Unhealthful News 185 - Cancer: screening is generally a bad idea; what about vaccines? (Part 1)

Yesterday, and several earlier times in this series, I wrote about how cancer screening tests have been shown to be inefficient (i.e., too expensive for the benefit they provide, as compared to competing uses of resources) or even more harmful than beneficial (a bad idea, even if they are free).  This is not quite always true, and refers only to mass screening of people with no symptoms and who are otherwise not believed to be at higher than average risk.  Cervical cancer screening is cheap and easy, and detects incipient cases of the disease rather than just finding growing cancers a bit earlier.  Most other cancer screening looks pretty bad on close inspection.  Maybe three or four mammograms between ages 50 and 65 are worthwhile.  

Finding a vaccine for cancer and administering it to everyone is possibly a different story.  About two weeks ago, the US government finally got around to recommending the HPV vaccine for boys.  It has been recommended for girls for a while because it was recognized that just about all cases of cervical cancer are caused by sexually-transmitted papilloma virus strains that the vaccine can prevent.  But beyond cervical cancer, it has long been recognized that HPV apparently causes many other cases of epithelial cancers (i.e., the body parts that are in contact with the outside world), specifically the bits that experience sexual contact, the oral cavity, esophagus, and anus.  So some commentators have asked "what took so long?"

That is indeed a good question.  There are a few answers that occur to me, one is fairly compelling, one is very compelling but no one seems aware of it, one is interesting and understandable but misguided, and two are deplorable.

The fairly compelling answer is that the vaccine is extremely expensive and prevention is probably of fairly little value.  Recall that screening is usually too expensive to justify; this is primarily because it is expensive to screen millions of healthy people who do not have the disease to find the few who have it.  It is also expensive to vaccinate millions of people who will never get cancer from HPV to save only the few who will.  However, when the prevention happens, in contrast with screening, there are no costs of false positives, no further treatment cost, and no damage from the disease at all.

Can screening ever look so good?  Not for cancer, given the high cost of treatment, but for other diseases it can.  The example that comes to mind is syphilis screening, which is mandatory in the US for pregnant women, in the sense that if they do not get the recommended screen during pregnancy then it is a mandatory part of medically supervised delivery.  Of all the screening tests I have ever done cost-benefit calculations for, this is the only one that came out clearly positive (note: I have never run the numbers for cervical cancer screening).  The reason that it is done is that syphilis transmitted to a newborn, if undetected, can be devastating.  But the benefits go beyond that, since non-symptomatic syphilis is unlikely to be detected, and so when it is detected the mother and her partner(s) are treated, eliminating those cases and all those they would cause in the future.  This almost certainly lowers the equilibrium prevalence of the disease in the population by a lot, perhaps by half or more.  But what makes this such a bargain is the treatment is relatively harmless and dirt cheap (a simple antibiotic shot), the cost of a false positive is equally low, and the cure is near certain and complete.  Unlike most cancer screening, this is a very good deal.

But a problem with the current HPV vaccine is that even though it does not entail further treatment like screening does, it is still very expensive.  It costs payers hundreds of dollars, in contrast with the few tens of dollars it costs to get a flu shot.  Merck charges $300 for the drug itself, and there is clinical time for three injections, plus whatever markup the medics add.  (Incidentally, the official claim is that the cost for vaccinating just the boys as they come of age would cost $140 million/year.  But since there are 2 million boys in every US one-year age cohort, this seems low by about a factor of five.  The estimate must be based on the assumption that few will follow the recommendations.  Or someone is lying. I will leave the investigation of that discrepancy to the news reporters.  Hahaha - just kidding.  They will never pursue it, and probably not even think enough to realize there is a problem.  Unfortunately, I will probably not investigate it either -- if anyone does, please clue me in with a comment.)

What almost none of the unhealthful news reports bother to mention, however, is that this huge total expenditure is mostly not real resource costs.  That is, it does not consume actual stuff or labor, but rather just moves money from one entity to another.  Because drugs always are priced at greater than their manufacturing cost, because the price has to amortize development costs, and especially because the US does not negotiate drug prices like every other country and thus allows monopoly profit on top of that, that $300 price undoubtedly includes a huge profit for Merck.  While we might not want to give money to Merck, at least it is not an actual cost.  The same is true for the physicians' fees for administering the injections.

A more important and compelling argument against the vaccine than purchase price is that this vaccine prevents cancers that, except in very rare cases, will occur at least 35 years in the future, and often decades more than that.  Those of you who know a bit about these analyses might point out that any proper cost-benefit analysis discounts future costs and benefits, so this delay is already accounted for.  But that process of discounting assumes that events will occur, and merely adjusts for the fact that we weight the present more heavily than the future in our decision making.  But is that a reasonable view?  Should we assume that oral cancer or cervical cancer (this observation applies to the girls too) will still be expensive or deadly disease that far in the future?  I certainly hope not.  Four decades might not get us to the point that a single injection will reliably eliminate the disease, as with syphilis now, but it seems safe to assume that there will be much better treatment than currently exists.  

Remarkably, I have never seen anyone else make this observation, other than me and some of my students (who were thinking this through as an assigned exercise with the advantage, compared to most people writing about such topics, of having been taught what to think about).  The strongest argument against this vaccine is that we should really hope it will not matter much for those who are currently young enough to benefit from it.  Consider how much the value of an expensive syphilis vaccine for children would have been overestimated in, say, 1930 (it became easy and pretty cheap to treat once penicillin was developed and proven in the early 1940s).

That is the good argument against the vaccine.  But I have never seen anyone make it, so it cannot explain what took so long.  In Part 2, I will look at the not-so-good reasons that probably do explain it.