92 Percent? 99 Percent?

Josh Barro has a useful piece drilling down on quite how effective Truvada is in preventing HIV-infection. It’s a dense statistical article and worth reading in full. The basic gist is that while a big study found zero HIV infections among those taking Truvada as prescribed, the sample size is too small to declare it 100 percent effective. So they use statistics to come up with the 99 percent figure. One reason to qualify it: there may have been other factors preventing HIV infection in those taking the pill regularly (like having fewer sex partners) than those who didn’t adhere closely to the regimen. If you add in those not taking the medication every day, but almost every day, you can get a statistic of 92 percent effectiveness.

Does this matter? Psychologically, there is a difference between a one percent chance and an eight percent chance of getting infected. But in reality: not so much. Anything well above 90 percent (and indeed, so far as we have actually measured, 100 percent) is a huge deal, especially since regular condom use only seems to lower the chance of HIV transmission by 70 percent. But it’s worth reminding ourselves of the statistical nuances here.