Those in care for HIV infection are strikingly different from the general U.S. population—they are disproportionately male, black, unemployed, poor, and lacking private health insurance.

The cost of care, $6.7 billion or about $20,000 per patient per year, is less than one percent of all direct personal health care expenditures in the United States. That is "not excessive," considering that HIV infection accounts for about 7 percent of the total potential years of life lost in the U.S., and suggests that the crisis in expenditures for patients with HIV is "one of financing, not cost."

These are the principal conclusions in the initial report from the HIV Cost and Services Utilization Study (HCSUS), the first large-scale effort to collect information on a nationally representative sample of people in care for HIV infection. The paper appears in today's issue of The New England Journal of Medicine.

Sponsored by HHS' Agency for Health Care Policy and Research (AHCPR), other federal agencies, and several private-sector partners, HCSUS is being led by RAND in conjunction with a broadbased consortium of private and public research institutions. The co-directors of the project are Dr. Samuel A. Bozzette of RAND, the University of California, San Diego, and the Veterans Affairs San Diego Healthcare System, and Dr. Martin F. Shapiro, of RAND and the University of California, Los Angeles. Bozzette is the lead author and Shapiro the senior author of today's paper.

The paper is based on data from the study's opening round of interviews with 2,864 patients who were randomly selected to accurately represent the study's "reference population"—adults with known HIV infection who received medical care during the first two months of 1996 in the contiguous 48 states. In addition to excluding adults treated exclusively outside the continental U.S., the study excluded adults treated only at military, prison, and emergency room facilities and also excluded children.

The study team found that there were 231,400 persons in the reference population. From this, the team estimates that 292,000 to 372,000 adults with known HIV infection saw a doctor at least once during any six-month period (the definition of regular medical care) in 1996, with the best estimate being 335,000. This compares to the Centers for Disease Control (CDC) estimate that 650,000 to 900,000 Americans are infected.

Here are some other key findings of the study (references to "patients" are to HIV-infected adults under care):

AIDS: 59 percent of patients met the CDC case definition for AIDS, the most advanced stage of HIV disease. This represents about 85 percent of all adults thought to be living with "full-blown" AIDS.

Care settings: 30 percent of patients received care at major teaching hospitals, the balance from office-based physicians and community hospitals and clinics. One third made at least one visit to an emergency department and 20 percent were hospitalized during the six months prior to interview. Their stays accounted for about one percent of all hospital days in the U.S.

Costs and types of care: Hospital care accounted for 43 percent of patients' direct medical care costs, pharmaceuticals for 39 percent, emergency department care for 2 percent, other outpatient care and associated costs for 15 percent.

Pharmaceutical use: 85 percent of patients used at least one HIV medication and 79 percent used an antiretroviral drug during the six months prior to being interviewed. The pattern of use of protease inhibitors and non-nucleoside reverse transcriptase inhibitors—recently developed drugs that are highly effective in suppressing HIV infection—changed rapidly during 1996. At the beginning of the year, 16 percent of patients had used one of the newer drugs. By December, that proportion shot up to 55 percent.

Population characteristics: 77 percent of patients were men and 89 percent were less than 50 years old. About half were non-Hispanic whites, one third were non-Hispanic blacks, and almost one sixth were Hispanic. Women were more likely than men to be young, black, less educated, unemployed, impoverished, and underinsured. But they were also less likely to have AIDS. Forty eight percent of patients had more than a high school education and 37 percent were employed. Forty six percent had annual household incomes of less than $10,000, 72 percent had less than $25,000 per year.

Insurance coverage: 20 percent of patients had no health insurance, proportionately only a fifth higher than the general U.S. population. But 29 percent of patients were insured only by Medicaid, a proportion three times greater than in the general population, and 19 percent had Medicare coverage, a proportion nine times greater than is typical for persons under 65 years old. Private insurance was most common in the Midwest and West while being uninsured was most common in the South.

Regional variations: The Midwest had the smallest caseload, with 11 percent of HIV-infected adults, the South had the largest with 36 percent, and the Northeast and West had 25 percent and 28 percent respectively. Note that these are regional caseloads, not population-based rates.

Risk factors: 49 percent of patients were men who had had sex with men but no injection-drug use, 24 percent reported injection-drug use with or without other risk behaviors. The remainder included 18 percent who reported only heterosexual sex and 9 percent who reported no known risk factors.

"These findings send two powerful messages," summarized Dr. Bozzette, an infectious diseases specialist and HIV expert. "First, it is deeply disturbing that up to two-thirds of persons with HIV infection are not getting regular care and that even fewer are getting highly active anti-HIV therapy. Clearly, we need to do better in identifying the HIV infected, getting them into care, and financing that care, particularly for the increasingly affected poor and minority communities. Second, our data explode the widespread belief that care for the HIV infected is extraordinarily costly. Although it is a large and growing burden on the public sector, HIV care is less expensive than care for many serious diseases."

"The HCSUS will address many additional questions before its scheduled end in mid-1999," said Dr. Shapiro. "These include disparities in access to care across socioeconomic and racial groups, the impact of variation in insurance coverage, the extent of mental health and substance abuse disorders, oral health, rural care, and the extent to which HIV is becoming resistant to antiretroviral drugs."

"This unique study will help public health planners and legislators to make data-based critical decisions that will affect HIV care into the new century," said AHCPR Administrator Dr. John M. Eisenberg.

Dr. Claude Earl Fox, administrator of HHS' Health Resources and Services Administration (HRSA), a major HCSUS co-sponsor, stated: "The data clearly underscore the critical issues HRSA has faced for the past decade in integrating the growing need for HIV/AIDS care into all our programs for low-income and medically underserved individuals. Our continuing challenge is to improve access to care by effectively reaching out to people as early in their infection as possible, and by getting them into high-quality care.

Besides AHCPR and HRSA, other HHS agencies supporting HCSUS include the Office of the Assistant Secretary for Planning and Evaluation, the National Institute of Mental Health, the National Institute on Drug Abuse, the National Institute on Aging, the National Institute of Dental and Craniofacial Research, and the Office of Research on Minority Health of the National Institutes of Health. Additional support was provided by the Robert Wood Johnson Foundation, Merck and Company, Glaxo-Wellcome, Inc., and Quest Diagnostics.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis.

##