Tuesday, April 1, 2014

Dr. Jay's Magical Math

You would expect pediatricians, especially Fellows of the American Association of Pediatrics, to know at least a little bit about epidemiology and to give others a proper, fact-based picture of what vaccine-preventable diseases can do. At the very least, you would not expect them to get things so spectacularly wrong that you wonder how they ever managed to get their license, let alone their degree. Yet that is exactly what one pediatrician does on a fairly regular basis. In fact, the things I'm about to describe I actually wrote about three years ago. I'm speaking of Jay Gordon, MD, FAAP, and in the time since that 2011 post, he doesn't seem to have learned anything. You'll see what I mean in a bit.

For those who don't know, Dr. Gordon is a California pediatrician who regularly downplays the risks of disease and advocates alternative vaccination schedules, as well as skipping vaccines altogether as "unnecessary". He is a darling of the anti-vaccine movement, since he supports their views that vaccines may be, somehow, dangerous. In fact, he is (or was) the pediatrician for Jenny McCarthy's son. Jenny, as you may or may not known, made quite a name for herself as the celebrity spokesperson for the anti-vaccine group Generation Rescue. Yet Dr. Gordon also appears to desperately crave the acceptance of his science-based peers.

As I mentioned, Dr. Gordon has this exceptionally annoying habit of downplaying the risks of disease, for example calling hospitalizations due to measles "soft hospitalizations", citing The Brady Bunch to support his notion that measles is not dangerous, and getting the epidemiological idea of "incidence" stupendously wrong. Whenever there is an outbreak of disease, he fairly consistently confuses incidence and prevalence.

It's that latter aspect of his poopooing the risk of disease that prompted this post. Last week, he offered his opinion about the current outbreak of measles in his home state of California, which as of March 28 had reached 49 cases in the state. When asked on Twitter by my friend @EpiRen to explain incidence in the context of the current measles outbreak, here's what Dr. Jay had to say:

The text reads, "@EpiRen for instance, in California, divide new cases of a disease by 38,000,000 to derive incidence rate. This makes no headlines, though" and "@EpiRen This will help you. Measles presently is creating very little risk and a lot of headlines scaring parents. en.m.wikipedia.org/wiki/Incidence..." So, according to Dr. Jay, the current incidence of measles in California is 49/38,000,000, or about 1 in 775,000. In other words, more people have caught measles than will likely die from drowning in the bath within the next year.

Do you see the error that Dr. Gordon made? He is using the total population of California (roughly 38 million) as the denominator to determine incidence. But you can't use the total population when trying to figure out disease incidence. As I explained to him in 2011 and as EpiRen explained to him in 2012 (as well as countless times on Twitter), incidence is the number of new cases divided by the total population at risk. In fact, the link that he included in his tweet states, quite clearly (emphasis added):
Incidence rate is the number of new cases per population at risk in a given time period.
Given that definition, if we assume that Dr. Gordon understands the definition of incidence, it would appear that he is saying that every single Californian is at risk of getting the measles. In fact, he confirms this when EpiRen points out his error:

The text reads: EpiRen: "@JayGordonMDFAAP WRONG! Incidence rate = new cases / at risk. Who is at risk in CA, doc?" Dr. Gordon: "'@EpiRen: WRONG! Incidence rate = new cases / at risk. Who is at risk in CA, doc?' Actually, 38,000,000. Huge denominator. Not just a few."

When EpiRen asks a clarifying question, Dr. Jay confirms he thinks that every single Californian, regardless of immunization status or disease history, is at risk of infection:

Text reads: "@EpiRen: So, by your statements, ALL of California is at risk for #measles?" Of course, Rene. Vaccine is not 100% and measles exists"

By stating that all Californians are at risk, Dr. Gordon implies that the vaccine does not work. But then he contradicts himself a bit later, saying that the vaccine is 90% effective:

Let's leave aside, for a moment, the fact that Jay Gordon, MD, FAAP is wrong about the efficacy of MMR vaccine, since after two doses more than 99% of individuals will develop immunity (and even only one dose is around 95% effective, so he's still wrong). On the one hand, Dr. Gordon says that all Californians are at risk, but on the other, he says that significantly fewer are at risk (i.e., 90% of those who are immunized are not at risk). Let's use that 90% efficacy figure for a moment.

For the 2012-2013 school year, 92.7% of kindergarteners received the full two doses of MMR vaccine. For simplicity's sake, let's round that off to 90% and further grant that that number reflects the vaccination coverage of the entire California population. That gives us 34.2 million people who are immunized against MMR, of which (using Jay's erroneous efficacy number) 30.78 million are protected. That leaves 7,220,000 people unprotected against measles (i.e., "at risk"), assuming we don't have anyone who is immune due to prior infection.

What does that do to our incidence? Instead of 49 out of 38,000,000, we would get 49 out of 7,220,000, or around 1 in 147,000. That's a five-fold greater incidence than Jay would have his followers believe. But, like I said, he's still wrong because he low-balls the effectiveness of the vaccine.

To keep things simple, let's stick with 90% of the population having been immunized. At 99% effectiveness, we get 33,858,000 people protected and 4,142,000 at risk. We now have an incidence of 1 in 84,500, nine times greater than what Jay says is the incidence. More non-immune Californians have caught measles in this outbreak than are likely to be murdered in the next year.

I am, admittedly, oversimplifying the calculations. In addition to making an assumption about the proportion of the population that is immunized, I'm leaving out the number of people who are immune because they got measles the old fashioned way: getting infected, suffering through several weeks of illness and fortunately surviving, with or without complications. Those people would bring the total at risk down even more. I'm also assuming an even distribution of protected and at-risk individuals. In reality, there are pockets where there are greater numbers of unvaccinated individuals. The long and short of it is that the incidence rate is much, much larger than Dr. Jay would like you to believe.

The point of all of this is to show how Dr. Jay Gordon, a Fellow of the American Association of Pediatrics and who should know better, either just does not understand incidence despite being told numerous times over several years, or intentionally uses the wrong numbers to try to downplay the impact of disease. By constantly mischaracterizing disease outbreaks in this fashion, Dr. Gordon does a grave disservice to his patients and his community. Comments like his are irresponsible coming from a physician and are likely to convince parents that they don't actually need to get their child immunized when there is an active outbreak in their area. That could lead to the further spread of outbreaks, not to mention the costs to taxpayers associated with investigating and controlling the them. I don't think people should panic, but they should get immunized if they haven't been. Pediatricians should do their job, urging their patients to get vaccinated.

And when Jay Gordon, MD, FAAP says idiotic things like this:
The measles outbreak of 2014 does not pose a risk to your healthy child.
he is being grossly irresponsible and failing to uphold his responsibility as a pediatrician and as a member of the AAP. His advice puts others at risk.


  1. You really expect better from a doctor.

  2. The new California kindergarten uptake rates have just come out. I looked at the data for just Orange County. There are a total of 42,781 children enrolled in kindergarten (public and private, in schools having more than 10 students in the kindergarten class). Of those, 3,795 lack 2 doses of MMR (the requirement for school entry). Of those 3,795, some subset may have had one dose -- the data doesn't tell us that, just the number lacking 2 doses. That's 8.9% -- not great, but as I said before, may include a large cohort that have had one dose, but not the required two.

    Orange County has a total of 1,545 students with personal belief exemptions, or 3.6%. That might be a better number to work with in estimating the number of kindergarteners with 0 doses of MMR.

  3. As you mention, the rate is FAR higher as the number of susceptible persons is further decreased by the fact that the population group born prior 1957 (or 1970 depending on jurisdiction) are assumed to be immune from prior natural measles. This is a huge chunk of people that should rightly be removed from the denominator. His math is astoundingly bad and betrays a complete lack of either epidemiological knowledge or professional dishonesty, or both.

  4. Jay is, quite simply, an idiot.

  5. I am not sure if Jay is an idiot or is deliberately deceptive. After all, if moms decide to vaccinate their babies, then they don't need Jay to be their pediatrician - they are free to find a much better one. Either way, he should have his AAP membership and medical license revoked.


You need a Google, LiveJournal, WordPress, TypePad, AIM or OpenID account to comment. If you need one of those, click on the appropriate link. Spam comments will be deleted. Comments on posts older than 30 days will go into moderation, due to the activities of commercial spammers.