Late HIV Diagnosis: Bad Medicine and Worse Public Health

The author discusses a new study that models the impact of the CDC's new recommendation of opt-out HIV testing in all health care encounters in the US for persons 13-64 years old.


Perspectives
June 2007 | Volume 4 | Issue 6 | e200 I t is estimated that there are some 252,000-312,000 Americans living with HIV yet unaware that they carry this deadly virus [1]. Once they develop symptoms of HIV disease, most will be diagnosed. But by then, their immune systems will have experienced permanent damage, and thousands will have inadvertently infected their partners. One analysis reckons that annual HIV transmission rates in the United States are some 3.5 times higher among people with undiagnosed HIV infection compared to those who are diagnosed [2], due to the fact that knowledge of positive HIV serostatus is associated with substantial decreases in high-risk sexual behaviors among those so diagnosed [3].
Although HIV remains a very serious medical concern, with the advent of improved antiretroviral treatments it is no longer the hopeless condition it was in the early 1980s. Yet late diagnosis of HIV infection remains a problem in the US three decades after the syndrome was fi rst described, despite the substantial medical and public health benefi ts of early HIV diagnosis. In the US, about 40% of persons who were diagnosed with AIDS in 2005 had their fi rst positive HIV test less than 12 months before their AIDS diagnosis [4]-meaning that they were infected with HIV for years prior to their initial diagnosis. Nor can late HIV diagnosis be simplistically ascribed to lack of access to health care. A variety of studies have shown that failure to diagnose HIV in a timely manner occurs even among persons who have regular access to health care [5][6][7].

"Opt-Out" HIV Testing in Health-Care Settings
Arguably, addressing the persistent problem of late HIV diagnosis was the major driver in the recent push by the Centers for Disease Control and Prevention (CDC) for routine, opt-out testing in health-care settings, i.e., informing patients of the intent to perform HIV testing and inferring consent unless the patient declines [8]. Citing the time constraints of riskassessment and counseling, CDC's 2006 guidelines also note that "prevention counseling should not be required with HIV diagnostic testing or as part of HIV screening in health care settings" [8].
In a new study published in PLoS Medicine, David Holtgrave estimates the overall costs of CDC's routine HIV testing recommendations using standard methods of scenario and costeffectiveness analysis [9]. His fi ndings are provocative: in his scenarios, targeted HIV counseling and testing performed better than opt-out testing in terms of expense, number of newly identifi ed patients, and prevention of HIV transmission [9].
Another salient feature of Holtgrave's analysis is his focus on the role of counseling in the context of HIV testing. He reminds readers that a randomized controlled trial sponsored by CDC in the late 1990s demonstrated that interactive client-centered counseling could reduce risk behaviors and the incidence of new sexually transmitted infections [10]. Although earlier CDC guidelines had already taken steps to streamline requirements for pretest counseling-acknowledging providers' need for fl exibility based on client base, HIV prevalence in the healthcare setting, and available resources [11]-the 2006 guidelines go even further. Prevention counseling is no longer identifi ed as a requirement for HIV screening in health-care settings. One of Holtgrave's scenarios addresses this change, positing an increase in HIV incidence resulting from certain high-risk persons who, in the absence of counseling, might interpret negative test results as endorsing the "safety" of their behaviors.

Implications of the Study
Perspectives on how best to encourage early diagnosis of HIV infection will continue to evolve as additional program implementation data and improved input parameters become available to further refi ne costeffectiveness analyses. But it would be a serious mistake to narrow this multifaceted discussion to an artifi cial dichotomy of opt-out routine testing versus targeted HIV testing.
The Perspectives section is for experts to discuss the clinical practice or public health implications of a published article that is freely available online.

Ronald O. Valdiserri
Funding: The author received no specifi c funding for this article.

Competing Interests:
The author wishes to declare that he was employed by the United States Centers for Disease Control and Prevention at the time that the guidelines discussed in this article were developed and was involved in their review. Also, the author has in the past collaborated professionally on several occasions with Dr. Holtgrave. The opinions expressed in this commentary are those of the author and do not necessarily represent the views of the US Department of Veterans Affairs. Scenario and cost-effectiveness analyses found that for the same programmatic cost, targeted counseling and testing would diagnose more people living with HIV and prevent more HIV infections than opt-out testing.
A broader, and perhaps more relevant, question worthy of further study is "How can we confi gure and sustain health-care systems that are capable of promoting and incentivizing necessary prevention services such as early HIV diagnosis?" [12]. These systems should be able to address the needs of individual patients, such as those patients who might require more time to accept HIV testing, even in an opt-out scenario, or those patients who receive a negative HIV test result, but require active intervention to facilitate changes in ongoing high-risk behaviors. In our current systems of care, limited provider time and competing demands are well-recognized obstacles to the optimal delivery of a host of clinical preventive services-not just HIV counseling and testing [13]. And although prioritization of clinical preventive services at a population level is a practical strategy to allocate scarce prevention resources [13], it may not serve the prevention needs of individuals or subgroups.
Future research on the timely diagnosis of HIV infection must include operational studies and systems analyses that explicitly evaluate the impact of changes in the design of health-care delivery systems (including referral systems) on the receipt of lifesaving services like HIV testing and, when needed, ongoing risk reduction counseling [14].