Abstract
Background
Evidence-based strategies are needed to address the increasing healthcare system demands for people aging with HIV (PAWH). HIV case management models integrating access to primary care, specialty care, and supportive services represent a promising approach for healthy aging.
Methods
We emulated a target trial comparing mortality outcomes through 36-month follow-up among people with HIV (PWH) enrolled in the New York City revised Care Coordination Program (CCR) versus a contemporaneous group of PWH who met CCR eligibility criteria but did not enroll in the CCR. To estimate the effect of the CCR on all-cause and cause-specific mortality, we used hazards regression with inverse probability of treatment weights controlling for clinical and demographic differences.
Results
From August 2018 to March 2021, 3040 individuals enrolled in the CCR. At enrollment, 43% of CCR enrollees were aged 50 years or older, 27% had a CD4 count <200 cells/µL, and 19% were diagnosed prior to 1996. In adjusted analyses, CCR enrollees had lower all-cause mortality (adjusted hazard ratio [aHR]: 0.71 [0.60–0.84]), driven primarily by reductions in HIV-related deaths. The effects on all-cause and HIV-related mortality were consistent among PAWH and younger PWH, among those lacking evidence of viral suppression (VS), and among those with lower CD4 counts (<350 cells/µL).
Conclusions
The CCR appears to reduce all-cause and HIV-related mortality including among PWH at highest risk for poor outcomes. These findings suggest that integrated care coordination models—linking primary care, specialty care, and psychosocial services—offer survival benefits and multiple health outcomes improvements among PWH.
Keywords: care coordination, durable viral suppression, mortality, people aging with HIV, viral load suppression
Relative to usual care, the revised Care Coordination Program shows evidence of reducing all-cause and HIV-related mortality among populations at risk for poor outcomes, including people aging with HIV, younger people with HIV (PWH), and PWH lacking evidence of viral suppression.
The life expectancy of people with HIV (PWH) has dramatically increased due to improvements in HIV care and availability of antiretroviral treatment (ART) [1, 2]. Early treatment and improved survival on ART has led to an increased number of people aging with HIV (PAWH), generally defined as those aged 50 years or older [3, 4]. As the median age of PWH has increased, the promotion of healthy aging has become a priority of HIV care [5]. Among the 1.1 million PWH in the United States, the number of people aged 50 years or older is predicted to increase from 54% at the end of 2021 to 63%–75% by 2040 [6]. Risk of major contributors to mortality such as chronic illness is greater among PWH than people without HIV [7]. Evidence-based strategies are needed to address the growing complexity of caring for PAWH [8, 9].
Care coordination or other models integrating access to primary and specialty care, and including psychosocial and supportive services for PWH, represent one approach to promoting longevity. The New York City (NYC) Ryan White HIV/AIDS Program Part A (RWPA) HIV Care Coordination Program (CCP) combines multiple evidence-based strategies that aim to meet the medical and social service needs of clients: HIV case management by interdisciplinary teams, patient navigation, and health promotion and coaching around HIV self-management skills [10, 11]. In observational effectiveness analyses, the CCP improved short and long-term viral suppression (VS) relative to usual care for those consistently unsuppressed [12, 13]. On the basis of CCP effectiveness findings, the Centers for Disease Control and Prevention designated the CCP as an evidence-based intervention for linkage, retention, and re-engagement [14] and evidence-based structural intervention for VS [15, 16]. However, the CCP has not been evaluated for impact on mortality [12, 13, 16–18]. Furthermore, the broader evidence base on case management effectiveness has focused on more proximal outcomes (care engagement, adherence, and VS) than mortality.
In 2018, the NYC Health Department implemented a revised Care Coordination Program (CCR) to strengthen the intervention and reduce barriers to its delivery. The objective of our analysis was to examine the impact of the CCR on all-cause and HIV-specific mortality and among all persons enrolled and specific populations at risk for mortality, including PAWH.
METHODS
Care Coordination Intervention
The NYC Department of Health and Mental Hygiene (NYC Health Department) launched a multicomponent CCP in December 2009, using federal RWPA funding [17]. The CCP model implemented in 2009 combined several best practices, including a “medical home” approach; multidisciplinary care team communication; shared decision-making through case conferences; patient navigation, including accompaniment to primary care visits; ART adherence support; and a structured health promotion curriculum [17, 19, 20]. Based on effectiveness findings [12, 13, 16] and identified barriers, the NYC Health Department and their community partner on RWPA, the HIV Health and Human Services Planning Council of New York, outlined a redesign of the CCP in 2017–2018. The revised program (CCR) also aimed to meet the medical and social service needs of clients through interdisciplinary case management, patient navigation, health education, and promotion of HIV self-management skills, but incorporated features intended to increase program reach, flexibility, feasibility, and sustainability. The modifications included adding resources for HIV self-management; supporting videoconferencing options for directly observed therapy; replacing per-member-per-day reimbursement with a fee-for-service reimbursement model; emphasizing identification and recruitment of individuals with documented clinical need (eg, unsuppressed viral load [VL]); and removing enrollment tracks that prescribed fixed visit frequencies based on assessed client need, in favor of greater flexibility for client-centered care [11]. Intervention details have been published [10, 11].
CCR enrollment was open to PWH who were 18+ or emancipated minors, eligible for local RWPA services (based on residence within the New York eligible metropolitan area and household income ≤435% of federal poverty level [FPL] for enrollments through February 2019 and ≤500% of FPL for enrollments after February 2019), and demonstrating need for care coordination (based on recent HIV diagnosis, care or treatment initiation, absences from care, unsuppressed VL, hepatitis-C coinfection, pregnancy, ART regimen change, missed appointments, or provider-assessed risk of care interruption or treatment failure).
Data Sources and Population
We merged longitudinal, population-based surveillance and programmatic data sources. Revised Care Coordination Program programmatic data were drawn from the NYC Health Department's Electronic System for HIV/AIDS Reporting and Evaluation (eSHARE). Ryan White HIV/AIDS Program Part A service providers have been contractually required to submit programmatic data through eSHARE since March 2011. Using eSHARE data, we identified all PWH enrolled in the CCR at any point from 01 August 2018, to 31 March 2021.
The HIV Surveillance Registry contains demographic (sex, race/ethnicity, and age at diagnosis) and transmission category information on all diagnoses of HIV reported in NYC, as well as comprehensive HIV-related laboratory reporting (including all CD4 and VL dates and results for tests ordered by NYC providers). Vital status information, including cause of death, is updated through regular matches with local and national death data to ascertain deaths occurring in and outside of NYC. For mortality outcomes, we used Registry data reported through 31 March 2024, which allowed for complete cause of death reporting through 31 March 2022.
Research Ethics
The Institutional Review Board at the NYC Health Department approved this study (Protocol 18-009). For these secondary analyses of de-identified data, we received a waiver of informed consent under 45 CFR 46.116(d)(2).
Target Trial Specification and Emulation
This observational analysis emulated a hypothetical randomized experiment, the target trial, to determine whether the CCR reduced mortality relative to usual care [21]. We first specified the protocol of a target trial to estimate the effect of the intervention on mortality. The hypothetical target trial protocol is summarized in Supplementary Table 1. Next, we emulated the design and intention-to-treat (ITT) analysis of the target trial.
Contemporaneous Comparison Population
To assess the impact of the CCR on mortality, we identified a contemporaneous registry-based control group of PWH who met eligibility criteria for enrollment into the CCR but did not enroll in the CCR. We identified controls in a 3-step process [22]. First, we identified eligible persons and their eligibility windows, or ranges of time from 1 August 2018, to 31 March 2021, when an individual met eligibility criteria for the CCR. Registry-based eligibility criteria are defined in Supplementary Table 2.
Second, because we expect CCR effects to vary based on HIV treatment history, we ascertained baseline treatment status prior to each month of eligibility. The 4 baseline treatment status groups were defined in terms of diagnosis or VS in the 12 months prior to eligibility: (1) newly diagnosed, (2) consistently suppressed (≥2 VLs ≥90 days apart and all VLs ≤200 copies/mL), (3) lacking evidence of VS (all VLs reported >200 copies/mL or no VL report), or (4) inconsistently suppressed (≥1 VL ≤200 copies/mL, but not all VLs ≤200 copies/mL).
Finally, for each CCR enrollee within each baseline treatment status group, we randomly sampled 1 control from among those identified as eligible . We matched the distribution of starts of follow-up for eligible yet unenrolled controls with the distribution of enrollment dates (month and year) in the CCR. In this way, each selected control was assigned a pseudo-enrollment date based on a month and year in which they were eligible for CCR enrollment.
Measures and Variable Definitions
Outcomes
We examined (1) all-cause, (2) HIV-related, and (3) all non-HIV-related deaths. We also examined the 2 most frequently occurring non-HIV-related causes: (3a) major cardiovascular disease and (3b) malignant neoplasms. We classified as “HIV-related” deaths for which the underlying cause recorded in the death certificate was coded with one of the International Classification of Diseases Tenth Revision (ICD-10) codes representing HIV disease, B20-B24, O98.8 or R75 [23, 24]. These codes correspond with opportunistic infections (eg, pneumocystis carinii pneumonia and cytomegaloviral infection), malignant neoplasms (Kaposi sarcoma and Burkitt lymphoma), other specific diseases or infections associated with HIV infection, and unspecified conditions consistent with HIV disease or AIDS. We classified as “non-HIV-related” all underlying causes that did not represent HIV-related causes. Major cardiovascular disease includes codes representing I00–I78 and non-HIV-related malignant neoplasms representing C00–C97 [25].
Covariates
We used the Registry to ascertain age, gender, race/ethnicity, HIV transmission category, VL, CD4 count, HIV diagnosis year, and baseline HIV treatment status. For VL and CD4 count, we used the most recent measure in the year before enrollment/pseudo-enrollment. Age 50+ has been proposed as the cut-point for defining PAWH, in recognition that PWH 50+ have poorer immune recovery after starting ART and greater burden of comorbidities relative to younger PWH [3, 4]. HIV diagnosis year was categorized in relation to HIV treatment eras with diagnoses <1996 representing the pre-combination ART (cART) era, 1996–2005 representing the cART era, and 2006 and later representing the single-tablet era.
Residential ZIP-code-level HIV prevalence and poverty were assigned based on each participant's most recent available ZIP code of residence in the Registry prior to enrollment/pseudo-enrollment. ZIP-code-level poverty data were obtained from the American Community Survey.
For CCR participants, we used eSHARE data to ascertain insurance status, housing status, recent drug use, mental health (reported diagnosis of depression/anxiety, psychosis, or posttraumatic stress disorder [PTSD]), and status. These variables were chosen because they were collected by CCR providers in the intake assessment and therefore could be used to identify clients in need of closer follow-up.
Statistical Analysis
We used an ITI approach with stabilized inverse probability of treatment weights (IPTWs) to examine the effects of the CCR on mortality outcomes [26]. For the IPTW, we fit logistic regression models with CCR enrollment as the outcome for each baseline treatment status group, and we a priori–selected confounders as independent variables (Supplementary Table 3). We generated hazard ratios to estimate the effect of the CCR on all-cause and HIV-specific mortality. Individuals contributed time from enrollment to the earlier of 31 March 2022, 36 months (1095 days) postenrollment/pseudo-enrollment, or death. We looked at overall CCR effectiveness and effectiveness within subgroups based on treatment status, age (<50 and ≥50), race/ethnicity, diagnosis year, and baseline CD4 count. To estimate the effect of the CCR on mortality, we applied the stabilized IPTW to separate models for each outcome with a term for CCR enrollment, the covariate of interest, and an interaction term.
To assess the robustness of the results to unmeasured or uncontrolled confounding, we present the E-value for the all-cause mortality effect estimate [27]. The E-value is “the minimum strength of association, on the risk ratio scale, that an unmeasured confounder would need to have with both the treatment and the outcome to fully explain away a specific treatment-outcome association, conditional on the measured covariates.” [27]
For CCR participants, we also examined factors collected at intake that could increase hazards of mortality. The model included the variable of interest (eg, insurance status) and adjusted for age, gender, transmission risk, race, treatment era, and baseline CD4 and VS.
Analyses were conducted in SAS V9.4.
RESULTS
From 1 August 2018, to 31 March 2021, there were 4693 CCR enrollees, of whom 4581 (98%) completed a program intake assessment and 3040 (65%) met the surveillance-based eligibility criteria (Figure 1). There were 63 021 case records in the Registry that met the surveillance-based CCR eligibility criteria for at least 1 month during the study period. Of these, 58 418 (93%) were not enrolled in the CCR or the original CCP from December 2009 to March 2021. From this group, 3040 individuals were randomly selected and assigned a pseudo-enrollment date.
Figure 1.
Participant flow diagram.
At enrollment, 43% of CCR participants were aged 50 years or older, 27% had a CD4 count <200 cells/µL, and 19% were diagnosed <1996 (Table 1). In the eligible non-CCR group, 41% were aged 50 years or older, 22% had a CD4 count <200 cells/µL, and 16% were diagnosed <1996. The IPTWs had a mean of 1 and standard deviation between 0.4 and 1. The CCR and non-CCR groups were balanced on all characteristics after applying the IPTWs.
Table 1.
Characteristics of Non-CCR and CCR Participants—August 2018 to March 2021, New York City
| Total | ||
|---|---|---|
| Non-CCR | CCR | |
| Total | 3040 (100.00) | 3040 (100.00) |
| Age | ||
| 0–17 | 14 (0.46) | 3 (0.10) |
| 18–29 | 518 (17.04) | 511 (16.81) |
| 30–49 | 1262 (41.51) | 1216 (40.00) |
| 50–64 | 1010 (33.22) | 1104 (36.32) |
| 65+ | 236 (7.76) | 206 (6.78) |
| Age 50+ | ||
| No | 1794 (59.01) | 1730 (56.91) |
| Yes | 1246 (40.99) | 1310 (43.09) |
| Gendera | ||
| Cisgender female | 811 (26.68) | 945 (31.09) |
| Cisgender male | 2149 (70.69) | 1984 (65.26) |
| TGNCNB | 80 (2.63) | 111 (3.65) |
| Race/Ethnicity | ||
| API and other | 85 (2.80) | 74 (2.43) |
| Black | 1579 (51.94) | 1613 (53.06) |
| Hispanic | 1005 (33.06) | 1228 (40.39) |
| White | 371 (12.20) | 125 (4.11) |
| Transmission risk | ||
| Heterosexual | 659 (21.68) | 731 (24.05) |
| IDU and MSM-IDU | 424 (13.95) | 583 (19.18) |
| MSM | 1255 (41.28) | 1027 (33.78) |
| Otherb | 163 (5.36) | 212 (6.97) |
| Unknown | 539 (17.73) | 487 (16.02) |
| HIV prevalence/povertyc | ||
| High poverty/high HIV prevalence | 1014 (33.36) | 1239 (40.76) |
| High poverty/low HIV prevalence | 505 (16.61) | 622 (20.46) |
| Low poverty/high HIV prevalence | 283 (9.31) | 197 (6.48) |
| Low poverty/low HIV prevalence | 972 (31.97) | 939 (30.89) |
| Unknown | 266 (8.75) | 43 (1.41) |
| Baseline CD4d | ||
| <200 | 650 (21.38) | 804 (26.45) |
| 200–349 | 465 (15.30) | 546 (17.96) |
| 350–499 | 454 (14.93) | 395 (12.99) |
| 500+ | 1037 (34.11) | 777 (25.56) |
| Unknown | 434 (14.28) | 518 (17.04) |
| Baseline VLd | ||
| <200 | 1187 (39.05) | 848 (27.89) |
| 200+ | 1441 (47.40) | 1705 (56.09) |
| Unknown | 412 (13.55) | 487 (16.02) |
| Treatment era based on diagnosis year | ||
| Prior cART (<1996) | 477 (15.69) | 588 (19.34) |
| cART era (1996–2005) | 949 (31.22) | 976 (32.11) |
| Single-tablet era (2006+) | 1614 (53.09) | 1476 (48.55) |
| Treatment statuse | ||
| Consistently suppressed | 279 (9.18) | 279 (9.18) |
| Inconsistently suppressed | 1108 (36.45) | 1108 (36.45) |
| Lacking evidence of VS | 1179 (38.78) | 1179 (38.78) |
| Newly diagnosed | 474 (15.59) | 474 (15.59) |
| Insuredf | ||
| Uninsured | … | 183 (6.02) |
| Insured | … | 2705 (88.98) |
| Missing | … | 152 (5.00) |
| Stably housedf | ||
| No | … | 681 (22.40) |
| Yes | … | 2271 (74.70) |
| Missing | … | 88 (2.89) |
| Recent drug usef | ||
| No | … | 2023 (66.55) |
| Yes | … | 670 (22.04) |
| Missing | … | 347 (11.41) |
| Depressionf | ||
| No | … | 1894 (62.30) |
| Yes | … | 1146 (37.70) |
| Psychosisf | ||
| No | … | 2574 (84.67) |
| Yes | … | 466 (15.33) |
| PTSDf | ||
| No | … | 2835 (93.26) |
| Yes | … | 205 (6.74) |
| levelf | ||
| High school or lower | … | 2019 (66.41) |
| >High school | … | 924 (30.39) |
| Missing | … | 97 (3.19) |
Abbreviations: API, Asian, Pacific Islander; cART, combination antiretroviral treatment; CCR, revised Care Coordination Program; IDU, injection drug use; MSM, men who have sex with men; TGNCNB, transgender, gender nonconforming or nonbinary; VL, viral load; >, viral suppression.
aFewer than 5 participants were transgender men across both groups.
bIncludes people with perinatal infection and transgender people with sexual contact.
cResidential ZIP code-level HIV prevalence and poverty were assigned based on each participant's most recent available ZIP code of residence in the Registry and classified as high (prevalence greater than median in the year prior to enrollment/pseudo-enrollment) or low (prevalence below median in the year prior to enrollment/pseudo-enrollment).
dIn the year prior to enrollment/pseudo-enrollment.
eBaseline treatment status definitions: lacking evidence of VS: no VL results <=200 copies/mL in the year prior to enrollment/pseudo-enrollment; newly diagnosed: diagnosed with HIV in the year prior to enrollment/pseudo-enrollment; inconsistently suppressed: at least 1 VL result <=200 copies/mL and 1 VL result >200 copies/mL in the year prior to enrollment/pseudo-enrollment; consistently suppressed: all VL results <=200 copies/mL in the year prior to enrollment/pseudo-enrollment.
fCaptured in the programmatic database and data not available for non-CCR comparison population.
All-cause and HIV-related Mortality
Among CCR participants, the all-cause and HIV-related mortality rates were 3.31 per 100 person years (PY) and 0.70 per 100 PY, respectively. Among controls, the all-cause and HIV-related mortality rates were 3.69 and 1.17, respectively, per 100 PY. In adjusted analyses, the CCR participants had lower all-cause and HIV-related mortality relative to their counterparts receiving usual care, and these effects were consistent among PAWH and younger PWH, among those lacking evidence of VS in the year before enrollment/pseudo-enrollment, among those diagnosed in later years (2006+), and among those with baseline CD4 counts <200 and 200–349 cells/µL (Table 2). Among PWH diagnosed from 1996 to 2005 and among PWH who were inconsistently suppressed in the preenrollment/pseudo-enrollment year, CCR participants had lower HIV-related but not all-cause mortality.
Table 2.
All-cause and HIV-related Mortality Rates by Baseline Treatment Status Group, Age at Enrollment, Diagnosis Year, and Baseline CD4 Count Among PWH Enrolled in the CCR Relative to Usual Care (Non-CCR)—New York City, 01 August 2018 to 31 March 2022
| Total Persons | All-cause Mortality | HIV-related Mortality | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Non-CCR | CCR | Non-CCR | CCR | … | … | Non-CCR | CCR | … | … | |||||||
| Persons | PY | Persons | PY | Deaths | Crude Rate (Per 100 PY) | Deaths | Crude Rate (Per 100 PY) | Crude HR | Adj HR | Deaths | Crude Rate (Per 100 PY) | Deaths | Crude Rate (Per 100 PY) | Crude HR | Adj HR | |
| Total | 3040 | 8390 | 3040 | 8580 | 310 | 3.69 | 284 | 3.31 | 0.90 (0.76, 1.05) | 0.71 (0.60, 0.84) | 98 | 1.17 | 60 | 0.7 | 0.60 (0.43, 0.83) | 0.43 (0.31, 0.60) |
| Treatment status | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … |
| Newly diagnosed | 474 | 1399 | 474 | 1403 | 8 | 0.57 | 9 | 0.64 | 1.12 (0.43, 2.90) | 0.73 (0.28, 1.91) | 3 | 0.21 | 2 | 0.14 | 0.66 (0.11, 3.98) | 0.34 (0.05, 2.09) |
| Consistently suppressed | 279 | 787 | 279 | 771 | 28 | 3.56 | 39 | 5.06 | 1.41 (0.87, 2.29) | 1.20 (0.75, 1.93) | 8 | 1.02 | 5 | 0.65 | 0.63 (0.21, 1.94) | 0.45 (0.15, 1.39) |
| Lacking evidence of VS | 1179 | 3085 | 1179 | 3302 | 178 | 5.77 | 119 | 3.60 | 0.63 (0.50, 0.79) | 0.55 (0.43, 0.69) | 64 | 2.07 | 33 | 1 | 0.48 (0.32, 0.74) | 0.39 (0.25, 0.60) |
| Inconsistently suppressed | 1108 | 3120 | 1108 | 3104 | 96 | 3.08 | 117 | 3.77 | 1.22 (0.93, 1.60) | 0.86 (0.66, 1.13) | 23 | 0.74 | 20 | 0.64 | 0.87 (0.48, 1.59) | 0.54 (0.29, 1.00) |
| Age 50+ at enrollment | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … |
| No | 1794 | 5130 | 1730 | 5013 | 100 | 1.95 | 95 | 1.90 | 0.97 (0.73, 1.29) | 0.75 (0.56, 0.99) | 36 | 0.7 | 27 | 0.54 | 0.77 (0.47, 1.26) | 0.47 (0.28, 0.78) |
| Yes | 1246 | 3260 | 1310 | 3567 | 210 | 6.44 | 189 | 5.30 | 0.82 (0.68, 1.00) | 0.70 (0.57, 0.86) | 62 | 1.9 | 33 | 0.93 | 0.49 (0.32, 0.74) | 0.41 (0.26, 0.63) |
| Diagnosis year | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … |
| <1996 | 477 | 1238 | 588 | 1574 | 90 | 7.27 | 104 | 6.61 | 0.91 (0.68, 1.20) | 0.82 (0.62, 1.09) | 27 | 2.18 | 27 | 1.72 | 0.79 (0.46, 1.34) | 0.75 (0.43, 1.28) |
| 1996–2005 | 949 | 2563 | 976 | 2704 | 117 | 4.56 | 116 | 4.29 | 0.94 (0.73, 1.21) | 0.80 (0.62, 1.04) | 32 | 1.25 | 21 | 0.78 | 0.62 (0.36, 1.08) | 0.47 (0.27, 0.84) |
| 2006+ | 1614 | 4589 | 1476 | 4303 | 103 | 2.24 | 64 | 1.49 | 0.66 (0.49, 0.91) | 0.52 (0.38, 0.71) | 39 | 0.85 | 12 | 0.28 | 0.33 (0.17, 0.63) | 0.20 (0.10, 0.40) |
| Baseline CD4 | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … | … |
| <200 | 650 | 1632 | 804 | 2141 | 134 | 8.21 | 144 | 6.73 | 0.82 (0.65, 1.04) | 0.71 (0.56, 0.90) | 48 | 2.94 | 42 | 1.96 | 0.67 (0.44, 1.01) | 0.55 (0.36, 0.83) |
| 200–349 | 465 | 1273 | 546 | 1541 | 54 | 4.24 | 46 | 2.99 | 0.70 (0.48, 1.04) | 0.57 (0.38, 0.85) | 16 | 1.26 | 6 | 0.39 | 0.31 (0.12, 0.79) | 0.24 (0.09, 0.68) |
| 350–499 | 454 | 1287 | 395 | 1135 | 36 | 2.80 | 31 | 2.73 | 0.98 (0.60, 1.58) | 0.78 (0.48, 1.26) | 10 | 0.78 | 4 | 0.35 | 0.45 (0.14, 1.45) | 0.41 (0.13, 1.30) |
| 500+ | 1037 | 3035 | 777 | 2239 | 35 | 1.15 | 47 | 2.10 | 1.81 (1.17, 2.81) | 1.48 (0.95, 2.30) | 9 | 0.30 | 7 | 0.31 | 1.05 (0.39, 2.82) | 0.59 (0.21, 1.67) |
| Unknown | 434 | 1163 | 518 | 1525 | 51 | 4.39 | 16 | 1.05 | 0.24 (0.14, 0.42) | 0.27 (0.15, 0.49) | 15 | 1.29 | 1 | 0.07 | 0.05 (0.01, 0.39) | 0.05 (0.01, 0.44) |
Bold text indicates confidence intervals do not contain the null value of 1. The 4 treatment status groups were defined in terms of diagnosis or viral suppression in the 12 m prior to enrollment/pseudo-enrollment: (1) newly diagnosed, (2) consistently suppressed (≥2 VLs ≥90 days apart and all VLs ≤200 copies/mL), (3) lacking evidence of VS (all VLs reported >200 copies/mL or lacking any VL report), or (4) inconsistently suppressed (≥1 VL ≤200 copies/mL, but not all VLs ≤200 copies/mL). Diagnosis years were categorized in relation to HIV treatment eras with <1996 representing diagnoses prior to combination antiretroviral treatment (cART), 1996–2005 representing the cART era, and diagnoses since 2005 representing the single-tablet era. We classified as “HIV-related” deaths for which the underlying cause recorded in the death certificate was coded with one of the ICD-10 codes representing HIV disease, B20-B24, 098.7, or R75.18. These codes correspond with opportunistic infections (eg, pneumocystis carinii pneumonia and cytomegaloviral infection), malignant neoplasms (Kaposi sarcoma and Burkitt lymphoma), other specific diseases or infections associated with HIV infection, and unspecified conditions consistent with HIV disease or AIDS.
Abbreviations: Adj, adjusted; CCR, revised Care Coordination Program; HR, hazards ratio; non-CCR, persons eligible yet not enrolled in the CCR; PWH, persons with HIV; PY, person years; VS, viral suppression.
Non-HIV-related Mortality
Non-HIV-related mortality was lower among CCR participants than controls, although the difference was not statistically significant (aHR: 0.85 [0.70–1.03]). The 2 most common causes of non-HIV-related mortality among CCR participants and controls were major cardiovascular disease and malignant neoplasms. Risk of death from major cardiovascular disease was significantly lower among CCR participants (aHR: 0.64 [0.42–0.98]) and risk of death from malignant neoplasms was nonsignificantly lower (aHR: 0.63 [0.39–1.04]) (Supplementary Table 4).
Assessment of Unmeasured Confounding
With an observed hazards ratio (HR) of 0.71 for all-cause mortality, an unmeasured confounder that was associated with both all-cause mortality (outcome) and enrollment into CCR (exposure) by a risk ratio of 2.2-fold each, above and beyond the measured confounders, could explain away the estimate, but weaker joint confounder associations could not explain away the estimate. For the confidence interval to include the null, an unmeasured confounder that was associated with the outcome and exposure by a risk ratio of 1.7-fold each could do so, but weaker joint associations could not.
Factors Associated With All-cause and HIV-related Mortality
All-cause mortality risk was higher among CCR participants with unstable (vs stable) housing (aHR: 1.32 [1.01–1.73]) and among CCR participants with PTSD (aHR: 1.51 [1.01–2.24]) (Table 3). Insurance status, recent drug use, depression/anxiety, psychosis, and education level were not significantly associated with all-cause mortality. No variables were significantly associated with HIV-related mortality.
Table 3.
Baseline Characteristics Associated With All-cause and HIV-related Mortality Among CCR Enrollees
| Total Persons | All-cause Mortality | HIV-related Mortality | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N | Sum PY | Deaths | Crude Rate (Per 100 PY) | Crude HR | Adjusted HR | Deaths | Crude Rate (Per 100 PY) | Crude HR | Adjusted HR | |
| … | 3040 | 8580 | 284 | 3.31 | … | … | 60 | 0.70 | … | … |
| Housing status | … | … | … | … | … | … | … | … | … | … |
| Unstable | 681 | 1896 | 82 | 4.32 | 1.43 (1.11, 1.85) | 1.32 (1.01, 1.73) | 12 | 0.63 | 0.90 (0.48, 1.71) | 0.77 (0.40, 1.49) |
| Stable | 2271 | 6437 | 194 | 3.01 | Ref | Ref | 45 | 0.70 | Ref | Ref |
| Unknown | 88 | 248 | 8 | 3.23 | 1.07 (0.53, 2.17) | 1.32 (0.65, 2.70) | 3 | 1.21 | 1.74 (0.54, 5.59) | 2.85 (0.86, 9.40) |
| Recent drug use | … | … | … | … | … | … | … | … | … | … |
| No | 2023 | 5713 | 185 | 3.24 | … | … | 38 | 0.67 | … | … |
| Yes | 670 | 1874 | 71 | 3.79 | 1.17 (0.89, 1.54) | 1.02 (0.77, 1.34) | 15 | 0.80 | 1.20 (0.66, 2.19) | 1.06 (0.57, 1.96) |
| Unknown | 347 | 994 | 28 | 2.82 | 0.87 (0.58, 1.29) | 0.94 (0.63, 1.40) | 7 | 0.70 | 1.06 (0.47, 2.37) | 1.23 (0.54, 2.77) |
| Depression/anxiety | … | … | … | … | … | … | … | … | … | … |
| No | 1894 | 5364 | 164 | 3.06 | … | … | 33 | 0.62 | … | … |
| Yes | 1146 | 3216 | 120 | 3.73 | 1.22 (0.96, 1.54) | 0.95 (0.74, 1.21) | 27 | 0.84 | 1.36 (0.82, 2.27) | 1.36 (0.82, 2.27) |
| Psychosis | … | … | … | … | … | … | … | … | … | … |
| No | 2574 | 7291 | 233 | 3.20 | … | … | 53 | 0.73 | … | … |
| Yes | 466 | 1290 | 51 | 3.95 | 1.24 (0.91, 1.67) | 1.20 (0.83, 1.63) | 7 | 0.54 | 0.75 (0.34, 1.64) | 0.68 (0.31, 1.51) |
| PTSD | … | … | … | … | … | … | … | … | … | … |
| No | 2835 | 8019 | 256 | 3.19 | … | … | 54 | 0.67 | … | … |
| Yes | 205 | 561 | 28 | 4.99 | 1.56 (1.06, 2.31) | 1.51 (1.01, 2.24) | 6 | 1.07 | 1.59 (0.68, 3.69) | 1.39 (0.59, 3.27) |
| Education | … | … | … | … | … | … | … | … | … | … |
| >High school/GED | 924 | 2655 | 68 | 2.56 | … | … | 17 | 0.64 | … | … |
| High school/GED or lower | 2019 | 5652 | 207 | 3.66 | 1.43 (1.09, 1.88) | 1.33(0.97, 1.78) | 40 | 0.71 | 1.10 (0.63, 1.95) | 0.85 (0.47, 1.52) |
| Unknown | 97 | 273 | 9 | 3.30 | 1.29 (0.64, 2.58) | 1.18 (0.59, 2.53) | 3 | 1.10 | 1.72 (0.50, 5.87) | 2.31 (0.66, 8.12) |
Adjusted model includes age, gender, transmission risk, race, treatment era, baseline CD4, and baseline VL.
Abbreviations: CCR, revised Care Coordination Program; GED, general educational development; HR, hazards ratio; PTSD, posttraumatic stress disorder; PY, person years; VL, viral load.
DISCUSSION
Relative to usual care, the CCR may reduce all-cause mortality among populations at risk for poor outcomes, including PAWH and younger PWH, PWH lacking evidence of VS, and PWH with CD4 <350 cells/µL. All-cause mortality reductions among CCR participants were primarily driven by decreases in HIV-related mortality and accompanied by some decreases in non-HIV-related causes of mortality. Comparative effectiveness studies do not usually focus on mortality, as interventions target gaps further upstream in the care continuum that would lead to mortality. Care coordination has emerged as an effective intervention for retention in care and for VS and durable VS [12, 13, 16, 22, 28], and our data suggest that the CCR can also be an effective intervention for reducing HIV-related mortality and mortality from some non-HIV-related causes. In the case of HIV-related mortality, the mechanism is likely through improved care retention and sustained VS. The finding of an effect on some non-HIV-related causes is notable, suggesting the CCR may be supporting timely identification and/or improved management of other conditions through effective coordination of care with non-HIV specialists.
Identifying evidence-based strategies for PAWH is essential for individual and public health, given that up to three-quarters of PWH in the United States will be >50 years of age by 2040. As of October 2025, the Compendium of Evidence-Based Interventions and Best Practices for HIV Prevention did not include any intervention with an evidence-based designation for PAWH [28, 29]. Programs such as the SF Golden Compass offer models of aging-focused care; however, comparative effectiveness data are unavailable [29, 30]. Cumulative HIV viremia has been associated with increased risk of myocardial infarction, stroke and cardiovascular disease in some cohort studies [31–34]. Thus, achieving and maintaining VS is an important factor in preventing and treating comorbidities associated with HIV and may be another mechanism through which care coordination addresses some non-HIV-related causes of mortality [35, 36]. Relative to the general population, PWH have a greater burden of chronic conditions, such as cancers, metabolic disorders [2, 37], and mental health disorders [38, 39] and a greater prevalence of risk factors for chronic conditions, such as smoking and substance use [40, 41]. Further increasing life expectancy, especially for PAWH, will require treating and preventing chronic conditions other than HIV [42]. Models, such as care coordination, that integrate HIV, primary, and specialty care, along with support for psychosocial and basic needs, may yield broader health and survival gains beyond the HIV continuum.
Among enrollees, assessing for PTSD or experience of unstable housing could help to identify those at higher risk for mortality. Study results highlight the importance of stratified care and treatment planning at the time of enrollment. Future research should explore how addressing barriers such as housing instability may reduce risk of mortality. We did not observe significant risk factors for HIV-related causes of death, but this may be due to small samples and reduced power.
Our study's strengths include the application of a target trial emulation, which enhances validity by ensuring alignment of eligibility and follow-up across CCR and non-CCR PWH, thus avoiding common design flaws that contribute to selection bias [21]. Another strength was the use of population-based outcome data, ensuring we had local and national mortality and cause of death captured among CCR and non-CCR PWH. However, individuals were not randomized to the CCR, and unmeasured confounding remains a concern. We were unable to control for mortality risk factors, such as smoking or substance use, which may differ between CCR and non-CCR-enrolled PWH. Such risk factors would need to be associated with enrollment into the program and the outcome by a risk ratio of 2.2-fold each to explain away the observed effects of the CCR. In addition, our ability to detect small to moderate effect sizes may be limited by the sample size of some smaller subgroups of interest. Finally, our results are based on an ITI framework, which would move effect estimates closer to the null value.
Relative to usual care, the CCR shows benefits for reducing all-cause mortality among populations at risk for poor outcomes, including PAWH, younger PWH, and PWH lacking evidence of VS in the prior year. The impact on all-cause mortality was primarily driven by reductions in HIV-related mortality. These findings suggest that the CCR may be directly or indirectly addressing barriers to HIV treatment and adherence, as well as helping to treat and prevent other chronic conditions among PWH.
Supplementary Material
Notes
Acknowledgments. The authors are indebted to Gina Gambone, Grace Herndon, Graham Harriman, Jennifer Carmona, Guadalupe Dominguez Plummer, and the Ryan White HIV/AIDS Program Part A Care Coordination Program staff for their dedication to the delivery of comprehensive services, their continual participation in reporting on this intervention, and their shared commitment to rigorous evaluation and the integration of findings into practice. The authors also wish to recognize the members of the study Community Advisory Board, for their guidance and critical input at various stages of this work. This work was completed as part of the Program Refinements to Optimize Model Impact and Scalability based on Evidence (PROMISE) Study.
Financial support. This work was supported by the National Institute of Mental Health (grant no. R01 MH117793) and (for the intervention itself) the Health Resources and Services Administration (grant no. H89HA00015).
Data availability. The data underlying this article cannot be shared publicly due to legal restrictions (New York Public Health Law Article 21, Title III) and the confidential nature of HIV surveillance data.
Contributor Information
McKaylee M Robertson, Institute for Implementation Science in Population Health (ISPH), City University of NewYork (CUNY), New York City, New York, USA; Graduate School of Public Health and Health Policy, City University of NewYork (CUNY), New York City, New York, USA.
Tigran Avoundjian, Bureau of Hepatitis, HIV, and Sexually Transmitted Infections, NewYork City Department of Health and Mental Hygiene, New York City, New York, USA.
Faisal Abdelqader, Bureau of Hepatitis, HIV, and Sexually Transmitted Infections, NewYork City Department of Health and Mental Hygiene, New York City, New York, USA.
Sarah L Braunstein, Bureau of Hepatitis, HIV, and Sexually Transmitted Infections, NewYork City Department of Health and Mental Hygiene, New York City, New York, USA.
Denis Nash, Institute for Implementation Science in Population Health (ISPH), City University of NewYork (CUNY), New York City, New York, USA; Graduate School of Public Health and Health Policy, City University of NewYork (CUNY), New York City, New York, USA.
Mary K Irvine, Bureau of Hepatitis, HIV, and Sexually Transmitted Infections, NewYork City Department of Health and Mental Hygiene, New York City, New York, USA.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
References
- 1. Samji H, Cescon A, Hogg RS, et al. Closing the gap: increases in life expectancy among treated HIV-positive individuals in the United States and Canada. PLoS One 2013; 8:e81355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Marcus JL, Leyden WA, Alexeeff SE, et al. Comparison of overall and comorbidity-free life expectancy between insured adults with and without HIV infection, 2000-2016. JAMA Netw Open 2020; 3:e207954. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Sánchez-Conde M, Díaz-Alvarez J, Dronda F, Brañas F. Why are people with HIV considered “older adults” in their fifties? Eur Geriatr Med 2019; 10:183–8. [DOI] [PubMed] [Google Scholar]
- 4. The Lancet Healthy Longevity . Ageing with HIV. Lancet Healthy Longev 2022; 3:e119. [DOI] [PubMed] [Google Scholar]
- 5. Althoff KN, Stewart CN, Humes E, et al. The shifting age distribution of people with HIV using antiretroviral therapy in the United States. AIDS 2022; 36:459–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. O’Connor SM, Song R, Viguerie A. et al. Projected increases in older people with HIV in the United States through 2040. Open Forum Infect Dis 2025; 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yang H-Y, Beymer MR, Suen S-C. Chronic disease onset among people living with HIV and AIDS in a large private insurance claims dataset. Sci Rep 2019; 9:18514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Brooks JT, Buchacz K, Gebo KA, Mermin J. HIV infection and older Americans: the public health perspective. Am J Public Health 2012; 102:1516–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Centers for Disease Control and Prevention . Prevention Research Synthesis (PRS) HIV Compendium of Best Practices. PRS compendium intervention search. Accessed March 17, 2025. Available at: https://wwwn.cdc.gov/HIVCompendium/SearchInterventions.
- 10. Irvine MK, Levin B, Abdelqader F, et al. Evaluation of the revised versus original Ryan white part A HIV care coordination program in a cluster-randomized, stepped-wedge trial. J Acquir Immune Defic Syndr 2023; 92:325–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Irvine MK, Levin B, Robertson MM, et al. PROMISE (Program Refinements to Optimize Model Impact and Scalability based on Evidence): a cluster-randomised, stepped-wedge trial assessing effectiveness of the revised versus original Ryan White Part A HIV Care Coordination Programme for patients with barriers to treatment in the USA. BMJ Open 2020; 10:e034624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Robertson MM, Penrose K, Irvine MK, et al. Impact of an HIV care coordination program on durable viral suppression. J Acquir Immune Defic Syndr 2019; 80:46–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Nash D, Robertson MM, Penrose K, et al. Short-term effectiveness of HIV care coordination among persons with recent HIV diagnosis or history of poor HIV outcomes. PLoS One 2018; 13:e0204017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Centers for Disease Control and Prevention . Compendium of evidence-based interventions and best practices for HIV prevention: LRC chapter – HIV Care Coordination Program (CCP). Accessed March 17, 2025. Available at: https://stacks.cdc.gov/view/cdc/149351.
- 15. Centers for Disease Control and Prevention . Compendium of evidence-based interventions and best practices for HIV prevention: Structural interventions chapter – HIV Care Coordination Program (CCP). Accessed March 17, 2025. Available at: https://stacks.cdc.gov/view/cdc/149351.
- 16. Robertson MM, Penrose K, Nash D, et al. Impact of an HIV Care Coordination Program on the timeliness of viral suppression and immune recovery among clients newly diagnosed with HIV. AIDS Behav 2020; 24:1237–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Irvine MK, Chamberlin SA, Robbins RS, et al. Improvements in HIV care engagement and viral load suppression following enrollment in a comprehensive HIV care coordination program. Clin Infect Dis 2015; 60:298–310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Robertson MM, Irvine MK, Penrose K, et al. Mortality among clients in the New York city HIV Care Coordination Program (CCP): incidence and associated clinical factors. Ann Epidemiol 2021; 64:161–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Irvine MK., Chamberlin SA, Robbins RS, et al. Come as you are: improving care engagement and viral load suppression among HIV Care Coordination clients with lower mental health functioning, unstable housing, and hard drug use. AIDS Behav 2017; 21:1572–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.American Academy or Pediatrics. Medical home. Accessed March 17, 2025. Available at: https://www.aap.org/en/practice-management/medical-home/.
- 21. Hernán MA, Sauer BC, Hernández-Díaz S, Platt R, Shrier I. Specifying a target trial prevents immortal time bias and other self-inflicted injuries in observational analyses. J Clin Epidemiol 2016; 79:70–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Avoundjian T, Robertson MM, Abdelqader F, et al. Impact of a revised Ryan White Part A HIV Care Coordination Program on viral suppression and durable viral suppression. J Acquir Immune Defic Syndr 2025; 101:77–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Bosh KA, Johnson AS, Hernandez AL, et al. Vital signs: deaths among persons with diagnosed HIV infection, United States, 2010-2018. MMWR Morb Mortal Wkly Rep 2020; 69:1717–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. CDC . HIV diagnoses, deaths, and prevalence: 2025 update. Accessed March 17, 2025. HIV Data Available at: https://www.cdc.gov/hiv-data/nhss/hiv-diagnoses-deaths-and-prevalence-2025.html.
- 25. Heron M. Deaths: leading causes for 2017. Natl Vital Stat Rep 2019; 68:1–77. [PubMed] [Google Scholar]
- 26.Hernán MA, Robins JM. Causal inference: What if (the book). Accessed March 17, 2025. Miguel Hernan’s faculty website, 2020. Available at: https://www.hsph.harvard.edu/miguel-hernan/causal-inference-book/.
- 27. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med 2017; 167:268–74. [DOI] [PubMed] [Google Scholar]
- 28. Irvine MK, Robertson MM, Nash D, et al. HIV Care Coordination promotes care re-engagement and viral suppression among people who have been out of HIV medical care: an observational effectiveness study using a surveillance-based contemporaneous comparison group. AIDS Res Ther 2021; 18:70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Greene M, Myers J, Tan JY, et al. The Golden Compass program: overview of the initial implementation of a comprehensive program for older adults living with HIV. J Int Assoc Provid AIDS Care 2020; 19:2325958220935267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Tan JY, Greene M, Blat C, et al. Examining the impact of the Golden Compass clinical care program for older people with HIV: a qualitative study. AIDS Behav 2022; 26:1562–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Delaney JA, Nance RM, Whitney BM, et al. Cumulative human immunodeficiency viremia, antiretroviral therapy, and incident myocardial infarction. Epidemiology 2019; 30:69–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Harding BN, Avoundjian T, Heckbert SR, et al. HIV viremia and risk of stroke among people living with HIV who are using antiretroviral therapy. Epidemiology 2021; 32:457–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Elvstam O, Marrone G, Engström G, et al. Associations between HIV viremia during antiretroviral therapy and cardiovascular disease. AIDS 2022; 36:1829–34. [DOI] [PubMed] [Google Scholar]
- 34. Morton ZP, Christina Mehta C, Wang T, et al. Cumulative human immunodeficiency virus (HIV)-1 viremia is associated with increased risk of multimorbidity among US women with HIV, 1997-2019. Open Forum Infect Dis 2023; 10:ofac702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Feinstein MJ, Hsue PY, Benjamin LA, et al. Characteristics, prevention, and management of cardiovascular disease in people living with HIV: a scientific statement from the American heart association. Circulation 2019; 140:e98–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Webel AR, Schexnayder J, Cioe PA, Zuñiga JA. A review of chronic comorbidities in adults living with HIV: state of the science. J Assoc Nurses AIDS Care 2021; 32:322–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Gallant J, Hsue PY, Shreay S, Meyer N. Comorbidities among US patients with prevalent HIV infection-A trend analysis. J Infect Dis 2017; 216:1525–33. [DOI] [PubMed] [Google Scholar]
- 38. Do AN, Rosenberg ES, Sullivan PS, et al. Excess burden of depression among HIV-infected persons receiving medical care in the united states: data from the medical monitoring project and the behavioral risk factor surveillance system. PLoS One 2014; 9:e92842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Beer L, Tie Y, Padilla M, Shouse RL, Project MM. Generalized anxiety disorder symptoms among persons with diagnosed HIV in the United States. AIDS 2019; 33:1781–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Edelman EJ, Tetrault JM, Fiellin DA. Substance use in older HIV-infected patients. Curr Opin HIV AIDS 2014; 9:317–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Deren S, Cortes T, Dickson VV, et al. Substance use among older people living with HIV: challenges for health care providers. Front Public Health 2019; 7:94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Edwards JK, Cole SR, Breger TL, et al. Five-year mortality for adults entering human immunodeficiency virus care under universal early treatment compared with the general US population. Clin Infect Dis 2022; 75:867–74. [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.