So once again this year, we have an apparent “shortage” in the supply of influenza vaccine. For some reason, this is considered a big deal. And for some reason, there seems to be very little discussion that’s I’ve ever noticed on the basic numbers on influenza infection, death, hospitalization, etc. Many dramatically and obviously false claims are pumped through the media as common-knowledge fact, and an air of mild hysteria seems to be surrounding the issue. Even on what I generally find to be generally informed and rational discussion fora, such as Eric’s excellent Educated Guesswork, there seems to be a substantial lack of basic information on which to base discussion. This is going to be a bit of a brief attempt at a survey of relevant bits of literature on the subject, interspersed with my own rantings.

Morbidity

First, I think it’s important to have an understanding of what the casualty rates and other costs of influenza are, in order to understand what we’re trying to control. One oftens hears how “36,000 people on average die each year of influenza in the US”, or I’ve even heard more than once on the TeeVee “36,000 americans die every year from the flu”. Any person of moderate education will know that if the latter were true, or the former true in the way it’s implied (which is basically the same sense as the 2nd), that everybody would know at least a couple of people who had died of influenza. I don’t know anyone who’s died of influenza. I’ve asked my wife, and she doesn’t either. If you’re slightly more educated, you’ll also know that immediately following the first world war, there was a severe worldwide influenza pandemic which killed millions. And then you’d ask yourself, since those deaths are obviously included in the “average”, how many people typically die each year. Well, I have some answers.

The American Lung Association in August published a survey[pdf] of morbidity from pneumonia and influenza in the US, covering 1979-2001. That’s a nice snapshot of recent, non-catastrophic years. Table 4 reproduced here in part, with a couple summary columns and rows, gives a nice breakdown for each year by age:

The annual median number of deaths is 1137, of which 85% are over 65, and 2% are 15 or younger. About 10 infants a year die from influenza, which is probably comparable to the death rate among infants from lightning strikes. Also worth noting is that in the most at-risk population, those 85+, deaths from influenza account for 2.2 per 100,000 of population.

Costs

Kids

A decent survey and collections of statistics for school-aged children, with to my mind dubious assumptions and therefore dubious results, is here. Assumptions they make which I consider dodgy are:

• Cost of vaccine. They give a price of $4 or $10 as the basis numbers used, depending on method of administration of the vaccine, either individually or in a group. I’ve read from reliable source (via other reliable source) that a 10-dose vial of vaccine (presumably used for group administration of vaccine) costs a pharmacy $67, or $6.70 per dose. I would estimate that the decrease in cost associated with bulk purchases for a school district or county would very likely be offset by the increased bureacracy cost in operating such a program (let’s say 30% = $2 savings per dose, times 50,000 doses for a school district of medium size = $100,000 which is about what you’d probably pay for the salary of the county under-manager for flu vaccination and her 5 underlings). Add the other costs of administering these doses (nurses, school administration, lost teaching time, etc) and you’re probably up around the $20 per dose mark that they use at the top of their sensitivity analysis.
• “Best case” efficacy — they quote a 75% effectiveness rating referencing a study which found a 50% effectiveness rating. Not sure where they get 75% from. Most of the studies I dug up where actual effectiveness was measured (as opposed to antibody production at detectable levels) seem to indicate something in the ballpark of 50% as pretty well normal. 56% which they use as “base case” actually seems a little high. Table 3 in the study shows that variability in the efficacy of the vaccine is critical in determining the cost savings here. More work should really have been done in trying to find a better number for this parameter.
• The cost-of-lost-work-by-caretaker number is calculated in a way likely to push it higher than it probably really is. It does not factor in the likelihood that in a dual-parent household, the lower wage earner (whether male of female) is probably more likely to stay home to care for the infant. They provide “It is certainly possible that an employed male in the household might take off work to be with a sick child rather than a woman. However, we used the salary of working women rather than working men to provide a more conservative estimate of the indirect costs of influenza”, however this doesn’t address the fact that while the average salary of men is certainly higher than the average for women, not all men make more money than their spouses. The cost-of-lost-work also assumes that infection rates for children by influenza are distributed uniformly across the parental income scale, by using average numbers for the incomes of parents, and average employment levels (97% for fathers!) for parents. In reality, I imagine that disease is heavily tilted towards the lower end of the economic scale, and that they probably therefore have a heavy overestimate of the average income of employed parents, a heavy overestimate of the number of families with 2 parents (and thus overestimate child->parent infection numbers), and a heavy underestimate of the unemployment rate of those parents. These combined assumption errors will result in an increased direct loss of income calculated, and an increased loss of income due to secondary infection of the parents.

Ok, so overall, if we take these assumption corrections into consideration, we’d probably find that flu vaccination in a group setting in fact does not come near the $34.79 number they promote in their headline, but is more likely quite near break even.

Adults and old people

I’m prepared to believe that vaccinating healthy adults is not very useful, which seems to be the consensus. Old people, it’s probably worth it, since they tend to get a lot of complications from flu infection, and rack up the hospital costs as a result, whether they end up dying or not. (Medical Care Of Old People Is Damned Expensive left as a topic for another day).

Conclusions

Ok, took a while to pull all that together, so I haven’t gone in and pulled out more that I was interested in pursuing. But that’s good for now. Bottom line: the media hysteria telling you you’re going to die if you don’t get the flu vaccine is crap. As is the suggestion that it’s definitely a good idea to vaccinate kids universally. I think though that it probably is a good idea to vaccinate “at risk” people, like those with existing respiratory complications like asthma or being old. But in the end, getting the flu is to my mind just a part of growing up, and something that builds character.