Abstract
Despite known benefits, only 19–28% of HIV-infected Americans are virologically suppressed (defined as ≤200 copies/mL). Engagement in HIV care represents a continuum from patients unaware they are infected to virological suppression. The electronic medical record of all newly diagnosed HIV-infected MSM seen at Fenway Health between 2000 and 2010 were reviewed. Patients were “engaged” if they had one negative HIV test and/or one physical exam within 24 months prior to their HIV diagnosis (n=291). All others were considered “new” (n=463). MSM engaged in care prior to HIV diagnosis were more often identified in acute retroviral syndrome or on routine screening, more rapidly linked to care, and less often diagnosed with a concomitant STI than those who were not engaged in care. Nearly 19% of all patients were diagnosed with AIDS the same time they were diagnosed with HIV. Blacks and those with higher CD4 counts at diagnosis were less likely to be virologically suppressed at 1 year. Between 2000 and 2010, patients retained in care were more likely to initiate ART and be virologically suppressed within 1 year independent of initial HIV viral load and CD4 count. Engagement in care prior to seroconversion influences important HIV outcomes. Programs that care for at risk populations should institute routine opt-out HIV testing and test-and-treat programs to optimize HIV care and prevention.
Introduction
Studies estimate that only 51% of HIV-infected Americans are receiving regular HIV care and just 19–28% are virally suppressed. 9,10 Those who are most likely to transmit HIV are least likely to be engaged in care. 9,10 Engagement in care represents a spectrum of distinct processes as defined by the Health Resources and Services Administration, ranging from patients not in care and unaware of their HIV status to patients fully engaged in comprehensive HIV primary medical care. 7 Significant deficits in engagement in HIV care, such as late diagnosis, suboptimal linkage to and retention in care, inappropriate ART use, and poor adherence to therapy, present major barriers to achieving optimal outcomes. 8 Furthermore, these obstacles present extensive barriers to the successful implementation of healthcare provider dependent public health strategies, including widespread opt-out testing and test-and-treat programs for HIV care and prevention. 11 –13 Despite increasing evidence for the importance of engagement in care, little is known about the characteristics of engaged patients and how engagement prior to seroconversion influences HIV outcomes. Fenway Health is a medical facility in Boston that provides healthcare to >20,000 patients, of which >25% are MSM. These demographics offer a unique opportunity to study the sociodemographic and behavioral characteristics of newly diagnosed HIV-infected MSM. The present study aims to explore how engagement in care prior to diagnosis influences specific HIV outcomes, such as identifying patients in acute retroviral syndrome (ARS) or AIDS at diagnosis, and virological suppression at 1 year. This investigation also seeks to determine sociodemographic, clinical, and behavioral associations with identifying patients in ARS or AIDS at diagnosis, and virological suppression at 1 year.
Methods
Study design, population, and setting
We collected data from the electronic medical record (EMR) database of all newly diagnosed HIV-infected MSM seen at Fenway Health, Boston, between January 2000 and December 2010 (n=754). Transgender women (male at birth) who reported sex with men were also included. Patients transferring HIV care, women at birth, and those reporting heterosexual contact, transfusion, or occupational exposure as their only HIV risk factors were excluded from this analysis. The study was approved by the Institutional Review Board of Fenway Health.
Variables collected
EMRs were individually reviewed to retrieve sociodemographic data systemically collected on Fenway Health patients including age, race, relationship status, insurance status, employment, housing status, as well as standardized clinical information concerning date of initial positive HIV antibody test, location of and reason for initial test, HIV viral load, and CD4 count at the initial visit and 1 year, risk factors for HIV acquisition (e.g., MSM, MSM, and IDU), the presence of acute retroviral syndrome (ARS), AIDS, or opportunistic infections; ART initiation, retention in HIV care, and evidence for other STIs at the initial visit and 1 year. For all patients, chart review was undertaken to extract current risk behavior data at the initial HIV visit concerning recreational drug use, alcohol use, number and gender of sexual partners, being in an HIV serodiscordant primary relationship, and other sexual practices. For patients engaged in care at Fenway Health, the reason (e.g., annual physical exam, urgent visit) and number of Fenway Health visits within 24 months prior to HIV diagnosis were also recorded. Missing data points were excluded from the analysis. Data was confirmed for accuracy by secondary review of patient records.
Definitions and primary outcome variables
Patients were considered “engaged in care” if they had at least one prior HIV negative test and/or one prior routine physical exam with a Fenway Health provider within 24 months prior to HIV diagnosis. Mental health or urgent visits that may have included a physical exam were excluded. All other newly diagnosed HIV-infected MSM were considered “new” patients if they presented for an initial HIV visit without previous engagement with another HIV provider.
The location of HIV test was categorized into “testing service at Fenway” for patients tested at the Fenway Health walk-in HIV counseling and testing service, “Fenway provider” for patients tested by a provider during an office visit, “home kit” for patients using over the counter HIV testing kits, “inpatient hospital” for patients tested during a hospital admission, “outside PCP” for patients tested by a non-Fenway primary care provider, and “walk-in clinic” for patients tested at a non-Fenway related testing service.
The primary reason for the HIV test was categorized into “exposure” for patients reporting an episode of unprotected sex, “random” for patients reporting an irregular testing history, “routine” for patients reporting a regularly scheduled testing regime, and “symptomatic” for patients reporting any symptoms, including that of an STI, opportunistic infection, or ARS.
Patients were considered “retained in care” if they were maintained in HIV primary care at Fenway Health with a routine follow up visit, including HIV surveillance labs noted between 11 and 13 months after the initial HIV visit. This was also the window for 1-year viral load data. As a definitive reason for loss to follow-up was not always documented, patients who died or transferred care were considered lost to follow-up and included as patients not retained in care.
Our primary outcomes were ARS and AIDS being diagnosed at the initial presentation, and virological suppression at 1 year. Patients with ARS were identified by the presence of detectable plasma HIV RNA without a positive ELISA and/or Western blot, or a physician diagnosis of ARS in the EMR. 14 A physician diagnosis of ARS included patients diagnosed at outside institutions where laboratory data were unavailable for confirmation, but ARS was documented by a healthcare provider and referenced appropriate laboratory data. Patients diagnosed with AIDS were recognized by a CD4 T-lymphocyte count <200 cells/μL or CD4 T-lymphocyte percentage of total lymphocytes of <14 or documentation of an AIDS-defining condition as outlined by the CDC. 14 Viral suppression at 1 year was defined as an HIV plasma viremia <75 copies/mL or physician documentation of an undetectable viral load in the EMR.
Statistical analysis
All analyses were performed using SPSS 19 (IBM Corporation, Armonk, NY), with statistical significance determined at the alpha 0.05 level.
Data on newly diagnosed HIV-infected patients were reported and stratified by level of engagement. Continuous variables were summarized using means, medians, and standard deviations, while categorical variables were summarized using proportions. T-tests/Wilcoxon rank sum tests and chi squared/Fisher Exact tests were utilized to examine associations between engaged patients and new patients.
Logistic regression procedures were used to determine factors independently associated with the primary outcomes. Linear regression procedures were used to compare trends in variables over the data collection period. Trends were also stratified by engagement. For all final fitted regression models, all variables with p values<0.10 on bivariate analyses were included.
Results
Characteristics of study population and engagement in care
We identified 291 “engaged” and 463 “new” HIV-infected patients at Fenway Health who presented for an initial comprehensive care visit between January 2000 and December 2010 (Table 1). Engaged patients were more often white, privately insured, and employed full-time compared to new patients. In terms of HIV diagnostics, engaged patients were more often identified in symptomatic ARS or on routine screening by a Fenway Health provider, and at an older age, than new patients. Engaged patients reported fewer days between initial diagnosis and linkage to care, more often disclosed unprotected sexual practices and substance use, and less often diagnosed with a concomitant STI at the initial HIV visit than new patients. Engaged patients were also more often retained in care at one year (Table 2).
In contrast, new patients were more often Hispanic, with regard to minority patients, and received free care compared to engaged patients (Table 1). In terms of HIV diagnostics, new patients were more often identified at a walk-in clinic or by an outside primary care provider (PCP) on random screening at a younger age than engaged patients, and then referred to Fenway Health.
Acute retroviral syndrome
Ninety-three patients were diagnosed with acute retroviral syndrome (ARS) over the decade, representing 12.3% of new HIV diagnoses (Tables 1, 3, and 4). Over 63% of these diagnoses were in engaged patients with 17% presenting with a concomitant STI, most frequently syphilis (10.8%).
The correlates of presenting with ARS are analyzed in Table 5. In the multivariable model, HIV testing at any location other than Fenway Health was associated with a decreased odds of being diagnosed with ARS (AOR=0.36; 95% CI: 0.21–0.64). In addition, individuals diagnosed with ARS were more likely to have been tested after a high-risk exposure (AOR=17.32; 95% CI: 1.81–165.67) or to present with symptoms (AOR=191.81; 95% CI: 26.16–1406.4) compared to patients who underwent routine testing.
AIDS
Nearly 19% of all patients had CD4 counts below 200 or an opportunistic infection at the time of their initial HIV diagnosis, accounting for 16.5% of engaged patients and 21.1% of new patients (Table 1). The prevalence of AIDS in newly diagnosed patients remained unchanged over time. Almost 66% of patients diagnosed with AIDS at their first HIV-related visit were new patients, and nearly half presented with a symptomatic opportunistic infection (Tables 3 and 4).
Table 6 shows the correlates of initially presenting with an AIDS diagnosis. In the multivariable analysis, older patients at the time of their HIV diagnosis (AOR=1.04 95% CI: 1.01–1.06) and those who were symptomatic when they tested (AOR=2.22, 95% CI: 1.42–3.46) were more likely to present with AIDS. Patients who were diagnosed by routine HIV testing (AOR=0.20, 95% CI: 0.08–0.53) and those in a primary serodiscordant relationship (AOR=0.47, 95% CI 0.23–0.98) were less likely to present with AIDS when they were diagnosed with HIV.
Viral suppression at 1 year
Of 595 newly diagnosed HIV-infected patients who were retained in care after 1 year, 47.2% achieved an undetectable viral load, and of patients initiated on ART, 99.3% achieved an undetectable viral load (Tables 2 –4).
On multivariable analysis, identifying as Black (AOR=0.37, 95% CI: 0.15–0.91) and having an increased CD4 count on initial HIV labs (AOR=0.99, 95% CI: 0.99–0.99) were associated with a decreased odds of viral suppression at 1 year (Table 7). Having a concomitant STI at the time of HIV diagnosis was associated with an increased odds of viral suppression at one year (AOR=2.81, 95% CI: 1.62–4.86). No patient in ARS was offered immediate ART and those presenting in ARS did not have an increased odds of virological suppression at 1 year.
Clinical and demographic trends in study population in 2000–2010
Over the data collection period from 2000 to 2010, there were no significant trends in patient demographics, engagement, plasma HIV RNA and CD4 count at time of presentation, CD4 count at 1 year, the prevalence of ARS or AIDS at time of HIV diagnosis, or rates of retention in care, substance use, sexual behaviors, or STIs.
However, there was a significant increased trend in the percent of retained patients achieving viral suppression at 1 year (1.12, 95% CI: 1.06–1.17), such that by 2010, nearly 65% of retained patients demonstrated an undetectable viral load after 1 year of engagement. Over the decade there was a trend of decreased clinic viral load at 1 year (0.09, 95% CI: 0.09–0.09) in conjunction with increased ART initiation (1.08, 95% CI: 1.03–1.14). Figure 1 shows mean and median initial viral load and CD4 count (A) compared with percent of retained patients who were virologically suppressed at 1 year (B) during the data collection time period.
Mean and median initial viral load and CD4 count at the initial visit
Discussion
Despite the advent of safe and effective therapy over the past 16 years, incomplete engagement in HIV care is common. Of the 1.1 million individuals living with HIV in the United States, an estimated 21% of are unaware of their infection, with knowledge rates being as low as 44% in some subpopulations (e.g., Black MSM). 15,16 About 50% of known HIV-infected individuals are not engaged or retained in regular HIV care. 9 Poorly engaged, retained, or undiagnosed HIV-infected individuals do not have regular access to the benefits of ART, screening, and prophylaxis for opportunistic infections and STIs, and other medical services. Those unaware of their HIV infection have a significantly higher risk of transmitting HIV, while HIV-infected patients who are virologically suppressed are less likely to transmit the virus to their partners. 17,18 Incompletely engaged and undiagnosed individuals account for the largest proportion of HIV-infected persons with detectable viremia and significantly contribute to ongoing transmission to uninfected partners. 9
The current study reviewed trends in HIV diagnosis and engagement in care among newly diagnosed HIV-infected MSM at the largest ambulatory care center for MSM in New England. We found that MSM already engaged in primary care at Fenway Health at the time of their HIV diagnosis were more often identified with symptomatic ARS or by routine screening than those referred from other providers. Patients who had a preexisting relationship with Fenway Health were more rapidly linked to HIV primary care, retained in care at 1 year, and were less often diagnosed with a concomitant STI at the initial HIV visit than new patients.
From a public health perspective, the risk of infection from patients with acute HIV appears to be 10–20 times higher than that from patients with chronic infection. 19,20 Up to 50% of new infections among MSM in high-income nations may be attributable to acute HIV infection. 20 In the face of an intractable rate of new infections among MSM, identifying acutely infected patients with high levels of viremia has considerable benefits at the individual and community level, and underscores the importance of educating primary care providers in the recognition of ARS. The provision of comprehensive care for MSM may be an important means to prevent and detect more acute infections, as we found that MSM who presented for more physical exams and medical sick visits within the 24 months prior to HIV diagnosis were more likely to have their HIV infection detected prior to seroconversion.
Fenway Health's comprehensive EMR is a major strength of the present study since it included all patient demographic data, physician, nursing, and mental health notes, notes from tertiary referral hospital, labs, medications, and other relevant information. However, substance use and sexual behavior were not systematically collected, and the high prevalence of methamphetamine and club drug use among many Boston area MSM is probably underestimated among the participants in this study. 21,22 The seemingly greater prevalence of substance use and unprotected sexual behaviors in engaged patients may be an issue of disclosure, as existing patients under the care of culturally sensitive providers may be more likely to disclose HIV risk behaviors, and have more visits in which to do so, and thus be screened for HIV and other STIs more frequently.
Despite engagement in care prior to diagnosis, a concerning 19% of existing and new patients were identified with AIDS at the time their HIV was diagnosed, and this disturbing figure remained unchanged over time. Moreover, half of the engaged patients diagnosed with AIDS presented with a symptomatic opportunistic infection. This result is surprising, as an existing relationship with a provider was not protective against AIDS, and there was no association with type and number of visits within the 24 months prior to diagnosis. These findings underscore the importance of ongoing HIV testing in primary care settings and suggest that even in specialized care settings, more frequent routine testing is needed, without waiting for the onset of symptoms. 23 –25 However, patients who disclosed a primary serodiscordant relationship demonstrated a decreased likelihood to be diagnosed with AIDS at HIV diagnosis, likely reflecting more frequent monitoring by their provider.
We also found that older patients were more likely to present with AIDS at their HIV diagnosis. Other studies have shown older populations to be at a greater risk of late presentation or advanced HIV disease at diagnosis. 23,26 –29 The reason for this finding is likely multi-factorial, including cultural factors, patient lack of appreciation for HIV risk, and provider bias in screening older populations for risk behaviors and offering HIV and STI testing. Future studies and public health interventions should address risk perceptions in patients and providers regarding older populations.
Additionally, partnered relationships were not protective of ARS or AIDS. Most patient notes did not specify whether these relationships were monogamous, but the high prevalence of partnered patients with newly diagnosed HIV suggests that all sexually active MSM should be regularly screened for HIV and other STIs regardless of reported relationship status. The high incidence of concomitant STIs at the initial visit and within 1 year further supports this recommendation. Patients diagnosed with concomitant STIs at the time of their HIV diagnosis were more likely to be virologically suppressed after 1 year, which may reflect the provider's preference to initiate ART in the presence of perceived increased risky sexual behaviors as a secondary benefit to the community. Future studies should explore this relationship further in a longitudinal design.
We also found that Black MSM were less likely to be virologically suppressed after 1 year, corroborating other concerning observations of the continued prevalence of racial disparities with regard to HIV prevalence, diagnostic timing, and engagement in care. 10,30 –32 Although there were few Black patients in the present study, these results highlight the role of healthcare disparities in HIV outcomes. The findings confirm the need for further research in this area and culturally-tailored programs to engage Black MSM in their care environments, as is a central point of the National HIV/AIDS Strategy.
In comparison to national data, all patients at Fenway Health significantly benefitted from higher levels of retention in care, ART initiation, and viral suppression. 9,15 –17 However, between 2000 and 2010, there were no appreciable trends in presenting levels of plasma viremia and CD4 count, suggesting the need for new programs to facilitate earlier and more frequent screening of at risk MSM. More encouraging was the finding that between 2000 and 2010, with increasing evidence for early initiation of ART and guideline changes, there was a trend of a greater percentage of retained patients to initiate HAART and to be virally suppressed within 1 year. 33 Of patients retained in care and initiated on ART, over 86% achieved viral suppression throughout the entire data collection period, and by 2010, nearly 99% achieved viral suppression. These impressive findings may have direct implications for public health programs to reduce community viral load such as test-and-treat, where the full engagement of HIV-infected individuals in care will be required for HIV prevention.
However, our broad definition for engagement in care may not accurately reflect true engagement, as two patients satisfying the definition for engagement were diagnosed by an outside PCP. Moreover, sensitivity analysis of patients diagnosed by an outside PCP did not reveal an increased odds of ARS or AIDS. This finding may support the development of an acute HIV screening protocol in addition to opt-out HIV antibody testing. This protocol may capture all patients with any risk factor for HIV acquisition in the setting of symptoms suggestive of viral infection. However, as a convenience sample, these results may not be applicable to patients engaged in primary care prior to diagnosis at centers without comprehensive care for MSM or HIV. Despite this limitation, patients receiving care at a comprehensive care site for MSM resulted in a higher proportion of ARS diagnoses compared to non-Fenway related diagnoses. This benefit cannot be highlighted in isolation as the proportion of AIDS diagnoses at Fenway did not change throughout the data collection period and was not different than patient diagnosed at non-Fenway related location. This represents an area for tremendous improvement.
The present study confirms the positive influence of engagement in primary care prior to seroconversion on important HIV outcomes at diagnosis and within 1 year. Fenway Health is an integrated community-based medical facility with a focus on culturally sensitive, high quality healthcare for sexual and gender minorities, as well as HIV medicine, and may explain why outcomes in Fenway Health patients measured as retention in care and viral suppression are considerably better than national statistics. 8 –10 Fenway Health may also represent an ideal healthcare delivery model for HIV primary care and the successful implementation of widespread opt-out HIV testing and test-and-treat programs for HIV care and prevention. Future studies and public health efforts should focus on more accurately defining engagement and retention, especially as it pertains to healthcare provisions for MSM, and on the development of HIV prevention methods specific to MSM to reduce the high incidence of AIDS when first diagnosed with HIV.
Footnotes
Acknowledgments
The authors would like to thank Stephen Boswell, MD, President and Chief Executive Officer of Fenway Health, Assistant Professor in the Department of Medicine at Harvard Medical School, and Assistant Professor in the Division of Infectious Disease at the Massachusetts General Hospital for his valuable comments in the writing of this research.
Sources of support: Harvard Center for AIDS Research (NIAID 5P30AI06354-08).
Author Disclosure Statement
No competing financial interests exist.