Ling Hsu

Late diagnosis of HIV infection : Trends, prevalence, and characteristics of persons whose HIV diagnosis occurred within 12 months of developing AIDS

Background:Persons diagnosed late in the course of HIV infection may be unknowingly transmitting infection and once diagnosed may have worse outcomes and greater medical expenses. Methods:Persons diagnosed with AIDS in San Francisco between 2001 and 2005 were included. Late testers were persons diagnosed with HIV 12 months or less before their AIDS diagnosis. Prevalence trends, demographic and risk correlates, and predictors of late testing were measured. Results:Among 2139 persons included, 830 (38.8%) were late testers. The prevalence of late testing was stable between 2001 and 2005. Late testing was more likely among persons <30 years old (Odds ratio [OR]: 1.99, 95% confidence interval [CI]: 1.4, 2.8), heterosexuals (OR: 1.88, 95% CI: 1.1, 3.1), persons without a reported risk (OR: 2.88, 95% CI: 1.7, 5.0), persons with private insurance (OR: 1.82, 95% CI: 1.4, 2.4), no insurance (OR: 1.83, 95% CI: 1.4, 2.4), born outside of the United States (OR: 1.64, 95% CI: 1.2, 2.2), and whose initial AIDS diagnosis was an opportunistic infection (OR: 2.24, 95% CI: 1.8, 2.8). Conclusions:A large proportion of persons with AIDS have tested late in the course of HIV infection and this proportion has not declined in recent years. Routine testing in medical settings, and use of rapid oral-fluid testing in traditional and nontraditional settings may increase early HIV diagnosis.

Disparities in engagement in care and viral suppression among persons with HIV.

Background:Engagement across the spectrum of HIV care can improve health outcomes and prevent HIV transmission. We used HIV surveillance data to examine these outcomes. Methods:San Francisco residents who were diagnosed with HIV between 2009 and 2010 were included. We measured the characteristics and proportion of persons linked to care within 6 months of diagnosis, retained in care for second and third visits, and virally suppressed within 12 months of diagnosis. Results:Of 862 persons included, 750 (87%) entered care within 6 months of diagnosis; of these, 72% had a second visit in the following 3–6 months; and of these, 80% had a third visit in the following 3–6 months. Viral suppression was achieved in 50% of the total population and in 76% of those retained for 3 visits. Lack of health insurance and unknown housing status were associated with not entering care (P < 0.01). Persons with unknown insurance status were less likely to be retained for a second visit; those younger than 30 years were less likely to be retained for a third visit. Independent predictors of failed viral suppression included age <40 years, homelessness, unknown housing status, and having a single or 2 medical visits compared with 3 visits. Conclusions:Socioeconomic resources and age, not race or gender, are associated with disparities in engagement in HIV care in San Francisco.

Declining incidence of AIDS-defining opportunistic illnesses: results from 16 years of population-based AIDS surveillance.

Objective:To measure the incidence and risk factors of AIDS-defining opportunistic illnesses (AOIs) in the pre-highly active antiretroviral therapy (HAART) (1993–1995), early-HAART (1996–2000), and late-HAART (2001–2008) periods. Design:Prospective cohort analysis of AIDS surveillance data. Methods:Individuals living with, or diagnosed with AIDS from 1993 through 2008 were included. Poisson regression models were used to estimate annual incidence rates of the eight most frequently occurring AOIs, and to compare these rates in the pre-HAART (1993–1995), early-HAART (1996–2000), and late-HAART (2001–2008) periods. Results:There were 18 733 individuals with AIDS included; 5788 were diagnosed prior to 1993 and 12 945 were diagnosed between 1 January 1993 and 31 December 2008. The incidence rates of Pneumocystis jiroveci pneumonia, wasting syndrome, Kaposis sarcoma, HIV encephalopathy, cytomegalovirus retinitis, cytomegalovirus, and esophageal candidiasis decreased during the study period, with the largest declines observed between the pre-HAART and early-HAART periods. Incidence rates also decreased between the early-HAART and late-HAART periods, though not as sharply. Incidence rate reductions between the earliest and latest period ranged from 84 to 99%. Conclusions:Steep declines in incidence of AOIs were found following the introduction of HAART and continued into the late-HAART era. These declines reflect the impact of HIV diagnosis and treatment on a population level.

Mortality Risk After AIDS-Defining Opportunistic Illness Among HIV-Infected Persons—San Francisco, 1981–2012

OBJECTIVE To examine whether improved human immunodeficiency virus (HIV) treatment was associated with better survival after diagnosis of AIDS-defining opportunistic illnesses (AIDS-OIs) and how survival differed by AIDS-OI. DESIGN We used HIV surveillance data to conduct a survival analysis. METHODS We estimated survival probabilities after first AIDS-OI diagnosis among adult patients with AIDS in San Francisco during 3 treatment eras: 1981-1986; 1987-1996; and 1997-2012. We used Cox proportional hazards models to determine adjusted mortality risk by AIDS-OI in the years 1997-2012. RESULTS Among 20 858 patients with AIDS, the most frequently diagnosed AIDS-OIs were Pneumocystis pneumonia (39.1%) and Kaposi sarcoma (20.1%). Overall 5-year survival probability increased from 7% in 1981-1986 to 65% in 1997-2012. In 1997-2012, after adjustment for known confounders and using Pneumocystis pneumonia as the referent category, mortality rates after first AIDS-OI were highest for brain lymphoma (hazard ratio [HR], 5.14; 95% confidence interval [CI], 2.98-8.87) and progressive multifocal leukoencephalopathy (HR, 4.22; 95% CI, 2.49-7.17). CONCLUSIONS Survival after first AIDS-OI diagnosis has improved markedly since 1981. Some AIDS-OIs remain associated with substantially higher mortality risk than others, even after adjustment for known confounders. Better prevention and treatment strategies are still needed for AIDS-OIs occurring in the current HIV treatment era.

Do people who develop AIDS within 12 months of HIV diagnosis delay HIV testing

Objective. Individuals diagnosed with AIDS within 12 months of HIV diagnosis have been considered “late testers.” Prevalence estimates of late testers have been made using HIV/AIDS surveillance data, and high rates of late testing have been reported. However, studies evaluating this definition have not been conducted. We measured the degree of misclassification of delayed testing based on this surveillance definition of late testing. Methods. We used dates of negative HIV tests among people who met this definition of late testing in San Francisco from 2007 to 2008 to reclassify people as “verified non-late testers” if there was a negative HIV test within five years of HIV diagnosis, as “verified late testers” if there were no prior tests or if negative tests were recorded five or more years prior to diagnosis, or as “late-tester status not verified.” We measured misclassification of late-tester status and the prevalence of late testing using the different definitions of late testing. Results. Of the 270 people who developed AIDS within 12 months of HIV diagnosis, we found that 89 (33.0%) were verified late testers, 131 (48.5%) were verified non-late testers, and 50 (18.5%) were unverifiable. Using the surveillance definition (individuals who develop AIDS within 12 months of HIV diagnosis), the prevalence of late testing was 26.3%, whereas it was 9.0% when restricted to individuals verified as late testers. Conclusion. Defining people who develop AIDS within 12 months of HIV diagnosis without taking into consideration the dates of prior negative HIV tests leads to substantial misclassification of late testing.

Evaluation of a non-name-based HIV reporting system in San Francisco

Objective: To develop and evaluate a non‐name‐based HIV reporting system. Methods: A population‐based study of the accuracy of a set of non‐name codes and a prospective study of a laboratory‐initiated HIV surveillance system conducted at a county hospital (site 1) and a health maintenance organization (site 2). Participants were persons reported with AIDS in San Francisco and patients with a positive test result for HIV antibody, p24 antigen, viral load, or a CD4 count at the study sites. Results: Proper match rate was 95% for records with complete codes and records with at least 50% of the codes. Proper non‐match rate was 99% for records with all code elements and 96% for records with at least 50% of the elements. Completeness of reporting was 89% (site 1) and 87% (site 2). Median number of days between test and receipt of test report at the health department was 9 days at site 1 and 7 days at site 2. During 1999, 78% of HIV‐infected patients at site 1 and 87% at site 2 had an HIV‐specific laboratory test. Conclusions: A non‐name‐based laboratory reporting system for HIV is feasible.