A Data Visualization and Dissemination Resource to Support HIV Prevention and Care at the Local Level: Analysis and Uses of the AIDSVu Public Data Resource

BACKGROUND

AIDSVu is a public resource for visualizing HIV surveillance data and other population-based information relevant to HIV prevention, care, policy, and impact assessment.

OBJECTIVE

The site, AIDSVu.org, aims to make data about the US HIV epidemic widely available, easily accessible, and locally relevant to inform public health decision making.

METHODS

AIDSVu develops visualizations, maps, and downloadable datasets using results from HIV surveillance systems, other population-based sources of information (eg, US Census and national probability surveys), and other data developed specifically for display and dissemination through the website (eg, pre-exposure prophylaxis [PrEP] prescriptions). Other types of content are developed to translate surveillance data into summarized content for diverse audiences using infographic panels, interactive maps, local and state fact sheets, and narrative blog posts.

RESULTS

Over 10 years, AIDSVu.org has used an expanded number of data sources and has progressively provided HIV surveillance and related data at finer geographic levels, with current data resources providing HIV prevalence data down to the census tract level in many of the largest US cities. Data are available at the county level in 48 US states and at the ZIP Code level in more than 50 US cities. In 2019, over 500,000 unique users consumed AIDSVu data and resources, and HIV-related data and insights were disseminated through nearly 4,000,000 social media posts. Since AIDSVu's inception, at least 249 peer-reviewed publications have used AIDSVu data for analyses or referenced AIDSVu resources. Data uses have included targeting of HIV testing programs, identifying areas with inequitable PrEP uptake, including maps and data in academic and community grant applications, and strategically selecting locations for new HIV treatment and care facilities to serve high-need areas.

CONCLUSIONS

Surveillance data should be actively used to guide and evaluate public health programs; AIDSVu translates high-quality, population-based data about the US HIV epidemic and makes that information available in formats that are not consistently available in surveillance reports. Bringing public health surveillance data to an online resource is a democratization of data, and presenting information about the HIV epidemic in more visual formats allows diverse stakeholders to engage with, understand, and use these important public health data to inform public health decision making.

Quality of care for HIV infection provided by Ryan White Program-supported versus Non-Ryan White Program-supported facilities

BACKGROUND

The Ryan White HIV/AIDS Care Act (now the Treatment Modernization Act; Ryan White Program, or RWP) is a source of federal public funding for HIV care in the United States. The Health Services and Resources Administration requires that facilities or providers who receive RWP funds ensure that HIV health services are accessible and delivered according to established HIV-related treatment guidelines. We used data from population-based samples of persons in care for HIV infection in three states to compare the quality of HIV care in facilities supported by the RWP, with facilities not supported by the RWP.

METHODOLOGY/PRINCIPAL FINDINGS

Within each area (King County in Washington State; southern Louisiana; and Michigan), a probability sample of patients receiving care for HIV infection in 1998 was drawn. Based on medical records abstraction, information was collected on prescription of antiretroviral therapy according to treatment recommendations, prescription of prophylactic therapy, and provision of recommended vaccinations and screening tests. We calculated population-level estimates of the extent to which HIV care was provided according to then-current treatment guidelines in RWP-supported and non-RWP-supported facilities. For all treatment outcomes analyzed, the compliance with care guidelines was at least as good for patients who received care at RWP-supported (vs non-RWP supported) facilities. For some outcomes in some states, delivery of recommended care was significantly more common for patients receiving care in RWP-supported facilities: for example, in Louisiana, patients receiving care in RWP-supported facilities were more likely to receive indicated prophylaxis for Pneumocystis jirovecii pneumonia and Mycobacterium avium complex, and in all three states, women receiving care in RWP-supported facilities were more likely to have received an annual Pap smear.

CONCLUSIONS/SIGNIFICANCE

The quality of HIV care provided in 1998 to patients in RWP-supported facilities was of equivalent or better quality than in non-RWP supported facilities; however, there were significant opportunities for improvement in all facility types. Data from population-based clinical outcomes surveillance data can be used as part of a broader strategy to evaluate the quality of publicly-supported HIV care.

Improving the representativeness of behavioral and clinical surveillance for persons with HIV in the United States: the rationale for developing a population-based approach

The need for a new surveillance approach to understand the clinical outcomes and behaviors of people in care for HIV evolved from the new challenges for monitoring clinical outcomes in the HAART era, the impact of the epidemic on an increasing number of areas in the US, and the need for representative data to describe the epidemic and related resource utilization and needs. The Institute of Medicine recommended that the Centers for Disease Control and Prevention and the Heath Resources and Services Administration coordinate efforts to survey a random sample of HIV-infected persons in care, in order to more accurately measure the need for prevention and care services. The Medical Monitoring Project (MMP) was created to meet these needs. This manuscript describes the evolution and design of MMP, a new nationally representative clinical outcomes and behavioral surveillance system, and describes how MMP data will be used locally and nationally to identify care and treatment utilization needs, and to plan for prevention interventions and services.

Injection and syringe sharing among HIV-infected injection drug users: implications for prevention of HIV transmission. Supplement to HIV/AIDS Surveillance Group

Because HIV-infected injection drug users (IDUs) can transmit HIV infection, we investigated factors associated with sharing of syringes in the past year among IDUs infected with HIV. We analyzed data from an interview survey of 11,757 persons > or = 18 years of age with HIV or AIDS between June 1990 and August 1995 who were reported to 12 state or city health departments in the United States. Of the 1527 persons who had ever shared syringes and reported injecting in the 5 years before the interview, 786 (51%) had injected in the year before interview, and of these, 391 (50%) had shared during that year. IDUs who were aware of their HIV infection for >1 year were less likely to share (43%) than those who were aware of their infection for 1 year or less (65%, adjusted odds ratio=2.15, 95% confidence interval, 1.52-3.03). The only statistically significant time trend was that the proportion of IDUs from Connecticut who shared decreased from 71% in 1992 to 29% in 1995. This trend appears to be related to the 1992 changes in Connecticut laws that allowed purchase and possession of syringes without a prescription. Because many HIV-infected IDUs continue to inject and share, prevention efforts should be aimed at HIV-infected IDUs to prevent transmission of HIV. Early HIV diagnosis and access to sterile syringes may be important methods for reducing syringe sharing by HIV-infected IDUs.

Differences in participation in experimental drug trials among persons with AIDS

To measure participation in experimental drug trials among persons with acquired immunodeficiency syndrome (AIDS), we interviewed 4,604 persons at least 18 years of age who were reported to have AIDS to 11 state and city health departments in the United States. Ten percent reported that they were currently in a trial. Current enrollment differed significantly (p < 0.05) by race/ethnicity (blacks, 5%; whites, 14%; Hispanics, 15%), gender (women, 7%; men, 11%), exposure mode (injection drug use, 5%, men who have sex with men, 14%), annual household income (< $10,000, 8%, > or = $10,000, 14%), education (< 12 years, 6%; > or = 12 years, 12%), health care (no regular care, 1%, public care, 8%; private care, 17%), and time since AIDS diagnosis (< or = 6 months, 9%; > 6 months, 12%). Adjusting for all factors and time since AIDS diagnosis, blacks (adjusted odds ratio [AOR] = 0.35, 95% confidence interval [CI] 0.26, 0.47), persons with less than 12 years of education (AOR = 0.71, CI 0.53, 0.96), and those without regular health care (AOR = 0.24, CI 0.10, 0.61) remained less likely to be in a trial. Blacks, those with less than 12 years of education, and persons without regular health care were less likely than other persons with AIDS to be currently enrolled in AIDS trials. To increase enrollment of these persons, researchers must address barriers to participation for these groups.

The types of drugs used by HIV-infected injection drug users in a multistate surveillance project: implications for intervention

OBJECTIVES

This study sought to describe the drugs used by drug injectors infected with human immunodeficiency virus (HIV) and to determine factors associated with the primary injection drug used.

METHODS

A cross-section of persons 18 years of age or older reported with HIV or acquired immunodeficiency syndrome (AIDS) to local health departments in 11 US states and cities was surveyed.

RESULTS

Of 4162 persons interviewed, 1147 (28%) reported ever having injected drugs. Of these 1147 injectors, 72% primarily injected a drug other than heroin. However, the types of drugs injected varied notably by place of residence. Heroin was the most commonly injected drug in Detroit (94%) and Connecticut (48%); cocaine was the most common in South Carolina (64%), Atlanta (56%), Delaware (55%), Denver (46%), and Arizona (44%); speedball was most common in Florida (46%); and amphetamines were most common in Washington (56%). Other determinants of the type of drug primarily injected were often similar by region of residence, except for heroin use. Polysubstance abuse was common; 75% injected more than one type of drug, and 85% reported noninjected drug use.

CONCLUSIONS

Preventing the further spread of HIV will require more drug abuse treatment programs that go beyond methadone, address polysubstance abuse, and adapt to local correlates of the primary drug used.

Risk behaviors of persons with heterosexually acquired HIV infection in the United States: results of a multistate surveillance project

To describe past risk behaviors among persons with heterosexually acquired human immunodeficiency virus (HIV) infection, we interviewed 497 persons > or = 18 years of age with heterosexually acquired HIV infection reported to 11 state and city health departments in the United States. Thirty-nine percent of persons reported using noninjection drugs in the past 5 years; noninjection drug use was highest among men whose sex partners injected drugs (53%). Sixteen percent of all persons used crack, and 17% were classified as potential alcoholics; among men, 29% were classified as potential alcoholics. Of the 49% of men who reported paying a woman for sex, 86% did so multiple times. Most persons had multiple sex partners in the past 5 years; however, 35% of the women had only one sex partner. Thirty-four percent of the women and 50% of the men had been treated for a sexually transmitted disease in the past 10 years. Seventy-four percent of the women and 68% of the men had never used condoms in the 5 years before they knew they were HIV positive. Among these people with heterosexually acquired HIV, noninjection drug use was common, many men have paid someone for sex, and many women have not had multiple sex partners. These findings have important implications for the types of prevention programs that can most successfully lessen the spread of HIV among heterosexuals.

Health insurance coverage among persons with AIDS: results from a multistate surveillance project

To determine factors associated with health insurance coverage among persons with acquired immunodeficiency syndrome (AIDS), we interviewed 1958 persons 18 years of age or older who were reported to have AIDS in 11 states and cities. Overall, 25% had no insurance, 55% had public insurance, and 20% had private insurance. Factors associated with lack of insurance varied by current employment status. Employed persons with an annual household income of less than $10,000 were 3.6 times more likely to lack insurance than employed persons with a higher income. Unemployed persons diagnosed with AIDS for less than 1 year were two times more likely to lack health insurance than unemployed persons diagnosed for a longer time. Making insurance available to persons identified as most likely to lack insurance should improve access to care for persons with AIDS.

Sociodemographics and HIV risk behaviors of bisexual men with AIDS: results from a multistate interview project

OBJECTIVE

To describe the sociodemographic characteristics and sexual and drug use behaviors of men with AIDS who engage in bisexual activity.

METHODS

We interviewed 2120 men aged > or = 18 years who were reported with AIDS in 11 states and cities. Men were considered bisexual if they reported having had sex with a man and a woman in the previous 5 years.

RESULTS

Of the 2020 men with AIDS who reported being sexually active in the previous 5 years, 1150 (57%) had had male partners only, 522 (26%) had had female partners only and 348 (17%) had had both. White men were least likely to report bisexual behavior (15%; 161 out of 1071). Men of Latin American descent were most likely to report bisexual behavior (24%; 37 out of 155), especially those born outside the United States who had lived there for < or = 10 years (38%; 11 out of 29). Bisexual Latin American men, regardless of birthplace, were more likely to be currently married than all other bisexual men (22 versus 7%; P < 0.05). HIV risk behaviors differed between men reporting bisexual and those reporting exclusively homosexual or heterosexual activity. Injecting drug use in the previous 5 years was more common among bisexual than homosexual men (12 versus 6%; P < 0.05). Bisexual men were more likely (P < 0.05) to have received money for sex (11%) than homosexual (4%) or heterosexual men (4%). This difference was even greater among injecting drug users receiving money for sex: bisexual (29%), homosexual (13%), heterosexual (3%).

CONCLUSIONS

Demographics and HIV risk behaviors of bisexual men with AIDS differ from those of homosexual and heterosexual men with AIDS. These findings indicate that special efforts are needed to prevent sexual transmission of HIV among bisexual men.