The Bayesian model on human papillomavirus vaccination in Italy lacks transparency.

To the Editor:

We refer to a recently published article1 that shows a new Bayesian method, applied to assess a vaccination strategy preventing human papillomavirus (HPV)-related diseases. The article basically describes a model for the economic evaluation of the quadrivalent HPV vaccine in Italy, concluding that it is a cost-effective strategy. Although any model, Bayesian or Frequentist, should be “populated” with reliable data,2 we felt some concern about many “inputs” regarding the Italian setting that could weaken the authors’ conclusions. We have listed some of the main ones.

Real data on Italian vaccination coverage are referenced by an abstract,3 without specifying that this refers to a very small region in Italy (Basilicata, 0.97% of the whole Italian population). Thereafter, table 1 refers for vaccination compliance and coverage to another article, published in Italian, focussed on the efficacy of the quadrivalent vaccine.4

Data on health states associated with HPV-related diseases refer to another abstract,5 then unspecified Italian utility weights for health states were applied, but to our knowledge, no utility tariffs have been validated so far in Italy.

Utilities of cervical cancer, genital and cervical lesions, all refer to an article on the costs of varicella-related hospitalizations in table 1.6

The vaccine price is not consistent with published data,7 and we could not find the figure used as a mean (€69.13, see table 1) in the references.8,9

More in general, the authors state that the cost-effectiveness of the quadrivalent vaccine is proven, ignoring the other, bivalent vaccine against HPV. As 3 recent critical reviews10–12 on economic evaluations regarding HPV vaccines—not cited in the article—concluded that long-term models on HPV vaccination lack transparency in key assumptions and methodological choices, we wonder whether the results of this model (producing a “virtual” follow-up of 90 y) can really be considered more reliable than the others already published.