nprotected oral sex is not well recognized as a risk factor for transmitting the human immunodeficiency virus, yet as evidence of its potential risk accrues, physicians must confront the question of how to counsel patients about oral sex risk. How do health care professionals feel about broaching this topic with their patients? Can the psychological literature help physicians understand their patients' risk perceptions and motivate healthful behaviors? In this article, we review the literature on oral sex risk, discuss barriers to communicating caution to patients, and propose recommendations and specific counseling strategies based on risk perception and behavior change theory that we believe will help physicians effectively intervene in the prevention of HIV transmission.
REVIEW OF THE LITERATURESince the advent of the AIDS epidemic, researchers have sought to define categories of risk for transmission of HIV in sexual behavior. Although receptive anal and vaginal intercourse have been clearly classified as the highest-risk behaviors, and kissing and masturbation as behaviors with the lowest risk, 1 the level of risk for oral sex remains a matter of debate. A variety of research methods-human and animal, epidemiologic studies and case reports, in vitro and in vivo-have yielded a growing literature addressing the question of oral sex risk. The comparative riskiness of oral sex in epidemiologic studies attempting to relate behavior to HIV seroconversion has been difficult to assess. These studies attempt the nearimpossible task of isolating specific sexual practices in research subjects, who usually engage in more than one type of risk behavior, and rely on retrospective self-report data, which can involve recall bias and responses tainted by social desirability (for commentary on problems in sexual behavior research, see R. C. Lewontin, Sex, lies and social science, The New York Review , April 20, 1995:24-9).Subsamples of subjects engaging only in oral sex are likely to be too small to provide adequate statistical power to assess the relative risk. 2,3 Some investigators have asked subjects only about oral sex practices or have lumped oral sex together with other behaviors. 4 In addition, a myriad of uncontrolled factors-dates of infection, number of oral sex events, whether or not oral cuts or abrasions were present-prevent the clear quantification of risk. Most of these epidemiologic studies have sampled homosexual or bisexual men exclusively; accordingly, even less is known about the relative risk of oral sex in the context of lesbian or heterosexual practices.Results from epidemiologic studies have generally been interpreted as evidence against oral sex risk, relative to other behaviors. 3,[5][6][7][8] These studies are based on retrospective self-report data from questionnaires or interviews about risk behaviors over a previous time period, and are also longitudinal in that they follow initially HIV-negative men and test these men over time to see whether or not they undergo seroconversion. Laboratory stud...