NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Chou R, Huffman L. Screening for Human Immunodeficiency Virus: Focused Update of a 2005 Systematic Evidence Review for the U.S. Preventive Services Task Force [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2007 Apr. (Evidence Syntheses, No. 46.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Screening for Human Immunodeficiency Virus

Screening for Human Immunodeficiency Virus: Focused Update of a 2005 Systematic Evidence Review for the U.S. Preventive Services Task Force [Internet].

Show details


After reviewing new studies on HIV screening, we found insufficient evidence to change the main conclusions of our 2005 evidence synthesis. Specifically, the 2005 evidence synthesis found no direct evidence on the effects of HIV screening on clinical outcomes. However, it found good evidence that HIV screening can accurately identify infected persons and screening appears acceptable to most patients, though a significant proportion of persons are likely to decline routine screening, frequently because of a low perception of risk. Many patients are currently diagnosed at advanced stages of HIV disease, and a significant proportion does not receive test results or enter into medical care. Identification and treatment of asymptomatic HIV infection at immunologically advanced stages of disease can result in marked reductions in clinical progression and mortality, particularly with the use of HAART. Risk factor assessment can identify persons at increased risk of infection, though targeted screening would miss a significant number of infected persons with unidentified or unreported risk factors. The 2005 USPSTF and 2006 CDC recommendations are generally in agreement on each of these points.

For this brief update, we found no new studies directly evaluating the effects of HIV screening on transmission risk. The major difference between the 2005 USPSTF and the 2006 CDC recommendations appears related to conflicting interpretations of the strength of indirect evidence on effects of screening on HIV transmission. To support recommendations for routine HIV testing in most adolescents and adults, the CDC cites a new meta-analysis16 that found that HIV-infected persons reduce their sexual high-risk behaviors substantially when they become aware of their infection, and a modeling study22 predicting significant reductions in new HIV infections based on these estimates. However, we identified flaws in the studies included in the meta-analysis cited by the CDC, including reliance on self-reported changes in behavior, lack of defined inception cohorts, significant attrition, poorly described methods for assessing sexual risk behaviors, and unclear blinding status of outcomes assessors. These methodological shortcomings probably led to overestimates of reductions in sexual risky behaviors. The reliability of models for estimating effects of reductions in risky sexual behaviors on transmission is also unknown, because of the difficulty of capturing accurately all the factors that can affect transmission, such as sustainability of behavior changes, effects on different types of behaviors, differences in baseline risk that could affect likelihood of testing and subsequent behavior changes, and effects of screening on those testing negative).

We also found that the cost-effectiveness of universal screening in low-risk, low-prevalence (<=0.3%) settings remains uncertain because cost estimates may be sensitive to transmission benefits. Excluding transmission benefits, one study found one-time screening in the general population cost >$100,000 per quality-adjusted life-year.56 In addition, although routine screening appears cost-effective (<$50,000 per quality-adjusted life-year) relative to no screening when population prevalence of undiagnosed HIV infection is >=0.5%, the incremental cost-effectiveness of routine screening relative to targeted screening based on risk factors in low-prevalence settings has not been evaluated. Studies in federally-funded testing sites indicate that most patients at higher risk for HIV infection are likely to be identifiable through risk factor assessment even in lower-prevalence settings.35 Studies that assess cost-effectiveness of routine versus targeted screening would be substantially more informative for a comparison of these strategies in low-prevalence settings. In addition, future cost-effectiveness analyses should include estimates of long-term harmful cardiovascular effects of HAART, which appear to increase over time.88

Although the 2006 CDC recommendations advise screening high-risk persons at least annually and testing patients and prospective partners before they initiate new sexual relationships, we found no new clinical studies evaluating optimal frequency of screening. A cost-effectiveness analysis, however, suggests that repeat screening at any interval costs more than $100,000 per quality-adjusted life-year in low-prevalence and low-incidence settings; however, the study did not incorporate potential beneficial effects on transmission.56 Annual screening was not cost-effective (≥$100,000 per quality-adjusted life year) even in high-risk settings (prevalence of undiagnosed HIV infection 3% and incidence 1.20%). Screening every three years may be cost-effective (<$50,000 per quality-adjusted life-year) when prevalence is at least 1%, incidence is at least 0.09%, and potential transmission benefits are included.55

We also found no new studies evaluating the effects of routine screening in low-risk populations on the proportion of HIV-infected persons diagnosed at later stages of disease, on test notification rates, entry into care, or uptake of recommended interventions. Nor did we find new studies evaluating the effects of opt-out testing without prevention counseling in non-pregnant, low-risk persons, as advised in the 2006 CDC recommendations. One reason this is important is because currently available studies on the effects of HIV screening on risky behaviors have included standard pre-test counseling and risk assessment as suggested in previous CDC recommendations. The effects of eliminating routine prevention counseling on risky behaviors (either positive or negative) are unknown. The lack of data on screening uptake, linkage to care, and stage of diagnosis in low-risk, low-prevalence settings may be explained in part by the fact that, until recently, recommendations for HIV testing centered on high-risk and high-prevalence settings, with prenatal screening the major exception. Although opt-out testing appears acceptable to most pregnant women, it is not clear if such data are applicable to non-pregnant persons, because a high value is placed on preventing HIV infection in newborns and there is strong evidence that interventions are effective for preventing mother-to-child transmission.89

In summary, there remains no direct evidence on benefits of screening for HIV infection in the general population. In the 2005 evidence synthesis, we found screening likely to be beneficial in average-risk patients, based on the effects of HAART in patients meeting criteria for treatment; however, a large number of patients would need to be screened to prevent one case of HIV disease progression or death. We estimated that the numbers needed to screen to prevent one case of clinical progression or death after three years ranged from 1,210 to 13,800 in the general population (Table 2).1 At that time, we were unable to estimate effects of HIV screening on rates of secondary transmission. We concluded that evidence showing decreased secondary transmission following screening would greatly strengthen the case for routine screening in the general population. This is particularly important because expanded screening programs are likely to identify more HIV-infected patients at earlier stages of disease. In such patients, who are less likely to qualify for HAART, other beneficial effects—such as decreased transmission rates—assume greater relative importance when considering net benefits from screening.

Table 2. Outcomes Table of Counseling and One-time Screening for HIV Infection After Three Years in 10,000 Asymptomatic Adolescents and Adults.

Table 2

Outcomes Table of Counseling and One-time Screening for HIV Infection After Three Years in 10,000 Asymptomatic Adolescents and Adults.

After reviewing new evidence available since June 2004, we still cannot estimate the effects of screening on HIV transmission rates. Studies directly linking screening to decreased rates of transmission would require very large populations with long duration of follow-up, and are difficult to perform. Not surprisingly, no such studies have been published since the 2005 evidence synthesis was completed. Although reductions in risky behaviors are reported following HIV testing and counseling, self-reported behavior changes (an intermediate outcome) may be unreliable, and studies have probably overestimated reductions in risky sexual behaviors. In addition, even if estimates of changes in self-reported behaviors are accurate, predicting their effect on transmission rates remains problematic because of the complex relationships between risky behaviors, HAART use, differential effects in those testing negative compared to those testing positive, and other factors that affect transmission.

The 2005 USPSTF recommendations for HIV screening consist of testing based on individual risk factors as well as testing based on clinical setting, including criteria based on local prevalence of infection. For risk-based testing to be maximally effective, more studies are needed on effective and efficient methods of risk assessment and on ways of improving testing rates in those assessed as being at high risk. With regard to prevalence-based testing, the 2005 USPSTF recommendations cite the 2001 CDC threshold of 1% to routinely test, though recent cost-effectiveness studies suggest that a significantly lower threshold may be appropriate. Even if the threshold for testing is lowered, a persistent challenge for prevalence-based testing is that local prevalence data are often not readily available for practicing clinicians. One approach could be for clinicians to institute routine testing unless local prevalence data is available to guide further testing—a strategy advocated by the 2006 CDC recommendations.3

By eliminating the need for risk assessment or local prevalence information, universal testing is theoretically less burdensome for clinicians and easier to put into practice, though studies assessing implementation of routine opt-out testing in low-risk, low-prevalence settings are not yet available. Another potential effect of routine testing is to decrease the stigma associated with HIV screening and misperceptions about who may be at risk. However, the acceptability of routine testing and rates of test uptake in low- or average-risk adults and adolescents has not been evaluated. Even in higher-risk settings, less than one-third of patients were tested after implementation of routine voluntary screening programs.34, 38, 39, 41 Though difficult to quantify and based primarily on anecdotal evidence, potential harms from implementing routine screening in low-prevalence settings have also been identified,1 including anxiety, labeling, adverse effects on close relationships, and a higher proportion of false-positive results. For example, rapid HIV tests are suggested in the 2006 CDC recommendations as a method for increasing test result notification rates, particularly in settings in which continuing relationships with patients do not exist.3 However, based on the sensitivity and specificity of rapid tests, the calculated positive predictive value prior to confirmatory testing in a population with an HIV prevalence of 0.2% is 50%,90 though actual false-positive rates may vary and in some settings are substantially lower than predicted.91, 92 Laws in certain states, mandating specific informed consent or extensive pretest counseling, constitute an important barrier to implementing the 2006 CDC recommendations for streamlined, routine voluntary testing.93

Despite continuing educational efforts and longstanding recommendations for screening of high-risk persons and settings, HIV incidence remains steady in the U.S. Reducing the rate of new HIV infections is an important public health goal, and more effective implementation of screening programs could be an integral method for achieving this aim.94 Studies comparing outcomes between general and targeted HIV screening in low-prevalence settings are urgently required to help clarify the advantages and disadvantages of these alternative approaches. To increase the success of all screening programs, more studies are also needed on methods for increasing HIV test uptake rates, particularly among high-risk persons, and for improving entry into care, reducing risky behaviors, and increasing use of recommended interventions in those testing positive.23


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (219K)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...