) | 455 (69.5) | 184 (15.2) | 1026 (84.8) | 32 (26.5) | 89 (73.6) | 11 (11.0) | 89 (89.0) | ||||
| High | 2294 (35.3) | 161 (29.8) | 380 (70.2) | 189 (15.9) | 1000 (84.1) | 83 (26.5) | 230 (73.5) | 45 (17.9) | 206 (82.1) | ||||
| Racial/language homogeneity index | 0.1756 | 0.0023 | 0.7638 | 0.9260 | |||||||||
| Low | 2128 (32.8) | 25 (32.5) | 52 (67.5) | 244 (12.6) | 1698 (87.4) | 4 (19.1) | 17 (80.9) | 13 (14.8) | 75 (85.2) | ||||
| Moderate | 2163 (33.3) | 200 (31.9) | 427 (68.1) | 182 (15.1) | 1022 (84.9) | 25 (25.8) | 72 (74.2) | 38 (16.2) | 197 (83.8) | ||||
| High | 2200 (33.9) | 242 (27.7) | 633 (72.3) | 112 (17.9) | 513 (82.1) | 155 (26.2) | 437 (73.8) | 16 (14.8) | 92 (85.2) | ||||
Bivariate analysis conducted using chi-square test for categorical variable and Wilcoxon rank sum test for continuous variables.
Values in bold are statistically significant results.
ACA, Affordable Care Act; FPL, federal poverty level; IDU, injection drug use; MSM, men who have sex with men; NHB, non-Hispanic Black; NHW, non-Hispanic White; RWP, Ryan White Program; SD, standard deviation; SES, socioeconomic status.
Among the racial/ethnic groups, Hispanics were only significantly associated with neighborhood indices. Hispanics in low SES disadvantaged, low residential instability, and low racial/language homogeneity neighborhoods were more likely to achieve sustained viral suppression (Table 1). The psychosocial indices were all significantly associated with sustained viral suppression across all racial/ethnic groups.
Table 2 shows the results of the fitted multi-level models. Model 1 was the unconditional random intercept model, which calculated an ICC of 0.04. This indicates that 4% of the variation in sustained viral suppression is explained by the variation in neighborhoods or that the expected correlation between individuals in the same neighborhood is 4%. Model 2 shows the results of the model with only the individual-level variables. This shows that NHB and Haitians versus NHW, 18–49 versus ≥50, household incomes <100% FPL versus ≥200 FPL, unknown versus ≥200 provider's RWP client load, and having higher psychosocial indices value (poor mental health, high alcohol/drug use, and low income/SES) had lower odds of achieving sustained viral suppression (Table 2). Recipients of ACA had higher odds of sustained viral suppression than no ACA recipients (Table 2).
Table 2.
Multi-Level Mixed Effect Models for Sustained Viral Suppression for Ryan White HIV/AIDS Program Clients by Race/Ethnicity, in Miami, Florida, 2017
| Model 1: unconditional |
Model 2: individual-level factors |
Model 3: individual+neighborhood-level factors |
Model 4: all level factors+interactions |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Estimate | SE | Estimate | SE | Estimate | SE | Estimate | SE | |||
| Intercept variance | 0.1401 | 0.0391 | 0.00729 | 0.01154 | 0.00746 | 0.01231 | 0.01462 | 0.01566 | ||
| ICC | 0.04 | |||||||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | |||||
| Race/Ethnicity | ||||||||||
| NHB vs. NHW | 0.59 | (0.43–0.81) | 0.61 | (0.44–0.84) | 0.48 | (0.32–0.73) | ||||
| Hispanic vs. NHW | 0.95 | (0.69–1.31) | 0.91 | (0.66–1.25) | 0.82 | (0.55–1.21) | ||||
| Haitian vs. NHW | 0.46 | (0.32–0.67) | 0.47 | (0.32–0.69) | 0.54 | (0.30–0.96) | ||||
| Age group | ||||||||||
| 18–34 vs. 50+ | 0.39 | (0.32–0.47) | 0.39 | (0.32–0.47) | 0.39 | (0.32–0.46) | ||||
| 35–49 vs. 50+ | 0.57 | (0.49–0.67) | 0.58 | (0.49–0.68) | 0.58 | (0.49–0.68) | ||||
| Gender | ||||||||||
| Women vs. men | 1.11 | (0.93–1.33) | 1.11 | (0.93–1.33) | 1.11 | (0.93–1.33) | ||||
| US born | ||||||||||
| Yes vs. no | 0.88 | (0.72–1.08) | 0.89 | (0.73–1.10) | 0.89 | (0.72–1.09) | ||||
| Household income (FPL) | ||||||||||
| 100–199 vs. ≥200% | 0.76 | (0.62–0.93) | 0.75 | (0.61–0.93) | 0.76 | (0.62–0.93) | ||||
| <100 vs. ≥200% | 0.53 | (0.43–0.65) | 0.53 | (0.43–0.65) | 0.53 | (0.43–0.66) | ||||
| Mode of transmission | ||||||||||
| Heterosexual vs. MSM | 0.89 | (0.74–1.07) | 0.91 | (0.75–1.09) | 0.92 | (0.76–1.10) | ||||
| IDU vs. MSM | 0.78 | (0.56–1.09) | 0.78 | (0.56–1.08) | 0.78 | (0.56–1.10) | ||||
| AIDS diagnosis | ||||||||||
| Yes vs. no | 0.57 | (0.49–0.65) | 0.57 | (0.50–0.65) | 0.57 | (0.50–0.66) | ||||
| Provider's RWP client load | ||||||||||
| 1–99 vs. 200+ | 0.87 | (0.73–1.04) | 0.88 | (0.74–1.05) | 0.88 | (0.74–1.05) | ||||
| 100–199 vs 200+ | 0.92 | (0.78–1.08) | 0.92 | (0.78–1.09) | 0.93 | (0.78–1.09) | ||||
| Unknown vs. 200+ | 0.65 | (0.49–0.86) | 0.64 | (0.49–0.85) | 0.64 | (0.48–0.85) | ||||
| ACA | ||||||||||
| Yes vs. no | 1.37 | (1.09–1.73) | 1.36 | (1.08–1.71) | 1.36 | (1.08–1.72) | ||||
| Psychosocial indices | ||||||||||
| Mental health | 0.84 | (0.79–0.89) | 0.84 | (0.79–0.89) | 0.84 | (0.79–0.90) | ||||
| Alcohol/drug use | 0.87 | (0.82–0.93) | 0.87 | (0.82–0.93) | 0.88 | (0.83–0.93) | ||||
| Income/SES | 0.83 | (0.78–0.89) | 0.83 | (0.78–0.90) | 0.83 | (0.77–0.89) | ||||
| Neighborhood indices | ||||||||||
| SES disadvantage index | ||||||||||
| Moderate vs. low | 0.90 | (0.74–1.11) | 0.80 | (0.58–1.10) | ||||||
| High vs. low | 0.91 | (0.73–1.13) | 0.88 | (0.62–1.25) | ||||||
| Residential instability index | ||||||||||
| Moderate vs. low | 0.91 | (0.74–1.10) | 1.04 | (0.75–1.44) | ||||||
| High vs. low | 0.91 | (0.74–1.12) | 0.82 | (0.57–1.17) | ||||||
| Racial/language homogeneity index | ||||||||||
| Moderate vs. low | 0.82 | (0.67–1.01) | 0.89 | (0.58–1.35) | ||||||
| High vs. low | 0.85 | (0.68–1.05) | 0.93 | (0.61–1.41) | ||||||
| Interactions | ||||||||||
| SES disadvantage index | ||||||||||
| Low (NHB vs. NHW) | 0.39 | (0.20–0.74) | ||||||||
| Low (Hispanic vs. NHW) | 0.68 | (0.37–1.22) | ||||||||
| Low (Haitian vs. NHW) | 0.52 | (0.22–1.24) | ||||||||
| Moderate (NHB vs. NHW) | 0.53 | (0.27–1.02) | ||||||||
| Moderate (Hispanic vs. NHW) | 0.91 | (0.48–1.71) | ||||||||
| Moderate (Haitian vs. NHW) | 0.42 | (0.18–0.97) | ||||||||
| High (NHB vs. NHW) | 0.56 | (0.27–1.14) | ||||||||
| High (Hispanic vs. NHW) | 0.89 | (0.44–1.81) | ||||||||
| High (Haitian vs. NHW) | 0.71 | (0.28–1.79) | ||||||||
| Residential instability index | ||||||||||
| Low (NHB vs. NHW) | 0.54 | (0.25–1.16) | ||||||||
| Low (Hispanic vs. NHW) | 1.00 | (0.48–2.09) | ||||||||
| Low (Haitian vs. NHW) | 0.72 | (0.29–1.78) | ||||||||
| Moderate (NHB vs. NHW) | 0.31 | (0.15–0.65) | ||||||||
| Moderate (Hispanic vs. NHW) | 0.54 | (0.27–1.11) | ||||||||
| Moderate (Haitian vs. NHW) | 0.44 | (0.18–1.09) | ||||||||
| High (NHB vs. NHW) | 0.67 | (0.41–1.11) | ||||||||
| High (Hispanic vs. NHW) | 1.01 | (0.64–1.59) | ||||||||
| High (Haitian vs. NHW) | 0.50 | (0.24–1.04) | ||||||||
| Racial/language homogeneity index | ||||||||||
| Low (NHB vs. NHW) | 0.38 | (0.16–0.88) | ||||||||
| Low (Hispanic vs. NHW) | 0.88 | (0.43–1.78) | ||||||||
| Low (Haitian vs. NHW) | 0.61 | (0.16–2.37) | ||||||||
| Moderate (NHB vs. NHW) | 0.62 | (0.34–1.13) | ||||||||
| Moderate (Hispanic vs. NHW) | 1.01 | (0.55–1.86) | ||||||||
| Moderate (Haitian vs. NHW) | 0.67 | (0.31–1.46) | ||||||||
| High (NHB vs. NHW) | 0.49 | (0.26–0.93) | ||||||||
| High (Hispanic vs. NHW) | 0.62 | (0.32–1.19) | ||||||||
| High (Haitian vs. NHW) | 0.38 | (0.19–0.75) | ||||||||
Values in bold are statistically significant results.
ACA, Affordable Care Act; CI, confidence interval; FPL, federal poverty level; ICC, intraclass correlation; IDU, injection drug use; MSM, men who have sex with men; NHB, non-Hispanic Black; NHW, non-Hispanic White; OR, odds ratio; RWP, Ryan White Program; SE, standard error; SES, socioeconomic status.
Model 3 illustrates that none of the neighborhood indices was significantly associated with sustained viral suppression after adjusting for individual-level factors. Finally, model 4 shows the interaction between race/ethnicity and neighborhood indices and displays significant findings for the difference in race/ethnicity within each neighborhood tertile. In low SES disadvantaged neighborhoods, NHB had a lower odd of sustained viral suppression [adjusted odds ratio (aOR): 0.39; 95% confidence interval (CI): 0.20–0.74] compared to NHW in low SES disadvantaged neighborhoods. Haitians had lower odds of sustained viral suppression (aOR: 0.42; 95% CI: 0.18–0.97) in moderate SES disadvantaged neighborhoods when compared to NHW in moderate SES disadvantaged neighborhoods.
Lower odds of sustained viral suppression were observed for NHB (aOR: 0.31; 95% CI: 0.15–0.65) in moderate residential instability neighborhoods compared to NHW in same neighborhoods. Finally, NHB exhibited lower odds of sustained viral suppression in neighborhoods with low and high racial/language homogeneity (aOR: 0.38; 95% CI: 0.16–0.88) and (aOR: 0.38; 95% CI: 0.19–0.75), respectively, when compared to NHW in the same neighborhoods. In addition, Haitians residing in neighborhoods with high racial/language homogeneity had lower odds of sustained viral suppression (aOR: 0.49; 95% CI: 0.38–0.75) when compared to NHW in the same neighborhoods.
Discussion
Our study analyzed RWP data to measure the influence of neighborhood factors on the association between race/ethnicity and sustained viral suppression, finding several significant effects. Haitians and NHB exhibited lower odds of sustained viral suppression when compared to NHW. While neighborhood-level factors were overall not associated with sustained viral suppression, in neighborhoods with low or moderate SES disadvantage and residential instability, NHB and Haitians compared to NHW had lower odds of sustained viral suppression. Finally, in neighborhoods with low or high racial/language homogeneity, NHB and Haitians had lower odds of sustained viral suppression compared to NHW.
After controlling for individual-level factors, we did not observe an overall association between neighborhood-level factors and sustained viral suppression. Factors influencing sustained viral suppression might be different, relative to other HIV care outcomes such as retention in care and viral suppression. Maintaining a sustained viral load could prompt unique challenges, such as medication fatigue and drug resistance, which could attribute to poor adherence over time. In addition, perhaps individual-level factors are more important predictors than neighborhood factors for sustained viral suppression.
Notably, our study found differences in rates of sustained viral suppression within neighborhood characteristics for different racial/ethnic groups. Haitians and NHB in moderate and low SES disadvantage neighborhoods, respectively, had lower odds of sustained viral suppression when compared to NHW residing in same neighborhoods. These findings are in contrast with some prior literature that states residents in more disadvantaged neighborhoods have lower rates of HIV care outcomes,18 when compared to more affluent neighborhoods.
Even though SES is a strong predictor of poor health outcomes, racial disparities between Blacks and Whites are also prominent among wealthier populations.27,28 The presence of acute and chronic discrimination experienced by high-income Blacks residing in integrated neighborhoods27 could affect their stress responses and explain poor health outcomes observed in our study. Also, similar earning high-income minorities and Whites do not reside in comparable neighborhoods, due to lack of generational wealth, systemic discrimination in housing, and racial segregation,28 which further worsen differences in achieving optimal health outcomes.
Neighborhoods characterized by residential instability have been associated with poor health outcomes.5 Our study found that NHB in neighborhoods with moderate residential instability had lower odds of sustained viral suppression. We observed similar results in a post hoc analysis, where homicide was included instead of residential instability in the final model (since it was originally removed due to correlation). This showed that NHB in neighborhoods with moderate and high homicides had lower odds of achieving sustained viral suppression. Regardless of low or high neighborhoods' racial/language homogeneity, NHBs had lower odds of sustained viral suppression, while Haitians had lower odds of sustained viral suppression in neighborhoods with high racial/language homogeneity.
Racial residential segregation among racial/ethnic minorities leads to economic deprivation, which contributes to negative health consequences.29 In addition, irrespective of socioeconomic composition, Black neighborhoods are isolated and located in areas of severe disadvantage.28 Conversely, even though integration of neighborhoods might result in decreased bias and favorable perception of minorities due to increased exposure and interaction, Blacks face less discrimination and prejudice in homogenous neighborhoods.17 Discrimination and stigma, including enacted30 and cumulative enacted stigma,31 are posited to affect minority populations and lead to negative health consequences.32
Individual-level characteristics were also associated with lower odds of sustained viral suppression. Compared to Hispanics and Whites, Blacks have a lower rate of sustained viral suppression and experience more treatment-related disparities.15 Haitians carry a disproportionate burden of HIV among Caribbean-born Blacks in South Florida and are less likely to be linked and retained in continuous care,33 which subsequently affects sustained viral suppression. Younger adults experience worsening viral load status from first to last viral load,34 and have higher risk of viral rebound,35 which affects achieving sustained viral load. Poverty has also been associated with poor health outcomes,36 while having an AIDS diagnoses indicates higher viral load and belonging in transitional care.37
Our findings also indicate that clients who did not know their provider's name had poor rates of sustained viral suppression. Not knowing a physician's name might be an indication of the poor quality of relationship that the provider has with the patient, which is instrumental for treatment adherence, retention in care, and ultimately viral suppression.38 This could also be attributable to having a new provider or receiving care in a group clinic, where the client does not see the same provider during each visit. The presence of psychosocial factors such as depression, anxiety, domestic violence, drug/alcohol use on poor outcomes of medication adherence, retention in care, and viral suppression has been well noted in prior literature.39
In addition, income-related disparities such as lack of access to work and homelessness have also been associated with negative HIV care outcomes.40 Conversely, adults enrolled in ACA had a higher odd of sustained viral suppression. Clients with income <100% of the FPL and with an undocumented status are ineligible to receive ACA benefits. Hence, the ACA client population is less socially disadvantaged as they have higher incomes and access to comprehensive medical coverage through the ACA.41
This study is not without its limitations. First, due to the cross-sectional nature of this study, there might be residual confounding from variables that were not assessed in this study. Second, using ZIP codes to define neighborhoods can lead to spatial misclassification, as individuals encounter numerous neighborhoods beyond the geographic boundary of ZIP codes.42 Third, the use of indices to characterize neighborhoods might prohibit identifying pertinent neighborhood factors that might be associated with sustained viral suppression. Finally, all PWH in our study are engaged in care in the RWP in Miami-Dade county, which has a majority non-US-born and Hispanic population.
Consequently, the findings might not be generalizable to individuals who are not engaged in care, individuals who are not enrolled in the RWP or other programs such as Medicare/Medicaid and employer-based and private insurance, and residents of other geographic areas. In addition, we conducted a sensitivity analysis to assess the robustness of the exclusion of individuals with only one suppressed viral load in our definition of sustained viral suppression and found similar results.
Overall, we observed clients enrolled in the RWP had achieved high rates of sustained viral suppression. However racial/ethnic disparities existed across various individual and neighborhood characteristics, particularly among minorities who carry a higher disease burden. Moreover, racial/ethnic minorities and socioeconomically disadvantaged communities were disproportionally affected by COVID-19 pandemic,43 which may further exasperate HIV disparities and care outcomes among minorities in these communities. Hence, it is important to continue to study neighborhood-level factors that might be potential barriers to achieving a consistent viral load, especially for racial/ethnic minorities. Also, examining mechanisms for the association between neighborhood factors and racial/ethnic disparities in sustained viral suppression is imperative to implement structural and community-wide interventions.
Supplementary Material
Acknowledgments
The authors wish to gratefully acknowledge Carla Valle-Schwenk, RWP Administrator, and the entire Ryan White Part A Program in the Miami-Dade County Office of Management and Budget, for their active assistance, cooperation, and facilitation in the implementation of this study.
Authors' Contributions
R.D.: conceptualization, data analysis, writing-original, review, and editing, and final approval. M.J.T. and D.T.D.: conceptualization, data analysis, writing-review and editing, and final approval. T.L. and S.F.P.: data analysis, writing-review and editing, final approval. P.B., and R.A.L.: data curation, writing-review and editing, and final approval. D.M.S.: conceptualization, data analysis, supervision, writing-review and editing, and final approval.
Disclaimer
The content is solely the responsibility of the authors and does not necessarily represent the view of the National Institute of Health.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
This research is supported by awards from the National Institute on Minority Health and Health Disparities (NIMHD) of F31MD015234 from the National Institute of Health. This research was also supported, in part, by NIMHD grants (R01MD013563, R01MD012421, K01MD013770, U54MD012393, and R01MD013554); National Institute of Mental Health grant (R01MH112406), and Centers for Disease Control and Prevention grant (U01PS005202).
Supplementary Material
References
- 1. Diez Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci 2010;1186:125–145. [DOI] [PubMed] [Google Scholar]
- 2. Latkin CA, Williams CT, Wang J, Curry AD. Neighborhood social disorder as a determinant of drug injection behaviors: A structural equation modeling approach. Health Psychol 2005;24:96. [DOI] [PubMed] [Google Scholar]
- 3. Surratt HL, Kurtz SP, Levi-Minzi MA, Chen M. Environmental influences on HIV medication adherence: The role of neighborhood disorder. Am J Public Health 2015;105:1660–1666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Duncan DT, Kawachi I. Neighborhoods and Health. Oxford, UK: Oxford University Press; 2018. [Google Scholar]
- 5. Browning CR, Cagney KA. Neighborhood structural disadvantage, collective efficacy, and self-rated physical health in an urban setting. J Health Soc Behav 2002:383–399. [PubMed] [Google Scholar]
- 6. Kowitt SD, Donahue KE, Fisher EB, et al. How is neighborhood social disorganization associated with diabetes outcomes? A multilevel investigation of glycemic control and self-reported use of acute or emergency health care services. Clin Diabetes Endocrinol 2018;4:19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Marco M, Gracia E, Tomás JM, López-Quílez A. Assessing neighborhood disorder: Validation of a three-factor observational scale. Eur J Psychol Appl Legal Context 2015;7:81–89. [Google Scholar]
- 8. Ross CE, Mirowsky J. Neighborhood disadvantage, disorder, and health. J Health Soc Behav 2001:258–276. [PubMed] [Google Scholar]
- 9. Kerr JC, Valois RF, Siddiqi A, et al. Neighborhood condition and geographic locale in assessing HIV/STI risk among African American adolescents. AIDS Behav 2015;19:1005–1013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Cerdá M, Tracy M, Messner SF, et al. Misdemeanor policing, physical disorder, and gun-related homicide: A spatial analytic test of” broken-windows” theory. Epidemiology 2009:533–541. [DOI] [PubMed] [Google Scholar]
- 11. Williams DR, Collins C. Racial residential segregation: A fundamental cause of racial disparities in health. Public Health Rep 2001;116:404–416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Crepaz N, Tang T, Marks G, et al. Durable viral suppression and transmission risk potential among persons with diagnosed HIV infection: United States, 2012–2013. Clin Infect Dis 2016;63:976–983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Jefferson KA, Kersanske LS, Wolfe ME, et al. Place-based predictors of HIV viral suppression and durable suppression among men who have sex with men in New York city. AIDS Behav 2017;21:2987–2999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Marks G, Patel U, Stirratt MJ, et al. Single viral load measurements overestimate stable viral suppression among HIV patients in care: Clinical and public health implications. J Acquir Immune Defic Syndr 2016;73:205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Crepaz N, Dong X, Wang X, et al. Racial and ethnic disparities in sustained viral suppression and transmission risk potential among persons receiving HIV care—United States, 2014. Morb Mortal Wkly Rep 2018;67:113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Beer L, Bradley H, Mattson CL, et al. Trends in racial and ethnic disparities in antiretroviral therapy prescription and viral suppression in the United States, 2009–2013. J Acquir Immune Defic Syndr 2016;73:446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Crockett KB, Edmonds A, Johnson MO, et al. Neighborhood racial diversity, socioeconomic status, and perceptions of HIV-related discrimination and internalized HIV stigma among women living with HIV in the United States. AIDS Patient Care STDs 2019;33:270–281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wiewel EW, Borrell LN, Jones HE, et al. Neighborhood characteristics associated with achievement and maintenance of HIV viral suppression among persons newly diagnosed with HIV in New York City. AIDS Behav 2017;21:3557–3566. [DOI] [PubMed] [Google Scholar]
- 19. Eberhart MG, Yehia BR, Hillier A, et al. Individual and community factors associated with geographic clusters of poor HIV care retention and poor viral suppression. J Acquir Immune Defic Syndr 2015;69:S37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Rebeiro PF, Howe CJ, Rogers WB, et al. The relationship between adverse neighborhood socioeconomic context and HIV continuum of care outcomes in a diverse HIV clinic cohort in the Southern United States. AIDS Care 2018;30:1426–1434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Department of Health and Human Services. National HIV/AIDS Strategy for the United States 2017. Progress Report. Available at: https://files.hiv.gov/s3fs-public/NHAS_Progress_Report_2017.pdf (last accessed October 15, 2020).
- 22. Census Data. 2013-2017, American Community Survey 5 Years Estimate. Available at: https://data.census.gov/cedsci/ (Last accessed May 26, 2019).
- 23. Simply Analytics. Simply Analytics Data. Available at: https://simplyanalytics.com/ (Last accessed May 26, 2019).
- 24. Christopoulos KA, Das M, Colfax GN. Linkage and retention in HIV care among men who have sex with men in the United States. Clin Infect Dis 2011;52(Suppl_2):S214–S222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ulett KB, Willig JH, Lin H-Y, et al. The therapeutic implications of timely linkage and early retention in HIV care. AIDS Patient Care STDs 2009;23:41–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Trepka MJ, Sheehan DM, Dawit R, et al. Differential role of psychosocial, health care system and neighborhood factors on the retention in HIV care of women and men in the ryan white program. J Int Assoc Prov AIDS Care 2020;19:2325958220950087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Colen CG, Ramey DM, Cooksey EC, Williams DR. Racial disparities in health among nonpoor African Americans and Hispanics: The role of acute and chronic discrimination. Soc Sci Med 2018;199:167–180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Reardon SF, Fox L, Townsend J. Neighborhood income composition by household race and income, 1990–2009. Ann Am Acad Polit Soc Sci Med 2015;660:78–97. [Google Scholar]
- 29. Krieger N. Theories for social epidemiology in the 21st century: An ecosocial perspective. Int J Epidemiol 2001;30:668–677. [DOI] [PubMed] [Google Scholar]
- 30. Algarin AB, Sheehan DM, Varas-Diaz N, et al. Health care-specific enacted HIV-related stigma's association with antiretroviral therapy adherence and viral suppression among people living with HIV in Florida. AIDS Patient Care STDs 2020;34:316–326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Kalichman SC, Katner H, Banas E, et al. Cumulative effects of stigma experiences on retention in HIV care among men and women in the rural southeastern United States. AIDS Patient Care STDs 2020;34:484–490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Pellowski JA, Kalichman SC, Matthews KA, Adler N. A pandemic of the poor: Social disadvantage and the US HIV epidemic. Am Psychol 2013;68:197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Cyrus E, Dawson C, Fennie KP, et al. Disparity in retention in care and viral suppression for black Caribbean-born immigrants living with HIV in Florida. Int J Environ Res Public Health 2017;14:285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Crepaz N, Tang T, Marks G, Hall HI. Viral suppression patterns among persons in the United States with diagnosed HIV infection in 2014. Ann Intern Med 2017;167:446–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Palmer A, Gabler K, Rachlis B, et al. Viral suppression and viral rebound among young adults living with HIV in Canada. Medicine 2018;97:e10562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Cohen JK, Santos G-M, Moss NJ, et al. Regular clinic attendance in two large San Francisco HIV primary care settings. AIDS Care 2016;28:579–584. [DOI] [PubMed] [Google Scholar]
- 37. Olatosi BA, Probst JC, Stoskopf CH, et al. Patterns of engagement in care by HIV-infected adults: South Carolina, 2004–2006. AIDS 2009;23:725–730. [DOI] [PubMed] [Google Scholar]
- 38. Earl TR, Saha S, Lombe M, et al. Race, relationships, and trust in providers among black patients with HIV/AIDS. Soc Work Res 2013;37:219–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Sullivan KA, Messer LC, Quinlivan EB. Substance abuse, violence, and HIV/AIDS (SAVA) syndemic effects on viral suppression among HIV positive women of color. AIDS Patient Care STDs 2015;29(S1):S42–S48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Thakarar K, Morgan JR, Gaeta JM, et al. Homelessness, HIV, and incomplete viral suppression. J Health Care Poor Underserved 2016;27:145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. McManus KA, Rhodes A, Bailey S, et al. Affordable Care Act qualified health plan coverage: Association with improved HIV viral suppression for AIDS drug assistance program clients in a Medicaid nonexpansion state. Rev Infect Dis 2016;63:396–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Duncan DT, Kawachi I, Subramanian S, et al. Examination of how neighborhood definition influences measurements of youths' access to tobacco retailers: A methodological note on spatial misclassification. Am J Epidemiol 2014;179:373–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Karmakar M, Lantz PM, Tipirneni R. Association of social and demographic factors with COVID-19 incidence and death rates in the US. JAMA Network Open 2021;4:e2036462–e2036462. [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.