标题：Prevalence of HIV Infection Among Inpatients and Outpatients in Department of Veterans Affairs Health Care Systems: Implications for Screening Programs for HIV
摘要：Objectives. We sought to determine the prevalence of HIV in both inpatient and outpatient settings in 6 Department of Veterans Affairs (VA) health care sites. Methods. We collected demographic data and data on comorbid conditions and then conducted blinded, anonymous HIV testing. We conducted a multivariate analysis to determine predictors of HIV infection. Results. We tested 4500 outpatient blood specimens and 4205 inpatient blood specimens; 326 (3.7%) patients tested positive for HIV. Inpatient HIV prevalence ranged from 1.2% to 6.9%; outpatient HIV prevalence ranged from 0.9% to 8.9%. Having a history of hepatitis B or C infection, a sexually transmitted disease, or pneumonia also predicted HIV infection. The prevalence of previously undocumented HIV infection varied from 0.1% to 2.8% among outpatients and from 0.0% to 1.7% among inpatients. Conclusions. The prevalence of undocumented HIV infection was sufficiently high for routine voluntary screening to be cost effective in each of the 6 sites we evaluated. Many VA health care systems should consider expanded routine voluntary HIV screening. Early identification of HIV infection through screening substantially lengthens the life of the person identified and provides an important public health benefit from reduced HIV transmission. 1 HIV screening in health care settings is also cost effective, even when the prevalence of HIV infection is as low as 0.05% to 0.1%. 1 – 3 Newly revised guidelines for HIV screening from the Centers for Disease Control and Prevention (CDC) recommend one-time screening in all health care settings unless the prevalence of undiagnosed HIV infection is documented to be less than 0.1%. 4 The prevalence of undiagnosed HIV infection has been documented in few health care settings in the era of highly active antiretroviral therapy. The Sentinel Hospital Study, which assessed HIV prevalence in a probability sample of nonfederal US hospitals, showed that the prevalence of HIV infection ranged from 0.1% to 7.8% among sentinel hospital populations (inpatients and outpatients combined) in 1989; however, currently, few people with HIV are hospitalized, so both inpatient and outpatient prevalence rates may differ from previous estimates. Although total HIV prevalence has been evaluated in many settings, the most important factor in determining the usefulness of an HIV screening program is the prevalence of unidentified, rather than known, HIV infection. The preferred method for determining an unbiased estimate of prevalence in a population is a blinded (anonymous) serological survey, an approach recommended by the CDC. 6 , 7 In a blinded serological survey, blood that is drawn for other purposes is stripped of identifiers and tested for HIV. Because the patient’s identity cannot be determined, informed consent is not required. Other approaches that depend on patients’ acceptance of screening are limited by selection bias that occurs when patients decline testing. 8 – 10 This selection bias can be substantial, and the prevalence of HIV among patients who have declined testing may be either higher or lower than the prevalence among patients who accept testing. 11 Blinded serological surveys have been used widely, most notably in the Sentinel Hospital Study. 5 We conducted a blinded serological survey to determine the prevalence of HIV infection among outpatients and inpatients in 6 Department of Veterans Affairs (VA) health care systems. The VA is one of the largest integrated health care systems and the largest provider of HIV care in the United States. Our goal was to assess the prevalence of both documented and undocumented HIV infection and to assess demographic and clinical predictors of documented and undocumented infection. Although predictors of undocumented HIV infection have been evaluated in emergency departments, 12 , 13 they have not been well studied in VA populations.
