I was delighted to receive the letter from Goldie et al. because they have played such a key role in modeling the effects of various approaches to cervical cancer prevention. I was, however, disappointed to find that many of the articles they referenced were iterations of a modeling exercise using population-based data and epidemiological data. I have not seen new information from field tests supporting their claim that our study was inaccurate.
Goldie et al. mention that we are biased in our citing of the data. True, we used a very favorable early estimate from their research about the possible benefits of screening. On the other hand, the findings of many of their recent HPV models for developing countries feature estimates based on a price per vaccinated girl of $10—not per dose of vaccine. Because each girl needs three doses and the market price is around $120 per dose (not to mention program costs), this surely qualifies as a best-case scenario.
Goldie et al. criticize our argument for “ignoring the lack of data for VIA [visual inspection with ascetic acid] at repeated intervals,” but there is more long-term information on the effects of screening and treatment than on the effects of HPV vaccination.
Goldie et al. rightly note that loss to follow-up has a huge effect on projections. Interestingly, they reiterate their estimate of 15% loss per visit, which we also used in our article. Yet I fail to see how loss to follow-up can be similar for a three-visit vaccine and a one-visit screen-and-treat approach.
Finally, the major points of our article do stand and deserve to be repeated. In developing countries (and even in developed ones) many women continue to die from HPV infections already contracted, and many of them could be saved by being screened at least once (better still, twice or thrice) starting in their 30s. Even if HPV vaccine programs were up and running tomorrow, most of those women would not be helped. Would we accept a child health program that would take effect in 20 years?
This is not an either-or position, but rather one of setting priorities. Reaching consensus is a worthy goal, but even more important is continuing open discussion of the evidence supporting various approaches, especially because the lives of millions of adult women are at stake.