The Effectiveness of Ineffective Vaccines

I have discussed Geert vanden Bossche's debatable theories before. Briefly, he argues that mass vaccination in a pandemic 'selects' for mutants that escape the immunity conferred by the vaccine, and this threat is reason enough to just stop. He ignores selection for escape of what is loosely, but erroneously, termed 'natural immunity', and the enormous health and social costs attaching to the sledgehammering of this nut. Unfortunately for his theories, all variants so far have arisen in poorly vaccinated settings. 

Now, Bret Weinstein - who claims to be an evolutionary biologist, yet - attempts, in his podcast and substack, 'On Driving SARS-CoV-2 Extinct', to rescue vanden Bossche from the dumpster and brush the banana skins from his shoulders. 

Here's an excerpt - 'Rounding the earth' is the work of one Mathew Crawford:

(I am aware that a properly-conducted review would include links to sources. In my defence, I can't bear to! Do your research.)

Now, this is pretty desperate stuff. Rather than recognise the possibility that there are other causes for increase of a variant besides vaccine, it is enough that a vaccine has been used, on however small a number, within a few hundred miles of the epicentre! 

Selection requires a differential in output. If one variant produces exactly the same number of secondary cases (on average) than another in the same environment - if they possess the same R value, in epidemiological terms - there is no opportunity for selection. It cannot increase relative to the other (by selection) without a differential it can affect: to get into more bodies it must have the greater R. 

The role of vaccine effectiveness

Now, this selective advantage can only happen even in principle if the efficacy of the vaccine against infection (VE(i)) is greater than zero. Efficacy in ideal conditions represents the differential between a vaccinated and an unvaccinated group of the same size and composition, where no-one knows which group they are in. In real world use, the term 'effectiveness' is used instead of 'efficacy'. In the real world, people know if they are vaccinated or not, affecting behaviour, and it also becomes more difficult to equalise groups with respect to other differentials such as age, comorbidity, socioeconomic factors and so on. Because of this - news though this will be to countless amateur internet data analysts - you cannot naively count cases in the two groups, and claim this accurately captures VE.

So, an accurately determined VE(i) gives the percentage differential in case count between treated and untreated groups of equivalent composition. 100% effectiveness means no cases in the vaccinated group, 0% means an identical (or a statistically insignificant differential in) case count. Intermediate values give the percentage reduction - eg 40% VE means 6 cases in the vaccinated for every 10 in unvaccinated: a 40% reduction. From this and the above, then, it is clear that VE 0 cannot generate vaccine escape - there is no possibility of getting into more bodies if the vaccine is completely ineffective; there is no selective pressure if it doesn't work.

But the relationship is not simply dichotomous. The strength of selection in the vaccinated, minimal when VE is zero and maximal when 100%, must therefore vary continuously between those extremes. The phrase "does not stop you catching it or passing it on", regurgitated monotonously online and whenever anyone sticks a microphone in front of a sceptic, means only that VE(i) is less than 100% - whereas the intent of most regurgitators is to suggest VE(i) is therefore zero, and there are exactly as many cases in a vaccinated as an unvaccinated group. 

It is true that VE(i) is steadily going down, due to waning, antigenic drift and the generalised immune escape of variants. Many even claim (based on shoddy VE-hacking by those amateur analysts referenced above) that VE(i) is negative for omicron.That being so would mean that omicron, at least, is not the variant vanden Bossche is looking for. 

For how can it select for immune escape if it confers no immunity?

We can extend this argument backwards. Each variant going forwards has successively diminished VE(i) - and, as the 'does not stop'-ers continually remind us, immunity wanes anyway. Therefore, Vanden Bosschian Selection must have been getting steadily weaker as a force, if it ever existed. But because Crawford has managed to locate a handful of vaccinees in the vicinity of each step, he gives them the same assumed selective force as 'wild-type'. I'd bet a pound to a peanut that Crawford and Weinstein have been pushing the implicit-0% "does not stop" narrative throughout 2021 (God forbid anyone should have a reason outside themselves to get jabbed!). And yet, a classic have-cake-eat-cake scenario, they also argue as if selective power were maximal throughout the series. No waning, no antigenic drift. 

It is interesting to see how doggedly people have been pursuing the absolutist "does-not-stop" line, and they seem almost relieved that omicron and waning have come along to retrospectively justify the falsehood they have been pushing throughout 2021. "It's going down" is a tacit admission that it was once up, no? But you'd have struggled to find a sceptic able to concede that in mid 2021. To concede it would kick one of the legs out from under their no-vax-'cos-I'm-OK stance. Yet the graphs they gleefully point to now show clearly that, when they were vigorously arguing against there being any protective effect whatsoever, there quite clearly was.

Sure it wanes. But you don't start off waned. Anyone commencing a course now will not have to worry about waning for a good while. Regarding waning as a justification for not getting jabbed is akin to regarding a motor vehicle's Certificate of Roadworthiness as unnecessary, since it will expire next year. On the other hand, anyone who had the disease in early 2020 may have little protection left; renewal may be appropriate. 

The role of frequency

Ignorance of diminishing VE is not even the main thing wrong with Crawford's analysis. If we assume, for argument's sake, that a variant can gain a copy specifically by infecting a vaccinated person, and these are rare, how is this supposed to translate into widespread transmission? The individuals in these trials were jabbed then went back to their communities, spread out; little pinpricks in the broader ground of unvaccinated individuals. Our variant gets an extra copy from infecting such a vaccinee, but it may have a long wait to find another. Remember that it only gets an advantage in vaccinees, in this example. The rest of the time, it wanders round the unvaccinated in direct competition for bodies with its ancestor, against whom it has no advantage. So this restrictive scenario isn't going to work. 

OK, you say, let's give it an advantage in the rest of the population too. Perhaps it can evade innate immunity, or evade the immunity of the previously infected. Indeed it might, and does. But now, we have abandoned vaccination as a unique cause of selective advantage altogether. If we create a selective advantage outside of vaccinees, that could easily be a prime driver itself - especially if there are far more such individuals. It doesn't speak well of Weinstein's grasp of evolutionary biology that he failed to notice this effect of the rarity of the advantageous circumstance on the strength of selection.

This is what we see in practice. The variants thrown up so far do not differentially 'prefer' vaccinees. They are simply hyper-infectious, and evade innate, infection-acquired and vaccine-induced immunity with not much discrimination. Such discrimination as does exist appears to give greater advantage, if anything, to escape of infection-acquired immunity. This is the opposite of what is needed; people can't argue that 'natural immunity generates fewer cases' at the same time as arguing 'vaccines select more strongly'.

The role of mutation

Mutations do not necessarily arise where they find their advantage. They are a function of number of replications - the more infections there are, the more opportunity for mutation. Once arisen, mutations are metaphorically chucked against the wall to see if any stick. Outside of hosts that provide an advantage (in terms of additional bodies entered), they can only drift. Once they hit the right kind of host, they get a boost, and if there are a lot of such hosts, they get a lot of such boosts. This can be enough to drive out and replace the ancestral type, where 'the right kind of host' is common. But consider: if there were a way of reducing the number of replications, that would reduce the number of mutations. Do we have such a way? Well, yes. Vaccination. Sure it's not perfect. Sure it wanes. But it will help limit the problem. If people who spent last year arguing against the protective effect had instead embraced it, we might be in a better position now Even 20% effectiveness against infection is protective - 8 cases per 10 unvaccinated, instead of 10. That's two individuals not passing it on who otherwise would have. And because, unlike additive parameters like hospitalisation and death, the effect compounds, the benefit goes beyond one generation. With an R of 2, after 1 interval we have 16 instead of 20. After 2 we have 32 vs 40. After 3, 64 vs 80. And of course, the presence of other vaccinees in the population reduces that R anyway. Every little helps.