Abstract
Persons living with HIV (PLWH) with substance use disorders (SUD) remain a population difficult to engage in HIV care. Project HOPE (Hospital Visits as an Opportunity for Prevention and Engagement), a randomized controlled trial testing patient navigation with/without contingency management for PLWH with SUD, aimed to address this disparity. PLWH with SUD who were out of care were recruited from 11 hospitals across the United States from 2012 to 2014. Baseline socioeconomic factors, medical mistrust scores, and perceived discrimination surveys were collected at enrollment and evaluated for effects on linkage to care at the 6-month follow-up assessment. Linkage to care (attending an outpatient visit for HIV care), early linkage to care (attending first visit within 30 days of enrollment), and engagement in care (two HIV visits within the 6-month period) were determined by medical record abstraction, supplemented by self-report. Among 801 participants enrolled in the study (mean age 45 years, 33% women, and 73% African American), those who did not complete high school and with severe food insecurity had lower odds of being linked to care at 6 months. Those with high levels of medical mistrust, recent drug use, and who did not complete high school had lower odds of early linkage to care. Early linkage was associated with higher odds of engagement at 6 months and was mitigated by both patient navigator interventions (all p < .05). Addressing social determinants of health is critical to correct the disparity seen in HIV outcomes among PLWH with SUD. Identifying factors that alter the effect of interventions could help identify patients who would benefit most.
Keywords: HIV, substance use disorder, care continuum, linkage to care
Introduction
Asignificant percentage of persons living with HIV (PLWH) do not progress along the HIV care continuum to achieve to virologic suppression.1–3 Among persons in the United States diagnosed with HIV, ∼75% are linked to care within 1 month, 49% are retained in care, and a mere 51% are virally suppressed.4,5 Furthermore, according to the CDC National HIV Surveillance System, men, blacks, and those with substance use disorders (SUD) are less likely to have regular HIV outpatient care (as measured by obtaining ≥2 CD4 or viral load tests ≥3 months apart) compared to their counterparts.4
Evidence-based interventions to improve linkage to HIV care are limited.6–9 In brief, strengths-based case management proved efficacious in enhancing linkage to care within 6 months in ARTAS (Antiretroviral Treatment Access Study), a randomized trial to test the efficacy of the intervention, as well as in ARTAS II, a study to test the effectiveness of the intervention, both yielding approximately a 20% effect size relative to standard of care.10,11 Emerging evidence suggests linkage may be enhanced with more expedited care entry and rapid antiretroviral therapy (ART) initiation within days of diagnosis,12,13 with fully powered intervention trials utilizing this framework on the horizon.
Despite evidence-based, strengths-based case management becoming standard of care for linkage, PLWH with SUD continue to link to care at lower rates than those without,14–16—suggesting a need for novel approaches to linkage to care for this population. This is of particular significance given that up to half of PLWH in the United States have a co-occurring SUD.17,18 Furthermore, PLWH with SUD do not uniformly fail to link to care, suggesting a need for a more granular understanding of factors that affect linkage within this population. For example, a secondary analysis of the ARTAS data showed that participation in substance use treatment was associated with faster linkage to HIV care.19
Project HOPE (Hospital Visits as an Opportunity for Prevention and Engagement) (CTN-0049; NCT01612169) aimed to address this disparity for HIV outcomes for PLWH with SUD.20 The study enrolled 801 participants who were randomized to one of three intervention arms for 6 months: (1) patient navigator, (2) patient navigator with contingency management, and (3) treatment as usual. Although the primary outcome, viral suppression at 12 months, was not significantly different between the intervention arms, there were improved rates of viral suppression among both intervention arms at 6 months.
The aim of the current analysis was to investigate predictors for linkage to care at 6 months among hospitalized patients with poorly controlled HIV and co-occurring SUD. We hypothesized that low socioeconomic factors, medical mistrust, and having experienced discrimination within the health care setting decreased the likelihood of being linked to care at 6 months and that these individuals would be the most responsive to the interventions evaluated in Project HOPE.
Methods
Study population and data sources
The current study is a secondary analysis of data collected from the 801 study participants in Project HOPE. Project HOPE had a three parallel-group design and recruited from 11 hospitals across the United States from July 2012 through January 2014. Participants were living with HIV, had co-occurring substance use, uncontrolled HIV infection, and were hospitalized at the time of recruitment. Participants were enrolled while hospitalized at one of the 11 participating hospitals and were randomized to a 6-month intervention consisting of (1) patient navigation, (2) patient navigation with contingency management, or (3) treatment as usual.
Methods for the parent study have been previously reported.20 After providing written consent and being screened for eligibility, participants were enrolled, provided blood specimens, and completed a social and behavioral assessment through a computer-assisted personal interview at the time of enrollment, which occurred during their hospitalization. They were then randomly assigned in equal proportions to one of the three intervention arms described previously. These interventions were continued for 6 months with follow-up assessments being performed at 6 months (end of intervention) and 12 months. Outcomes pertaining to attendance of HIV care visits were determined by medical record abstraction, supplemented by self-report when the medical record was missing. The study was approved by the Institutional Review Boards of all participating institutions.
Study variables
Socioeconomic variables
Age was divided into a dichotomous categorical variable, <45 years old and ≥45 years old. Socioeconomic variables were obtained from the computer-assisted personal interview at the time of enrollment and included income, educational level, insurance status, homelessness, and food insecurity. Income was defined as the individual's personal annual income based on self-report and was divided into two tiers (≤$10,000 and >$10,000 per year, approximating the federal poverty limit). Educational level was defined as the individual's highest grade or level of school completed, or the highest degree received based on self-report, divided into low (not achieving a high school diploma/GED) or high (graduating from high school or attaining a higher degree). Insurance status was defined as the individual's report at the time of taking the baseline survey and was divided into three groups (no; yes; don't know). Housing was defined as within the 6 months before answering the survey by self-report and was divided into two groups, unstable (homeless, living in a shelter, transitionally housed, or staying with family/friends) or stable (permanent long-stay hotel, HIV group home, drug treatment facility, halfway house, renting a house/apartment, owning a home, or other). Food scarcity was based on responses to a previously validated survey and participants were divided into quartiles.21
Perceived discrimination and medical mistrust variables
Perceived discrimination within the health care community and medical mistrust were obtained from the computer-assisted personal interview at the time of enrollment and were defined as follows. Perceived discrimination was based on survey data (yes/no/don't know/refuse to answer) assessing whether or not the participant had perceived discrimination in the following categories, with at least one affirmative answer qualifying as the individual perceiving discrimination: HIV status, gender, sexual orientation or practices, race/ethnicity, and drug use. Medical mistrust was based on responses to a previously validated survey and divided into two categories, present (score >36) or absent (score ≤36), using a cutoff score of 36 based on the original validation study.22
Outcome variables
Time to linkage to care was defined as early (first appointment for HIV care was completed within 30 days of study enrollment) or late (first appointment for HIV care was completed after 30 days of study enrollment). Linkage to care was defined as attending a first follow-up outpatient visit for HIV care. Engagement in care was defined as having two clinic visits for HIV within the 6-month time period.5 Attendance of HIV care visits for all outcome variables was determined by medical record abstraction, supplemented by self-report when the medical record was missing.
Statistical analyses
All statistical analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC). Descriptive statistics were calculated for covariates across the entire population as well as for each individual treatment arm as mean values with standard deviation (SD), or frequency and proportion within categories. All variables with more than two categories were treated as ordinal categorical variables.
A bivariable analysis using logistic regression was performed to calculate the odds ratios (OR) and 95% confidence intervals (CI) between individual characteristics and the primary outcome of interest, linkage to care at 6 months. These analyses were performed on the treatment as usual arm to increase generalizability beyond the study population. A multivariable analysis was performed using complete case analysis to assess the effect measure modification of the baseline characteristics on the intervention arms compared to treatment as usual on the outcome of interest, linkage to care at 6 months. Variables considered for inclusion into the multivariable model had p-values of <.05 in the bivariable analysis. In addition, age, gender, education, income, baseline CD4 count, and baseline HIV-1 RNA were also considered for inclusion regardless of p-value. These variables were selected before statistical analysis to be of particular interest to evaluate for their effect measure modification on progression along the HIV care continuum. Iterative likelihood ratio tests (LRT) were performed to determine whether variables would be kept in the multivariable model and were repeated until all remaining interaction variables were considered to be statistically significant with LRT p-values <.05. Adjusted OR (aOR) for linkage to care at 6 months were then obtained for each variable in the model, adjusting for all other variables included in the analysis. Participants with missing data for any variable in the multivariable model were not included in the multivariable analysis as this was a complete case analysis.
A bivariable analysis using logistic regression was performed to assess the OR between individual characteristics and linkage to care within 30 days of enrollment. This was performed for the overall cohort as well as stratified by treatment arm. Bivariable analysis using logistic regression was then performed to assess the OR comparing early (linked to care within 30 days of enrollment) to late (linked to care after 30 days of enrollment) on engagement in care at 6 months. Participants who did not link to care by 6 months were excluded from this analysis as it would be impossible for an individual to attend a second visit without having attended a first. This analysis was performed for the overall cohort as well as stratified by treatment arm. A forest plot was created with JMP Pro, version 14 (SAS Institute, Cary, NC).
Sensitivity analysis
Missing values were infrequent, with two notable exceptions. There were 74 of the 801 study participants (9%) who were lost to follow-up, the majority of whom died before follow-up, for whom all outcome data were missing. These participants were excluded from all statistical analyses. There were 236 of the 801 study participants (29%) who chose not to answer the survey on annual income. The primary bivariable and multivariable analyses were performed using complete case analysis, including only participants without missing data. A sensitivity analysis was performed to evaluate the study participants with missing income data. Baseline demographics were obtained for participants with missing income data, which was then compared to the demographics of participants with low and high incomes using Cochran-Mantel-Haenszel tests.
Results
Study population and data sources
All 801 study participants from the original Project HOPE were included in this study's analysis. Baseline participant characteristics are shown in Table 1. The mean age at enrollment was 45 years, ∼33% of the study participants were women, and the majority were black (73%). Seventy-seven percent of participants met eligibility criteria for drug use and 59% for heavy alcohol use. Thirty-two percent admitted to ever using injection drugs, with 18% having done so within the 12 months before study enrollment. In addition, 33% had no form of health insurance, 40% had not completed high school, and 73% earned ≤$10,000 per year. At the time of study enrollment, 29% reported perceived discrimination within the health care setting and 16% had scores suggesting medical mistrust with scores greater than 36 (mean score: 28.7, SD: 7.78). Median CD4 was 109 cells/μL (Q1, Q3: 29–242 cells/μL) and median HIV viral load was 52,826 copies/mL (Q1, Q3: 5,180 to 199,218 copies/mL).
Table 1.
Baseline Characteristics Among Project Hospital Visits as an Opportunity for Prevention and Engagement for HIV-Infected Drug Users Study Participants (N = 801)
| Demographic | Overall (N = 801) | Navigation (N = 266) | Navigation + incentives (N = 271) | Usual treatment (N = 264) |
|---|---|---|---|---|
| Age (years) | 45 (9.98) | 45 (9.85) | 45 (10.04) | 44 (10.09) |
| Female | 261 (33) | 87 (33) | 94 (35) | 80 (31) |
| Race/ethnicity | ||||
| Black | 579 (72) | 195 (73) | 194 (72) | 190 (72) |
| White | 97 (12) | 27 (10) | 41 (15) | 29 (11) |
| Hispanic | 84 (10) | 26 (10) | 25 (9) | 33 (13) |
| Other | 38 (5) | 16 (6) | 10 (4) | 12 (5) |
| Missing | 3 (0.4) | 2 (0.8) | 1 (0.4) | 0 (0) |
| Eligibility due toa | ||||
| Drug use | 613 (77) | 213 (80) | 210 (77) | 190 (72) |
| Alcohol use | 471 (59) | 146 (55) | 155 (57) | 170 (64) |
| Ever IDU | 260 (32) | 90 (34) | 85 (31) | 85 (32) |
| IDU in the past 12 months | 147 (18) | 50 (19) | 51 (19) | 46 (17) |
| Insurance status | ||||
| Some | 534 (67) | 176 (66) | 182 (67) | 176 (67) |
| None | 261 (33) | 88 (33) | 88 (32) | 85 (32) |
| Unknown | 6 (0.8) | 2 (1) | 1 (0.4) | 3 (1) |
| Not completed high school | 319 (40) | 117 (44) | 105 (39) | 97 (37) |
| Unstable housing status | 357 (45) | 126 (47) | 116 (43) | 115 (44) |
| Annual income ≤$10,000 (N = 565) | 414 (73) | 143 (76) | 140 (70) | 131 (74) |
| Household food insecurity | ||||
| None (0) | 327 (41) | 110 (41) | 109 (40) | 108 (41) |
| Mild (0–2) | 89 (11) | 30 (11) | 31 (11) | 28 (11) |
| Moderate (2–11) | 191 (24) | 57 (21) | 72 (27) | 62 (23) |
| Severe (>11) | 194 (24) | 69 (26) | 59 (22) | 66 (25) |
| Perceived health care discrimination (N = 798) | 232 (29) | 76 (29) | 73 (27) | 83 (32) |
| Medical Mistrust Score | 28.7 (7.78) | 28.8 (8.11) | 28.1 (7.42) | 29.1 (7.78) |
| >36 | 128 (16) | 50 (19) | 30 (11) | 48 (18) |
| ≤36 | 673 (84) | 216 (81) | 241 (89) | 216 (82) |
| HIV-1 RNA copies/mLb | 52,826 (5,180, 199,218) | 54,028.5 (5,770, 192,246) | 53,133 (4,683, 199,218) | 49,445.5 (7,275, 222,502) |
| ≤200 copies/mL | 87 (11) | 30 (11) | 28 (10) | 29 (11) |
| >200 copies/mL | 714 (89) | 236 (89) | 243 (90) | 235 (89) |
| CD4 countb | 109 (29, 242) | 95.5 (27, 240) | 123 (35, 259) | 105.5 (24.5, 238) |
| ≤200 | 534 (67) | 174 (65) | 179 (66) | 181 (69) |
| >200 | 267 (33) | 92 (35) | 92 (34) | 83 (31) |
All values are listed as N (%) or as mean (standard deviation).
Patients could be eligible for the study by more than one criterion.
HIV-1 RNA copies/mL and CD4 count are reported as median (Q1, Q3).
IDU, injection drug use.
Linkage to care at 6 months
Overall, 565 (80%) of participants were linked to care within 6 months of study enrollment. A bivariable analysis was performed for each baseline characteristic to evaluate its effect on linkage to care at 6 months among the participants randomized to usual treatment (Table 2). Participants with high levels of medical mistrust (OR: 0.34; 95% CI: 0.13–0.91) and those that met eligibility criteria for drug use (OR: 0.49; 95% CI: 0.26–0.91) were found to have lower odds of linkage to care at 6 months. Individuals who had recently engaged in injection drug use within the 12 months before enrollment were more likely to be linked to care at 6 months (OR: 2.81; 95% CI: 1.38–5.75).
Table 2.
Factors Affecting Linkage to Care at 6 Months Among the Control Arm (N = 264).
| Predictor | Usual treatment |
|
|---|---|---|
| OR (95% CI) | p | |
| Age | ||
| <45 Years | 0.96 (0.53–1.74) | .89 |
| ≥45 Years | Reference | |
| Gender | ||
| Female | 0.62 (0.31–1.21) | .16 |
| Male | Reference | |
| Ethnicity | ||
| Black | 0.90 (0.33–2.42) | .83 |
| Hispanic | 0.86 (0.25–3.03) | .82 |
| Othera | 1.42 (0.27–7.52) | .68 |
| White | Reference | |
| Eligibility due to drug use | 0.49 (0.26–0.91) | .02 |
| Eligibility due to alcohol | 1.28 (0.67–2.43) | .56 |
| IDU ever | 1.82 (0.98–3.37) | .06 |
| IDU in past 12 months | 2.81 (1.38–5.75) | .005 |
| Insurance | ||
| None | 1.44 (0.78–2.66) | .25 |
| Some | Reference | |
| Education | ||
| Low | 0.78 (0.42–1.45) | .43 |
| High | Reference | |
| Housing | ||
| Unstable | 0.60 (0.32–1.10) | .10 |
| Stable | Reference | |
| Income | ||
| Low | 0.73 (0.31–1.72) | .47 |
| High | Reference | |
| Food insecurity | ||
| Mild (0–2) | 1.06 (0.40–2.82) | .91 |
| Moderate (2–11) | 1.36 (0.66–2.71) | .43 |
| Severe (>11) | 0.43 (0.17–1.07) | .07 |
| None (0) | Reference | |
| Perceived discrimination | ||
| Yes | 0.93 (0.49–1.77) | .82 |
| No | Reference | |
| Medical mistrust | ||
| >36 | 0.34 (0.13–0.91) | .03 |
| ≤36 | Reference | |
| Viral load | ||
| >200 copies/mL | 0.61 (0.25–1.52) | .29 |
| ≤200 copies/mL | Reference | |
| CD4 | ||
| ≤200 | 1.75 (0.89–3.43) | .10 |
| >200 | Reference | |
Bolded values indicate statistical significance with p < 0.05.
Other ethnicities include American Indian/Alaska Native, Native Hawaiian/Pacific Islander, and Other.
Age, gender, income, CD4 count, HIV viral load, medical mistrust, eligibility due to drug use, and injection drug use within the 12 months before enrollment were considered for inclusion into the multivariable model to evaluate their effect modification on the intervention arms compared to the usual treatment arm on linkage to care at 6 months. There were no statistically significant effect measure modifiers, with an LRT evaluating the aforementioned variables resulting a p-value of .09. Although not statistically significant with an LRT p-value of .26, there was a trend toward improved responses to the intervention arms when stratifying by income. Participants with low income (patient navigator aOR: 2.68; 95% CI: 1.44–5.00; and patient navigator with incentives aOR: 6.09; 95% CI: 2.93–12.66) appeared to be more sensitive to intervention arms than those with high income (patient navigator aOR: 1.81; 95% CI: 0.46–7.07; and patient navigator with incentives aOR: 2.17; 95% CI: 0.60–7.80). Not completing high school (aOR: 0.42, 95% CI: 0.26–0.69) and severe food insecurity (aOR: 0.46, 95% CI: 0.25–0.83) were associated with lower odds of being linked to care at 6 months after adjusting for the variables listed above. Participants randomized to the intervention arms (patient navigator aOR: 2.60; 95% CI: 1.50–4.51; and patient navigator with incentives aOR: 4.44, 95% CI: 2.44–8.10) were found to have higher odds of being linked to care at 6 months in the adjusted model. Results are shown in Table 3.
Table 3.
Multivariable Analysis
| Predictor | Usual treatment |
|
|---|---|---|
| aOR (95% CI)a | p | |
| Age | ||
| <45 Years | 1.42 (0.87–2.32) | .17 |
| ≥45 Years | Reference | |
| Gender | ||
| Female | 0.81 (0.48–1.37) | .44 |
| Male | Reference | |
| Eligibility due to drug use | 1.41 (0.78–2.53) | .26 |
| Insurance | ||
| None | 1.46 (0.84–2.53) | .18 |
| Some | Reference | |
| Education | ||
| Low | 0.42 (0.26–0.69) | <.001 |
| High | Reference | |
| Housing | ||
| Unstable | 0.62 (0.37–1.03) | .06 |
| Stable | Reference | |
| Income | ||
| Low | 0.97 (0.53–1.77) | .92 |
| High | Reference | |
| Food insecurity | ||
| Mild (0–2) | 0.54 (0.25–1.19) | .12 |
| Moderate (2–11) | 1.15 (0.58–2.27) | .69 |
| Severe (>11) | 0.46 (0.25–0.83) | .01 |
| None (0) | Reference | |
| Medical mistrust | ||
| >36 | 1.20 (0.66–2.21) | .55 |
| ≤36 | Reference | |
| Viral load | ||
| >200 Copies/mL | 0.74 (0.35–1.56) | .42 |
| ≤200 Copies/mL | Reference | |
| CD4 | ||
| ≤200 | 1.21 (0.73–2.01) | .45 |
| >200 | Reference | |
| Treatment intervention | ||
| PN | 2.60 (1.50–4.51) | <.001 |
| PN + I | 4.44 (2.44–8.10) | <.0001 |
| Usual care | Reference | |
Predictors for linkage to care at 6 months, Project HOPE (N = 490).
Bolded values indicate statistical significance with p < 0.05.
Adjusted for age, gender, eligibility due to drug use, insurance, education, housing, income, food insecurity, medical mistrust, viral load, CD4 count, and treatment intervention.
HOPE, Hospital Visits as an Opportunity for Prevention and Engagement; PN, patient navigation; PN + I, patient navigation with incentives.
Early linkage to care
There were 249 participants (35%) who linked to care within 30 days of study enrollment. A bivariable analysis evaluating factors affecting early linkage to care among all Project HOPE participants is shown in Figure 1. Participants not completing high school (OR: 0.67; 95% CI: 0.49–0.93), with high levels of medical mistrust (OR: 0.59; 95% CI: 0.38–0.93), and those that met eligibility criteria for drug use (OR: 0.69; 95% CI: 0.48–0.98) had lower odds of being linked to care within 30 days of study enrollment in the overall cohort. Participants with insurance at time of enrollment were found to have higher odds of being linked to care at 30 days (OR: 1.50; 95% CI: 1.08–2.08). Full results, including breakdown by intervention arm, are shown in Supplementary Table S1.
FIG. 1.
A forest plot of predictors for early linkage to care among Project HOPE participants who linked to care (N = 704). HOPE, Hospital Visits as an Opportunity for Prevention and Engagement; IDU, injection drug use.
The effect that early linkage to care had on engagement in care at 6 months is shown in Table 4. Participants who linked to care within 30 days of enrollment were found to have higher odds of being engaged in care at 6 months compared with participants who linked to care between 30 days and 6 months (OR: 4.03; 95% CI: 2.75–5.91). This effect was seen to be consistent across all three study arms.
Table 4.
Effect of Time to Linkage to Care on Engagement in Care at 6 Months (N = 486)
| Predictor | Proportion engaged in care at 6 months | OR (95% CI) | p |
|---|---|---|---|
| Overall (N = 486) | |||
| Time to linkage | |||
| ≤30 Days | 149/249 | 4.03 (2.75–5.91) | <.0001 |
| >30 Days | 64/237 | — | — |
| Navigation (N = 158) | |||
| Time to linkage | |||
| ≤30 Days | 49/86 | 3.01 (1.56–5.82) | .001 |
| >30 Days | 22/72 | — | — |
| Navigation + incentives (N = 188) | |||
| Time to linkage | |||
| ≤30 Days | 62/105 | 3.35 (1.82–6.15) | .0001 |
| >30 Days | 25/83 | — | — |
| Usual care (N = 140) | |||
| Time to linkage | |||
| ≤30 Days | 38/58 | 7.27 (3.40–15.54) | <.0001 |
| >30 Days | 17/82 | — | — |
Bolded values indicate statistical significance with p < 0.05.
Sensitivity analysis
Baseline characteristics for participants with missing income data are shown in Supplementary Table S2. The mean age at enrollment was 45 years (SD: 9.88 years), 39% were women, and 78% were black. Thirty-one percent admitted to ever using injection drugs, with 18% having done so within the 12 months before enrollment. In addition, 42% had no form of insurance, and 45% had not completed high school. At the time of study enrollment, 26% reported having perceived discrimination within the health care setting and 15% had scores suggesting medical mistrust. Median CD4 was 113 cells/μL, and median HIV viral load was 56,477 copies/mL.
The baseline demographics between participants with low, high, and missing income are compared in Supplementary Table S3. Baseline demographics were mostly similar among the three income groups with a few exceptions. Participants with missing income were more likely to be women and be unstably housed, and were less likely to have completed high school than those with low income, and they were less likely to have insurance than those who reported income (all p-value <.05).
Discussion
This work has identified several factors that affect linkage to care among PLWH with SUD. High levels of medical mistrust and recent substance use were associated with lower odds of linkage to care at both 30 days and 6 months, while low education was associated with lower odds of linkage to care at 30 days. Although many of these factors have previously been shown to be associated with poorer HIV care outcomes in other populations, this study was one of the largest to evaluate these factors in PLWH with SUD, a particularly vulnerable cohort.23–25—Not completing high school significantly lowered a participant's odds of linking to care by 6 months by 50%. Although low education has been seen in other populations to be associated with poorer HIV outcomes,26,27 this finding is inconsistent, with some studies not finding an association.28 This suggests the need for continued efforts to identify interventions to improve linkage to care in people with low education, and the need to consider providing educational opportunities for PLWH.
Unlike other analyses, we did not find an association between income level and linkage to care at 6 months.29–31–In this work, making less than $10,000 per year was not associated with statistically significant lower odds of linkage to care at 6 months. In addition, income was not found to be a significant effect measure modifier when evaluated in the multivariable analysis. These findings were surprising in light of work done in other disciplines showing that programs with financial incentives may be most beneficial in improving clinical outcomes among individuals with lower incomes.32
Participants with high levels of food insecurity were found to have 50% lower odds of linking to care at 6 months compared with those without food insecurity. This has been seen previously in other populations, but has not been studied extensively among PLWH with co-occurring SUD.33–35 This work augments the growing body of literature that highlights the interaction between food insecurity and health care outcomes.
Stratification by intervention arm shows that high levels of medical mistrust were associated with lower odds of early linkage to care (within 30 days) only among the treatment as usual group (OR: 0.34; 95% CI: 0.13–0.91), highlighting the integral role patient navigators have in overcoming medical mistrust in this population. Previous reports have linked medical mistrust to poorer progression along the HIV care continuum at several steps, including linkage to care36 and adherence to ART,37,38 but none has specifically focused on PLWH with SUD. Not only does our study support the importance of medical mistrust in the delivery of health care to PLWH with SUD but it also suggests that this association may be mitigated by patient navigators. It is possible that patient navigators, using a strengths-based case management approach and motivational interviewing techniques, with or without financial incentives, may have reduced the negative impact medical mistrust had on early linkage to care in this study.
There is overwhelming evidence of the benefits for early diagnosis, linkage to care, and viral suppression using ART.13,39–45 Our findings support the rapid entry concept, showing that patients who linked to care early (≤30 days) had four-times higher odds of being engaged in care at 6 months when compared with those who linked to care later (>30 days). It is unclear whether the strong effect of early linkage to care is due to an intrinsically more motivated study participant or whether the act of linking more quickly causes a patient to remain engaged in care at higher rates. In addition, this result is subject to lead-time bias, with individuals who were late in linking to care having less time available to return for a second visit. Ongoing implementation studies evaluating rapid linkage to care aim to answer these important questions.
This work had a number of limitations. First, patient characteristics used in our work were collected at the time of enrollment primarily by self-report. Although the analysis in this work made the assumption that these factors were static throughout the duration of the study, it is possible that some may have changed over time. In addition, as a secondary analysis, the original study was not powered to address the outcomes evaluated in this work. Since some analyses are not based on prior hypotheses, our multiple comparisons may identify some associations that could be spurious and must be confirmed by future targeted studies. Missing data were also a limitation. Although income was not found to be statistically significant, nearly one-third of participants did not report their income data. This limitation was addressed by performing a sensitivity analysis which found individuals who did not report their income were largely similar to those who provided income information. Finally, although overall educational level was evaluated in this work, there were no data available specifically on health literacy, which is an important topic warranting investigation in future research.
Our findings highlight factors such as poverty, educational level, and medical mistrust being associated with poor linkage to care and suggest an opportunity for targeted interventions with the PLWH with SUD population. It emphasizes the importance of social determinants of health in affecting health outcomes as well as health care access. Finally, this suggests the need for future research to build prediction tools to improve interventions, using targeted approaches for deploying limited human resource support, and novel interventions to overcome specific barriers for the ultimate improvement of progression along the HIV care continuum for a population that has continued to have poor outcomes.
Supplementary Material
Acknowledgments
The authors thank the Project HOPE study sites for their work during the initial study. The authors also thank Rui Duan for his help in obtaining the dataset for analysis and Mitchel Klein for his guidance in analysis.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
Supported, in part, by the NIH/NIAID (Emory CFAR P30 AI050409) and the NIDA CTN (3U10DA013720-06S3). N.A.S. received funding through the National Center for Advancing Translational Sciences of the National Institutes of Health (grant nos. UL1 TR002378, TL1 TR002382).
Supplementary Material
References
- 1. Vital signs: HIV prevention through care and treatment—United States. MMWR Morb Mortal Wkly Rep 2011;60:1618–1623 [PubMed] [Google Scholar]
- 2. Burns DN, Dieffenbach CW, Vermund SH: Rethinking prevention of HIV type 1 infection. Clin Infect Dis 2010;51:725–731 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gardner EM, McLees MP, Steiner JF, Del Rio C, Burman WJ: The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clin Infect Dis 2011;52:793–800 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bradley H, Hall HI, Wolitski RJ, et al. : Vital signs: HIV diagnosis, care, and treatment among persons living with HIV—United States, 2011. MMWR Morb Mortal Wkly Rep 2014;63:1113–1117 [PMC free article] [PubMed] [Google Scholar]
- 5. Centers for Disease Control and Prevention: Understanding the HIV Care Continuum. Available at https://www.cdc.gov/hiv/pdf/DHAP_Continuum.pdf (2014), accessed May2018
- 6. Higa DH, Crepaz N, Mullins MM, Prevention Research Synthesis P: Identifying best practices for increasing linkage to, retention, and re-engagement in HIV medical care: Findings from a systematic review, 1996–2014. AIDS Behav 2016;20:951–966 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Marks G, Gardner LI, Craw J, Crepaz N: Entry and retention in medical care among HIV-diagnosed persons: A meta-analysis. AIDS 2010;24:2665–2678 [DOI] [PubMed] [Google Scholar]
- 8. Centers for Disease Control and Prevention: Linkage to, retention in, and re-engagement in HIV care (LRC) Chapter. 2018. Available at: https://www.cdc.gov/hiv/research/interventionresearch/compendium/lrc/ accessed May, 2018
- 9. Thompson MA, Mugavero MJ, Amico KR, et al. : Guidelines for improving entry into and retention in care and antiretroviral adherence for persons with HIV: Evidence-based recommendations from an International Association of Physicians in AIDS Care panel. Ann Intern Med 2012;156:817–833, W-284, W-285, W-286, W-287, W-288, W-289, W-290, W-291, W-292, W-293, W-294. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Craw JA, Gardner LI, Marks G, et al. : Brief strengths-based case management promotes entry into HIV medical care: Results of the antiretroviral treatment access study-II. J Acquir Immune Defic Syndr 2008;47:597–606 [DOI] [PubMed] [Google Scholar]
- 11. Gardner LI, Metsch LR, Anderson-Mahoney P, et al. : Efficacy of a brief case management intervention to link recently diagnosed HIV-infected persons to care. AIDS 2005;19:423–431 [DOI] [PubMed] [Google Scholar]
- 12. Labhardt ND, Ringera I, Lejone TI, et al. : Effect of offering same-day ART vs usual health facility referral during home-based hiv testing on linkage to care and viral suppression among adults with HIV in Lesotho: The CASCADE Randomized Clinical Trial. JAMA 2018;319:1103–1112 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Rosen S, Maskew M, Fox MP, et al. : Initiating antiretroviral therapy for HIV at a patient's first clinic visit: The RapIT Randomized Controlled Trial. PLoS Med 2016;13:e1002015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bhatia R, Hartman C, Kallen MA, Graham J, Giordano TP: Persons newly diagnosed with HIV infection are at high risk for depression and poor linkage to care: Results from the Steps Study. AIDS Behav 2011;15:1161–1170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Giordano TP, Visnegarwala F, White AC Jr., et al. : Patients referred to an urban HIV clinic frequently fail to establish care: Factors predicting failure. AIDS Care 2005;17:773–783 [DOI] [PubMed] [Google Scholar]
- 16. Yehia BR, Stewart L, Momplaisir F, et al. : Barriers and facilitators to patient retention in HIV care. BMC Infect Dis 2015;15:246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Kuo I, Golin CE, Wang J, et al. : Substance use patterns and factors associated with changes over time in a cohort of heterosexual women at risk for HIV acquisition in the United States. Drug Alcohol Depend 2014;139:93–99 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Santos GM, Rapues J, Wilson EC, et al. : Alcohol and substance use among transgender women in San Francisco: Prevalence and association with human immunodeficiency virus infection. Drug Alcohol Rev 2014;33:287–295 [DOI] [PubMed] [Google Scholar]
- 19. Gardner LI, Marks G, Strathdee SA, et al. : Faster entry into HIV care among HIV-infected drug users who had been in drug-use treatment programs. Drug Alcohol Depend 2016;165:15–21 [DOI] [PubMed] [Google Scholar]
- 20. Metsch LR, Feaster DJ, Gooden L, et al. : Effect of patient navigation with or without financial incentives on viral suppression among hospitalized patients with HIV infection and substance use: A Randomized Clinical Trial. JAMA 2016;316:156–170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Coates J, Swindale A, Bilinsky P: Household Food Insecurity Access Scale (HFIAS) for Measurement of Household Food Access: Indicator Guide (v. 3). In: Food and Nutrition Technical Assistance Project AfED, ed. Washington, DC, 2007
- 22. Thompson HS, Valdimarsdottir HB, Winkel G, Jandorf L, Redd W: The Group-Based Medical Mistrust Scale: Psychometric properties and association with breast cancer screening. Prev Med 2004;38:209–218 [DOI] [PubMed] [Google Scholar]
- 23. Erickson A, Becker M, Shaw S, Kasper K, Keynan Y: Substance use and its impact on care outcomes among HIV-infected individuals in Manitoba. AIDS Care 2015;27:1168–1173 [DOI] [PubMed] [Google Scholar]
- 24. Gwadz M, de Guzman R, Freeman R, et al. : Exploring how substance use impedes engagement along the HIV Care Continuum: A Qualitative Study. Front Publ Health 2016;4:62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Meyer JP, Althoff AL, Altice FL: Optimizing care for HIV-infected people who use drugs: Evidence-based approaches to overcoming healthcare disparities. Clin Infect Dis 2013;57:1309–1317 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. McFall AM, Mehta SH, Srikrishnan AK, et al. : Getting to 90: Linkage to HIV care among men who have sex with men and people who inject drugs in India. AIDS Care 2016;28:1230–1239 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Sobrino-Vegas P, Rodriguez-Urrego J, Berenguer J, et al. : Educational gradient in HIV diagnosis delay, mortality, antiretroviral treatment initiation and response in a country with universal health care. Antivir Ther 2012;17:1–8 [DOI] [PubMed] [Google Scholar]
- 28. Cohen GM, Werner L, Gengiah S, Naidoo K: Role of education in HIV clinical outcomes in a tuberculosis endemic setting. J Int Assoc Provid AIDS Care 2014;13:402–408 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. 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. Clin Infect Dis 2016;63:396–403 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Sprague C, Simon SE: Understanding HIV care delays in the US South and the role of the social-level in HIV care engagement/retention: A qualitative study. Int J Equity Health 2014;13:28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Walcott M, Kempf MC, Merlin JS, Turan JM: Structural community factors and sub-optimal engagement in HIV care among low-income women in the Deep South of the USA. Cult Health Sex 2016;18:682–694 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Lasser KE, Quintiliani LM, Truong V, et al. : Effect of patient navigation and financial incentives on smoking cessation among primary care patients at an Urban Safety-Net Hospital: A Randomized Clinical Trial. JAMA Intern Med 2017;177:1798–1807 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Spinelli MA, Frongillo EA, Sheira LA, et al. : Food insecurity is associated with poor HIV outcomes among women in the United States. AIDS Behav 2017;21:3473–3477 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Weiser SD, Fernandes KA, Brandson EK, et al. : The association between food insecurity and mortality among HIV-infected individuals on HAART. J Acquir Immune Defic Syndr 2009;52:342–349 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Weiser SD, Yuan C, Guzman D, et al. : Food insecurity and HIV clinical outcomes in a longitudinal study of urban homeless and marginally housed HIV-infected individuals. AIDS 2013;27:2953–2958 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Beer L, Fagan JL, Garland P, et al. : Medication-related barriers to entering HIV care. AIDS Patient Care STDs 2012;26:214–221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Bogart LM, Wagner G, Galvan FH, Banks D: Conspiracy beliefs about HIV are related to antiretroviral treatment nonadherence among african american men with HIV. J Acquir Immune Defic Syndr 2010;53:648–655 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Kalichman SC, Eaton L, Kalichman MO, Grebler T, Merely C, Welles B: Race-based medical mistrust, medication beliefs and HIV treatment adherence: Test of a mediation model in people living with HIV/AIDS. J Behav Med 2016;39:1056–1064 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Philbin MM, Tanner AE, DuVal A, et al. : HIV testing, care referral, and linkage to care intervals affect time to engagement in care for newly diagnosed hiv-infected adolescents in 15 adolescent medicine clinics in the United States. J Acquir Immune Defic Syndr 2016;72:222–229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Hall HI, Tang T, Johnson AS, Espinoza L, Harris N, McCray E: Timing of linkage to care after HIV diagnosis and time to viral suppression. J Acquir Immune Defic Syndr 2016;72:e57–e60 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Halperin J, Katz M, Pathmanathan I, et al. : Early HIV diagnosis leads to significantly decreased costs in the first 2 years of HIV care in an Urban Charity Hospital in New Orleans. J Int Assoc Provid AIDS Care 2017;16:527–530 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Koenig SP, Dorvil N, Devieux JG, et al. : Same-day HIV testing with initiation of antiretroviral therapy versus standard care for persons living with HIV: A randomized unblinded trial. PLoS Med 2017;14:e1002357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Onyeajam DJ, Eke R, Stephens TG, Duffus WA. Time to linkage to care and viro-immunologic parameters of individuals diagnosed before and after the 2006 HIV testing recommendations. South Med J 2013;106:257–266 [DOI] [PubMed] [Google Scholar]
- 44. Pilcher CD, Ospina-Norvell C, Dasgupta A, et al. : The effect of same-day observed initiation of antiretroviral therapy on HIV viral load and treatment outcomes in a US Public Health Setting. J Acquir Immune Defic Syndr 2017;74:44–51 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Robertson M, Laraque F, Mavronicolas H, Braunstein S, Torian L: Linkage and retention in care and the time to HIV viral suppression and viral rebound—New York City. AIDS Care 2015;27:260–267 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.