In a recent editorial for The Lancet, pediatric epidemiologist Shamez Ladhani argued that vaccinating kids for COVID is a fairly low-benefit intervention and questioned its relative effectiveness. Ladhani acknowledges there is a net reduction in severe disease but says the fiscal and opportunity costs may be too high. He thus asks those in favor of vaccination: what are we trying to achieve?
I argue there is plenty to be gained. When this debate started last fall, the benefits were obvious. We reasonably believed during delta predominance that many infections could still be prevented, that MIS-C was a significant threat, that breakthroughs were mild and much less disruptive, and that the effect on social and institutional stability was large. The emergence of Omicron meant most people got infected regardless, and MIS-C seems to be greatly reduced with this variant.
Omicron’s effect on the cost-benefit ratio is central to Ladhani’s argument. He argues the main benefit now is the reduction of severe disease rates, which for kids are already low (in absolute terms). His primary source is a nationwide study in Italy capturing 644 severe cases and two deaths from January - April 2022.
I’m not sure how useful this is for the sake of precision as the extrapolated hospitalization rates yields 61,320 total hospitalizations in the US and 225 deaths. We’ve actually had 80,000-130,000 hospitalizations and 1000+ deaths, and that’s without every child having been infected. The extrapolation is poor because as one of his sources notes: “Due to a low number of severe events observed, our estimates…have poor precision” (quote refers to vaccination but the same logic applies). The other of Ladhani’s three rate sources says that for 78% of US hospitalizations COVID was the primary cause. As an aside, he misquotes their rate as a case-hospitalization rate when in reality it is a cumulative hospitalization rate (easy mistake).
That said, I’m OK to take his hospitalization rate at face value: 84 per 100,000 infections. This reads like a tiny number and Ladhani leverages that: “I mean, how low can you go?🤷🏻♂️” But there are two problems with doing this: 1) most kids will get COVID so 84 per 100,000 children is more likely and 2) kids hospitalizations rates are low for everything.
In the US for example, there are about 2 million pediatric hospitalizations per year discounting birth. When we look at the top ten causes, their rates are between 55.8 per 100,000 and 169.2 per 100,000. Given 70% of COVID hospitalizations in kids have happened in the past year, this would make it the #9 cause (rate = 58.8). Causes #4-6 are mood disorders, appendicitis and epilepsy which are hard to reduce. Discounting them would make COVID the #4 cause, and among all 10 causes, likely the most preventable. It may be an even larger burden given causes #1, #2, #3 and #8 are all complications or exacerbations of infections (pneumonia, bronchitis, asthma and dehydration) and thus do not represent a single disease.
Are these hospitalizations worth preventing? I think so. Hospitalization is a major life disruption for a parent and child that can be hard to recover from. And while rare, it can create long-term injury and disability. Here are two examples I came across which accord with what hospitalists have told me (interviews: 1, 2). A study in Hong Kong found COVID causes encephalitis at similar rates to influenza. Encephalitis is swelling of the brain, and while most recover it can cause lasting mental impairment. We also believe Omicron causes convulsions at a higher rate, with one case report series finding a complex seizure. Complex seizures can cause epilepsy. Notably, even the hospitalizations that occur in those vaccinated appear less severe.
Most hospitalizations are of course not severe, and around 61% of kids are admitted with fever or dehydration from diarrhea/vomiting (South Africa data). These kids are sometimes admitted on protocol and most are expected to recover fully. But do bear in mind that we recommend other vaccines like the rotavirus vaccine to prevent such complications. And as Harvard pediatrics professor Lakshmi Ganapathi notes: “even with other respi infections, kids do get hospitalized for complications (e.g. dehydration, febrile seizures, asthma exacerbation etc.).”
The more severe complications, however, highlight that hospitalization isn’t always benign. We know that COVID is a leading cause of death in kids (#8 or #9). The fact that this appears low as a rate does not mean reductions aren’t worthwhile. We as a society have steadily reduced most causes of pediatric hospitalization for years. Failing to curb a major spike when we have the tools available would be a clear divergence from public health trends.
I know Ladhani agrees that reducing death + disability is worthwhile. His argument is that this just isn’t an efficient way to do so. But here I also disagree.
Even without severe outcomes, infectious disease is burdensome. It is estimated that influenza costs the United States upwards of $10B a year, most of which is lost productivity. The average parent loses 6 hours of work when a kid is sick enough to see their doctor. ILI in a household with kids costs $400+ in lost productivity (school and work). None of this counts reduced productivity while working, which even for colds amounts to $16B a year (1, 2).
We also forget that people adapt to health threats. Policymakers commonly acknowledge that individuals make small but costly adjustments to lower risk which collectively amount to huge sums (we call this the value of statistical life). Data from my city at least suggests that parents are concerned about their kids' risk of COVID, they just don’t believe the vaccine is safe and/or effective. This same concern drove immense and costly behavior change in 2020 and may still be limiting parents. It is unique to the novel coronavirus as most people aren’t as scared of flu.
Fortunately, COVID vaccines reduce both risk and disease burden. A large study of symptoms in the UK found that one dose of mRNA vaccine significantly reduced the incidence of fever, nausea, myalgia & and other flu symptoms in 12-15s vs. Omicron (reduction was present but nonsignificant for 16-17s, likely significant with two doses). This is consistent with other studies which show an even greater benefit with two doses. A child is still likely to stay home with COVID, but is likely able to do schoolwork (esp. in the zoom era) and unlikely to need parental care or time-off. They are less likely to go to the doctor, which in the case of flu occurs 100x more commonly than hospitalization. They are also less likely to go to the ER, to which COVID sent 300K+ kids by the end of January. I thus believe the cost-benefit will favor vaccination, just as it does with influenza.
Vaccination of course has costs and can cause some illness. But the flipside of low efficacy in 5-11s has been generally mild reactions. And vaccination can be planned around, done at a pharmacy on the way home, etc.
If we extrapolate the cost of one vaccination to the whole population (73 million kids) we get a cost of about $2.9B. While sizeable, we spend $233B on kid’s health yearly with upper respiratory infections being the #7 largest cost. When one subtracts well-care, routine dental care, and other dental care, they are the #3 largest cost. The other two are ADHD and asthma which we can’t easily prevent with a cheap vaccine. Personal spending isn’t exactly analogous to public health spending but among other things we are willing to pay for, it’s up there. As a fraction of public health spending, vaccines for all children would be about 10% of yearly spending. We typically spend 4% of our yearly budget on communicable diseases, most of which impact a minority of people and many of which vanished during the pandemic.
None of this is even to speak of long-term benefits. Ashish Jha presciently argued in October 2021 that vaccination is an investment in a world where COVID is endemic. We don’t yet know how significant reinfections will be but Ladhani’s data show they appear to have a capacity to hospitalize (rates were similar between first and second infection, but that can’t be extrapolated as a case rate, and outcomes were less severe). Lab data in children suggests some mount a poor immune response, which is much more consistent in those vaccinated. This mirrors adult studies showing significantly elevated correlates of protection following a breakthrough infection as compared to Omicron infection alone. Given the likely cumulative effects of vaccination on reinfection, the benefits of diminished illness and fewer severe outcomes will grow in contrast to the front-loaded costs of vaccination.
I thus am not sure what other comparable intervention would be as cost-effective as COVID vaccination. Increasing rates of other vaccinations is a slog that comes in the form of modest improvements year-after-year. Conversely, some 20% of parents are willing to vaccinate 0-11s even before authorization. I might prioritize other vaccines like MMR where low, but in the US routine vaccination rates have not broadly dropped. There is thus lots of talk but little evidence that promoting this vaccine has hurt trust or uptake. As someone who advocates for immunization professionally, my fear is that those who have taken a hardline stance against vaccinating kids (e.g. Florida) are creating de novo hesitancy that could spread to other vaccines.
Because ultimately, I agree with Ladhani that this isn’t a massively consequential debate. I think childhood is relatively safe such that most interventions will yield modest benefits. Mainstream discussion of these nuanced efforts is bound to contribute more rhetoric than clarity and thereby exacerbate polarization. And for what?
Pediatric COVID vaccines can save lives, prevent severe disease, lessen symptoms for millions, are effective relative to other interventions, and may even pay for themselves. Knowing there is a moderate but positive effect on health we should agree and move forward, unless someone has a better way to spend the money. Kids fall needlessly through the cracks of our indecision.
Thanks to Edward Nirenberg and Anne Sosin for some brief feedback on this essay.