Dr. Wong suggests that we did not model the impact of false-positive HIV test results in our analysis. He also suggests that the principle of patient autonomy effectively argues against the implementation of any mandatory screening program.

Testing for HIV is a two-step process, so that one must apply the operating characteristics sequentially. The combined specificity is on the order of 0.9999926 [1, 2]. Thus, the number of possible false-positive results out of 1.6 million screened women is approximately 11, not 16 000 as suggested by Dr. Wong. Second, and probably more important, any patient with positive results on Western blot should have HIV RNA levels measured to determine the viral burden or to confirm HIV infection in someone with a suspected false-positive result. The HIV RNA or other available tests would probably identify any false-positive results. Therefore, the theoretical problem of a false-positive result should not affect our analysis.

Although the principle of patient autonomy is of great importance in most ethical dilemmas, it may not be decisive in any ethical conflict involving two parties. This is especially true when one of the parties cannot speak (that is, the fetus) and the best way of preventing HIV infection of the newborn is therapy during pregnancy. However, what often gets lost in the debate about mandatory testing is that although testing is a critical first step, it is not the only consideration. The most important part of the process is the counseling and subsequent care, which will maximize the likelihood of a woman getting and taking treatment and completing the entire course.

We agree with Dr. Wong's conclusions that mandatory screening should not be public policy, although we disagree with some of his reasons. The focus should move from discussion of mandatory testing to the greater need for excellent counseling and care for all pregnant women, HIV infected or not.