I’ve come to expect robust skeptical discussion in online forums like this one, but I was encouraged last week to witness some of it “in the field”, as an anthropologist might say. I was doing my regular morning grade school drop-off, when I chanced to overhear a debate between my son’s kindergarten teacher and the mother of one of his classmates. It was a debate about the risk of an anaphylactic emergency at school, and what caught my attention was that — against stereotype — it was the mother clamoring for science and reason.
She was complaining about the new school requirement that her child, who is allergic to peanuts, wear an EpiPen waist-pouch during the daily half-hour lunchtime recess. The child already had an EpiPen stored in the school office, in accordance with Sabrina’s Law (discussed below), and the mother was loathe to purchase an extra, especially when she believed exposure to the elements would make the rather expensive pen degrade 3-4x faster. After discussion with her son’s allergist, she was confident that in the unlikely event of a reaction during recess, there would be ample time to retrieve the pen from the office and administer it within the required time frame, which she said was about 5 minutes after onset of symptoms.
The teacher disagreed, citing not just safety concerns, but the fact that she has two allergic children in her class and in the event of a double reaction, 5 minutes might not be enough if she first had to retrieve the pens from the office. Hogwash, said the mother, the likelihood of one lunchtime reaction is small enough — two concurrently is infinitesimal.
Now I was well aware that death by peanut allergy is an extremely low risk, but as tempting as was to simply use the mother’s statements to feed my own confirmation bias, I decided to resist that urge. Partly because I heard the talk track in my own head that started, “This is ridiculous, back when I was a kid…” and recognized it for the appeal to antiquity that it was. But mostly it’s because from a policy perspective, the school’s response to risk must rely not only on raw probabilities, but also on the cost to lower them further. In other words, if it’s cheap enough, policymakers may well strive to make rare risks even rarer.
Anaphylactic Allergies, By the Numbers
We’re fortunate to have up-to-the-minute, Canadian data in the form of the SCAAALAR study just completed by researchers at McMaster & McGill, and sponsored by Health Canada and the AllerGEN research network. According to its findings, 1.68% of Canadian children are probably allergic to peanuts, 90% of which have had a severe reaction.
While that sounds bad — and indeed the study points out that Canadian food allergy rates are higher than those reported in most other countries — the reality is that the severest of consequences (death) is extremely rare. A study by Anaphylaxis Canada found only 32 food allergy related deaths in Ontario in the 15 year span from 1986-2000. Twenty of these were from peanuts and tree nuts, and only 6 occurred at a school or camp. That last number is the relevant one here, suggesting that the annual school day mortality rate across Ontario is about 0.4, or one death every 2.5 years.
Policy Response, On the Cheap
At least that was the situation before Sabrina’s Law, enacted in 2005, which formalized the requirement for school boards in Ontario to establish and maintain an anaphylaxis policy that reduces the risk of exposure to anaphylactic allergens; train teachers on the risk and how to respond; and keep detailed case files and EpiPens in the school for students with anaphylactic allergies.
The Toronto District School Board fulfills this legislation via its Operational Procedure PR563, but leaves it up to individual schools how to implement its requirements. Many schools, including my son’s, are completely nut-free, going far beyond what Sabrina’s Law dictates. In addition to not bringing any nut-based products, children are also not allowed to share food, so it’s a pretty restrictive regime.
There are two key things to note about these policy interventions. First, they’re cheap: other than training teachers, which may require an outside trainer once a year, all of them can be enacted with existing resources. The additional nut-free policy is completely free, putting all of the burden (really just an inconvenience “cost”) on parents. That’s certainly the right price for making a low risk even lower.
Second, there’s no way for an individual to achieve the same aims without such a policy. For example, an allergic student cannot themselves remove exposure risks or make teachers better prepared to deal with a reaction.
Quantifying the Reduction
Since Sabrina’s Law has only been in effect for 5 years, and the nut-free policies in many schools are even more recent, we won’t have actual data to show how much of a reduction we’ve seen for some time yet. However, we can make a reasonable estimate by analyzing each policy and the likelihood that it will make an impact.
While it’s tempting to think that the nut-free policies will reduce the risk to zero, that’s not likely the case. First, remember that nuts were the cause of only 62.5% (20/32) of the food-allergy related deaths in the study above. Second, we’ve got to discount that figure further for non-compliance with the policy, inadvertent or otherwise. Let’s say compliance with the “no-nuts” policy is 80%, and the “no-sharing” policy removes another half of the non-compliance risk. That means the whole policy might remove a total of (62.5% x 90% =) 56% of the risk, bringing the mortality rate down to one death every 5.7 years.
And that’s before taking into account Sabrina’s Law itself, especially the storage of student EpiPens and training of teachers to use them. It’s not a panacea. Given the time-sensitive nature of administration, and the possibility of human error, some proportion of students won’t get it in time. The Anaphylaxis Canada study provides some (albeit weak) indication of what this percentage might be, saying that 4 patients had epinephrine close by at the time of the reaction, and 2 died while trying to get to it. While these numbers aren’t large enough to provide statistical validity, 50% certainly seems like a reasonably conservative ballpark. If we’re correct, that brings the the death rate down to one death every 11.4 years.
Put another way, these policies should lower the number of deaths in each K-12 cohort from 5 to 1. Four children saved every 11.4 years isn’t a lot compared to interventions like seat-belts, but given the low cost, it’s hard to argue that the effort shouldn’t be made.
Getting back to the debate that started it all, let’s take a look at the statements made by the mother and the teacher, and see if they hold water. There are essentially two claims we need to consider:
- That the incremental benefit of wearing the waist pouch during recess justifies its costs; and
- That this is especially true because of the risk of multiple concurrent reactions.
We’ll take #2 first, and chalk up an easy win for the teacher. When looking at the probability of concurrency, you can only multiply probabilities when the two events are independent of one another. In this case, it’s not at all clear that they are, since the same anaphylactic allergen (say, a surreptitiously shared Snickers bar) could be the cause of both reactions. While that’s not the only scenario, it’s considerable enough that we should expect the likelihood of two concurrent reactions to be not much less than the risk of one. Nice work, Teach!
Claim #1 is a bit more nuanced, but just as easy to quantify. We first need to know what percentage of the overall risk can be attributed to that 30 minute period. It’s tempting to try to divide 30 minutes into a 6.5 hour school day, but that’s likely misleading, because the risk is not spread evenly over that time period. The majority of risk of food allergy can be expected to cluster around periodic “feeding events”. Let’s say there are 5 of those during an average day (3 meals, 2 snacks), two of which occur during the school day (1 meal, 1 snack), and one of those during or proximate to the relevant half-hour period. That means that ~50% of the exposure risk is in this period. But we’re not getting rid of that whole exposure risk, since as we saw above, the existing policies result in an EpiPen being administered half the time, so it’s just the other half that we’re reducing by 50%. That means that a policy of recess EpiPen carriage would reduce the mortality rate by another 25%, from one death every 11.4 years, to one death every 15.2 years.
Now we said above that interventions in low risk scenarios may be undertaken if the cost is cheap enough, but is that true here? Recall that the cost of the interventions above were not only absolutely low, but also spread out across the entire Ontario school-age population of 2.3 million kids, whereas the cost of this intervention falls squarely on the mother. An extra EpiPen is about $100 and lasts a year when stored properly. The mother contended it degrades 3-4x faster when exposed to the elements in the pouch, though I was unable to fully verify this contention. The manufacturer insert does say it should be stored at room temperature (25 degrees), with “excursions” permitted between 15 and 30 degrees, and that the pen should be replaced if the liquid is discolored. So it certainly seems plausible that a more rapid replacement schedule could be required due to extreme exposure, but the research from use by paramedics suggests otherwise.
So we’ll take the conservative position and say that the annual incremental cost of this intervention is merely the $100 to purchase the second EpiPen. Considering that it provides such a modest reduction in an already infinitesimal base rate, and since it’s something that an individual can do themselves even without such a policy, I’d have to fall on the side of making this an optional intervention at best. So chalk this one up for Mom.
It’s great to see some rational debate over risk in the schoolyard, and to hear parents standing on the side of the math. In the end, it seems clear that the parent is right in this case, but that’s not necessarily true with the prior interventions like Sabrina’s Law and nut-free schools, which make a rare risk even rarer at an appropriate price point.
As a parent of a (thankfully) non-allergic child, I may still grumble as I scour the ingredients list of every granola bar I pack, but I can’t argue that my being slightly inconvenienced is worth a child’s life.