BACKGROUND/RATIONALE:
Early identification of HIV infection is a crucial step that enables HIV-infected veterans to receive the maximum benefit from HIV care. Without early detection of HIV, many veterans will be diagnosed at advanced stages of HIV infection when irreversible immunologic damage has occurred. Currently, the VA has no guidelines on who should be screened for HIV. This study is part of a national VA initiative, the Quality Enhancement Research Initiative (QUERI).
OBJECTIVE(S):
Our research has the following specific aims: 1) to assess existing practice patterns for the identification of established HIV infection; 2) to evaluate the cost -effectiveness of published guidelines for the identification of established HIV infection when applied to VA populations; 3) to develop the scientific basis for HIV screening guidelines; and 4) to perform a pilot emonstration project that evaluates the usefulness of interventions to increase HIV screening rates.
METHODS:
To evaluate existing HIV testing practices patterns, we performed retrospective chart reviews on tested patients, and evaluated testing practice in at-risk patients at 4 VA medical centers. We also performed a blinded serologic survey at 5 VA medical centers to determine the prevalence of HIV. We used state-of-the-art methods, including computer-based reminders to implement a preliminary HIV screening guideline in a pilot study. We will evaluate the cost-effectiveness of current published guidelines, as applied to the VA. We will develop the foundation for an evidence-based guideline for HIV screening.
FINDINGS/RESULTS:
Our analysis examining testing practices and appropriateness of testing at four VA sites has been completed. Our findings showed that among 13,991 patients identified as at risk for HIV, 5076 (36%) had been tested for HIV; testing ranged from 32% to 40% by site (p<0.01). HIV testing was more likely in patients aged 30-39 OR 4.7 [3.7-5.9]), with hepatitis C or B (p<0.01), and who used cocaine or opiates (OR 1.6 [1.4=1.8]). The probability of being tested ranged from 4% to 93% depending on risk factors. Our review of 1100 medical records found that the prevalence of HIV varied from 1% to 20% among tested patients at the 4 sites. The proportion of tests deemed appropriate varied from 78% to 95% (p<0.01), with an overall rate of 90%. HIV infection was more likely in patients with history of opportunistic infection (adjusted odds ration OR 4.5 [1.8-11.5]), in men who have sex with men (p<0.01), and varied by site. Among tested patients at the VA, the vast majority of HIV testing was appropriate. In contrast, among patients at risk for HIV, almost two-thirds had not been tested. Further testing of this group could markedly benefit patients who have unidentified HIV infection by providing access to antiretroviral therapy.
The pilot implementation of a preliminary HIV screening guideline to determine its effectiveness in improving HIV screening and testing practices showed that computer-based reminders with feedback did not increase screening rates in our primary-care practices. Among patients who were tested, most were tested appropriately. However, 27% of patients who were not tested had HIV risk behaviors, a finding that highlights the importance of risk assessment. Because our providers strongly believed they had insufficient time to perform risk assessment, use of ancillary personnel may be a more successful strategy for increasing rates of screening.
We tested 3,750 unique outpatient and 3,575 unique inpatient specimens for our blinded serologic survey. Of the 7,325 patients, 314 (4.3%) had HIV infection. Inpatient prevalence ranged from 1% to 7% between centers, outpatient prevalence ranged from 1.7% to 8%. Of patients who tested positive, from 3% to 43% did not have documentation of known infection (absolute prevalence of undocumented HIV infection 0.14% to 2.3%). The HIV prevalence at all of our sites was substantially higher than the 1% prevalence at which routine screening for HIV is recommended by national guidelines. Although this finding indicates that these sites should consider routine voluntary HIV screening, the lower prevalence of undocumented HIV suggests that further evaluation of the cost-effectiveness of screening should also be a high priority.
Our analysis of the cost-effectiveness of screening at the VA is ongoing.
IMPACT:
Our study will develop the scientific basis for guidelines for the early identification of HIV infection among veterans. This project will develop the scientific basis of a screening guideline for the VA.
PUBLICATIONS:
Journal Articles
- Owens D, Sundaram V, Lazzeroni L, Douglass L, Tempio P, Sanders G, Holodniy M, Shadle V, McWhorter V, Agoncillo T, Haren N, Chavis D, Jensen P, Simberkoff M, Bozzette SA. HIV Testing appropriateness and predicators of HIV infection in Department of Veterans Affairs health care systems. Medical Decision Making. 2002; 22: 534-534.
