Getting HIV-Positive People Into Care -- and Keeping Them There

Clinical Impact of Late Entry to HIV Care

No one needs vast experience with HIV infection -- or much imagination -- to grasp why delaying the first visit to an HIV clinic portends ominous clinical outcomes for people who test positive. Infection with a virus that relentlessly unpins the immune system kills most people, in time, unless they start taking drugs to pin down that virus. Probably for these reasons, research on the clinical impact of late entry to HIV care is sparse. But when hard numbers go lacking, modelers gleefully fill the breach. A modeling study that evaluated starting antiretrovirals late -- at a CD4 count under 200 cells/mm³ -- figured that tardy treatment takes 24 years off a normal life span.²³

This novel analysis by modelers at Harvard and other centers tried to reckon life expectancy, compared with the general population, in four groups: (1) HIV-negative people with mortality risk profiles similar to people with HIV because of substance abuse and other high-risk behaviors, (2) HIV-positive people who begin antiretrovirals according to then-current guidelines, that is, when the CD4 count fell below 350 cells/mm³, and who go on to another regimen when one regimen fails, (3) HIV-positive people who do not start antiretrovirals till their CD4 count falls below 200 cells/mm³, and (4) people who do not start the next available antiretroviral regimen after failure of one regimen.

The investigators derived demographic data from HIV-positive people in several cohorts. At HIV seroconversion, this group averaged 33 years of age, had an average viral load of about 65,000 copies/mL, and averaged 534 CD4 cells/mm³. About half were black, 27% non-Hispanic white, and 21% Hispanic.

Figure 3. Compared with the general US population, HIV-negative people with a mortality risk profile similar to HIV-positive people (for example, because of substance abuse) would lose 8.3 years from the expected life of a 33-year-old (first bar), according to a modeling study.23 HIV-positive people who start antiretroviral therapy (ART) at the 350 CD4 threshold and continue treatment with new regimens when one regimen fails would lose an estimated 20.2 years compared with 33-year-olds in the general population (second bar). HIV-positive people who do not start antiretroviral therapy until their CD4 count stands between 50 and 199 cells/mm3 would lose 24.15 years of life compared with 33-year-olds in the general population (third bar). And HIV-positive people who start treatment at 50 to 199 cells/mm3 but quit after two combinations fail would lose 27.3 years of life compared with the general population (fourth bar).

For 33-year-olds in the general US population, estimated life expectancy was an additional 42.9 years at the time of this study. For the HIV-negative group with a risk profile similar to US residents with HIV, life expectancy at age 33 would be only 34.6 years (Figure 3). In other words, compared with the general population, this "HIV-like" group would lose 8.3 years in life expectancy. For the HIV-positive group that starts antiretrovirals at the 350-cell threshold, life expectancy at age 33 would be 22.7 years, so they would lose an estimated 20.2 years compared with the general population.

For 33-year-old HIV-positive people who start antiretroviral therapy at a CD4 count between 50 and 199 cells/mm³, life expectancy would be only another 18.75 years, meaning they would lose 24.15 years compared with the general population, that is, 3.95 years more than the HIV-positive group that started antiretroviral therapy at the 350-cell cutoff. In this group that did not start antiretrovirals until their CD4 count lay between 50 and 199 cells/mm³, life expectancy would be 18.2 years for those who took four antiretroviral regimens before quitting, 17.0 years for those who took three regimens, 15.6 years for those who took two, and 13.7 years for those who took only one. To put it another way, dropping out of care earlier and earlier robs more and more years from an HIV-positive person's life.

Of course much of the data summarized in the article on late HIV diagnosis in this issue of RITA! also suggest the clinical and monetary impact of delayed care after HIV diagnosis, because later diagnosis means later care.

Strategies to Unstitch Seams Between HIV Diagnosis and Care

Indulgent readers will condone one more restatement of the obvious: HIV diagnosis earlier in the course of infection (the topic of the first article in this issue) will promote faster entry to care -- or at least entry at a higher CD4 count and a less dire disease stage. But diagnosis without linkage to care is almost pointless. And as Michael Mugavero and University of Alabama colleagues note, "although the importance of linkage to care is emphasized in the CDC guidelines, implementation has often focused on increasing the number of tests performed, with considerably less programmatic emphasis on linking patients to HIV care."²

In 2010 the US National HIV/AIDS Strategy outlined a plan to torque up access to care.²⁴ Though the goals are laudable, the recommended "action steps" are (in the nature of political manifestos) broad (Table 2). Still, the document's first action step accurately embodies a key finding of research discussed above and below: smoothing out "seams" that hem the newly diagnosed from speedy care should be a priority. Studies show that intense case management after diagnosis, and/or housing testing and care services inside the same walls, can ease the way into care. But even targeted programs can yield disappointing results in recalcitrant populations.

Table 2. US National HIV/AIDS Strategy to Increase Access to HIV Care24

Apparently only one randomized trial, by the CDC's Antiretroviral Treatment and Access Study (ARTAS), has assessed entry-to-care tactics.²⁵ The ARTAS team randomized just-diagnosed people to case management or passive referral standard-of-care (which seems substandard when one considers the often lengthy lapse between a positive test and a handshake with an HIV clinician). This trial, published in 2005, involved 273 recently diagnosed people in Atlanta, Baltimore, Los Angeles, and Miami. The intervention included up to five contacts with a case manager over 90 days, while people in the passive referral group got only information about HIV and local care resources.

Significantly more people in the case-management group saw an HIV clinician at least once in 6 months (78% versus 60%, adjusted relative risk 1.36, P = 0.0005), and at least twice in 12 months (64% versus 49%, adjusted relative risk 1.41, P = 0.006).²⁵ People older than 40, those who had not used crack recently, Hispanics, and people who enrolled in the trial within 6 months of HIV diagnosis were significantly more likely to have at least two HIV clinic visits. The CDC team estimated that this type of case management costs $600 to $1200 yearly -- a pittance when weighed against the cost of delayed care. Although these results clearly demonstrate that one-on-one case management gets people into care more reliably than a stack of informational leaflets, the low 12-month visit rate in either study group suggests these people needed more help than they got.

Two other nonrandomized ARTAS studies evaluated case management and co-location, current jargon for having testing and treatment units under one roof. One of these later analyses focused on a larger group of 626 recently diagnosed people seen at 10 sites across the United States from 2004 through 2006.²⁶ All these people had up to five meetings with a linkage case manager over 90 days. In the first 6 months, 497 of the 626 study participants (79%) saw an HIV clinician, almost the same proportion as in the randomized trial.²⁵ People who had two or more case-management sessions and those seen at a site with testing and care under the same roof were significantly more likely to have an HIV office visit in 6 months. Other factors that favored linkage to care were age over 25, Hispanic ethnicity, stable housing, and no recent noninjection drug use. A separate analysis that used ARTAS data plus site visits and project director reports figured that sites with co-located testing and care had a substantially higher linkage rate than stand-alone sites, 87% versus 73%.²⁷

Four studies assessed how well outreach programs get underserved HIV-positive groups into care -- injection drug users,^28,29 nonwhites,^29,30 and people with unstable housing.³¹ Street outreach for injection drug users²⁸ and peer-based outreach to people of color and injection drug users at 21 California sites (the California Bridge Project)²⁹ got only about half of study participants, at best, into care. Another California Bridge Project study focused on 325 out-of-treatment HIV-positive people who averaged 1.5 years since HIV diagnosis and their first meeting with project staff.³⁰ Almost three quarters of these people were nonwhite, and half were men who have sex with men. Case workers managed to link only 29% of this group to care -- after an average 15.4 client contacts. Although these outreach programs linked half or fewer people to HIV care, the success rate surely reflects the hard-to-reach populations that the programs targeted.

An outreach program in New York City targeted 161 HIV-positive residents of single-occupancy hotels, 95% of whom were minorities and 59% of whom were active drug users.³¹ Ninety-five study participants were assessed before receiving the intervention, while 66 were assessed after receiving the intervention. These people had better baseline access to care than the groups in the studies summarized above.^28-30 Three quarters of the pre-intervention group and 91% of the postintervention group already had a regular health care provider.

In the pre-intervention approach, an outreach worker went door-to-door in eight single-occupancy hotels and asked residents if they needed services and wanted to join a harm-reduction program. The intervention consisted of adding a physician to the door-to-door outreach team and asking residents if they wanted to see a physician right now. Comparing data from preintervention and postintervention interviews, multivariate analysis that accounted for drug use, HIV severity, and other factors determined that the intervention independently raised chances of having a regular provider (OR 5.3, P = 0.02), taking antiretrovirals (OR 5.7, P = 0.02), and having a better perception of quality of care (OR 4.9, P = 0.003).

Besides case management, outreach, and colocation of testing and care services, a few studies have pinpointed specific tactics that may cut the time between HIV diagnosis and care (Table 3). The already-discussed University of Alabama HIV clinic study found that every 10 days between the call to make a first clinic appointment and the appointment date magnified the no-show risk about 30%.¹⁸ Other research shows the importance of having appointment times convenient for patients and having providers who speak the patient's language.^32,33 In a review article on improving US women's access to care, Mariam Aziz and Kimberly Smith of Chicago's Rush University Medical Center stressed the need to create a "woman-friendly environment" that offers child care and access to "case management, social workers, and gynecologic care, at a minimum."³⁴

Table 3. Getting Newly Diagnosed People Into HIV Care: Strategies That Work

After Michael Mugavero and University of Alabama colleagues figured out why HIV-positive people failed to show up for their first HIV clinic appointment,¹⁸ they devised a program, Project CONNECT, to help solve the problem.³⁵ Newly diagnosed people are scheduled for a clinic orientation visit within 5 days of their first call for an appointment (see Figure 1 in the Mugavero interview). That visit includes a semistructured interview, a psychosocial questionnaire, baseline lab tests, and (for the uninsured) a visit with a social worker. Patients needing substance abuse or mental health services get a prompt referral to appropriate services. Because the clinician has lab results and other data at the first patient encounter, care can begin immediately. During the first year that Project CONNECT was in place, the clinic's no-show rate dropped from 31% to 18% (P < 0.01).² CONNECT cut the risk of failure to establish HIV care almost 50% (OR 0.54, 95% CI 0.38 to 0.76).

San Francisco General Hospital (SFGH), which cares for large populations of gay men and poor, homeless, or uninsured people, created the Positive Health Access to Services and Treatment (PHAST) system to get newly diagnosed people into care.^36,37 All SFGH care settings use rapid HIV testing and a central diagnostic lab that pages positive results to a PHAST worker, who meets patients when they get a positive result.

The PHAST team member provides intensive on-the-spot support and education, schedules confirmatory testing, and performs clinic intake including CD4 count, viral load, and resistance testing. PHAST also helps newly diagnosed people with insurance applications and provides appointment reminders and primary care until the patient is transferred to a permanent HIV provider.