Last April, we reported on the failure of Truvada, an oral anti-HIV pill, to prevent infection in African women. The results of the trial were disappointing, and surprising, because Truvada had been found to prevent infection in 90% of gay men who took it religiously. We pointed out at the time that the researchers had yet to analyze blood samples they'd taken from the women in the study. Those samples would show whether the women had been taking the drug as prescribed, which would suggest that its failure was due to some biological factor, or whether they had been failing to take the drug. It looks like it's the latter. This week, the NYTimes reports, the researchers announced at an AIDS research conference in Seattle that of the women who got infected, only a quarter of them had any Truvada at all in their blood. The fact that nonadherence, rather than biological factors, caused the drug fail is important to know. Truvada might one of a large number of therapies that can help if taken properly but meet with clear failure when they are not. Antibiotics for tuberculosis, for example, are very effective if a regimen is completed, but patients frequently fail to take all the pills, because they start to feel better or because they just forget, something we can all sympathize with. These failures result in prolonged illness and in the creation of multidrug-resistant tuberculosis strains. For that reason, one of the most important breakthroughs in tuberculosis treatment in recent decades is making patients report to clinic to take their pills while a nurse or doctor is watching. Why the women were not taking Truvada is the next puzzle the researchers will have to investigate. Could it be that stigma of AIDS makes having a bottle of anti-HIV medication around the house socially awkward? Or did these healthy women not have a strong enough impetus to take the drug? As the example of tuberculosis shows, getting people to take medication---especially, perhaps, medication that's meant to prevent infection rather than cure it---can be as big a hurdle as developing a drug that works in the first place.