Thankfully, there seems to have been lots of smart discussion about speeding up the pace of the UK vaccine roll-out. There’s been the debate over prioritising first doses; efforts to remove bureaucracy; ideas to mobilise as many people and locations as possible; and thoughts on maximising supply. In contrast, I have seen almost no discussion about another idea: prioritising vaccination of those who haven’t already had coronavirus. So I wanted to share a few quick thoughts. To be clear, I have no expertise in this area (and am writing in haste), but in the absence of any apparent UK discussion I hope the below is a prompt for others to do some deeper thinking about this policy option.
First, the basic concept: if coronavirus infection typically confers some immunity, could we delay vaccination for those who have already been infected and thus – given a scarcity of vaccines/appointments – prioritise those who would benefit more?
In England, around 9% of the 16+ population – around 4 million people – “would have tested positive for antibodies to SARS-CoV-2” in November – and that number will have risen since. Plus, the number of people with detectable antibodies is, as I understand it, also only a lower bound for the number of people who have been infected and who might have immunity. So a significant (and still growing) chunk of the population – maybe even a fifth (if antibody tests only capture half) – has already been infected. It would be wrong to assume those people are all now totally immune from COVID-19, but I believe it is fair to say they are – all else equal – now at a lower risk than others of being struck down by it, at lower risk of transmitting coronavirus (probably), and presumably at lower risk of being host to the development of new mutations. To derive the most impact from vaccination, as soon as possible, then, deprioritising those who’ve already had coronavirus seems attractive.
A recent (preprint) academic paper modelled this strategy and found that “Preferentially vaccinating seronegative individuals yielded large additional reductions in cumulative incidence and mortality in locations with higher seroprevalence and modest reductions in locations with low sero-prevalence.” In the chart below, you can see (in the purple lines) that: 1) prioritising the vaccination of the 60+ population is a good strategy; but also 2) on top of that strategy, prioritising vaccination of seronegative individuals means a greater reduction in deaths and years of life lost (YLL).
What would that this strategy mean in practice? One option would be to require everyone to have an antibody test before receiving vaccination. That might be appropriate if we had precious few vaccines to distribute. However, it is likely that vaccine supply will soon exceed antibody testing capacity (apparently 120,000 a day), and in any case antibody testing is far from perfect. We could make use of unused antibody testing capacity for some-but-not-all vaccinations, without allowing that capacity to constrain the number of vaccinations, but asking people to have their blood tested simply so that they could have their potential vaccination delayed is a bit of an ask.
More straightforwardly (though here I’m departing from the academic paper above, which focuses on antibody testing), why not make use of the records that show that 2.8 million people in the UK (and rapidly rising) have tested positive for coronavirus? Could those people be deprioritised for vaccination?
To put some rough numbers on the potential impact of this:
- Let’s say the Government’s goal is to ultimately vaccinate 50 million people.
- Let’s say they manage to average 2 million first doses per week (optimistically assuming that second doses could be additional to that)
- This suggests a full time-frame of 25 weeks for first doses
- If we can identify 3 million people who’ve already had coronavirus (which will soon be a conservative number, even excluding children and those who’ve died), and we (crudely) assume those people are immune, moving them to the back of the queue would mean that the point at which all 50 million have some immunity would be moved forward by 1.5 weeks. That would be 6% of the total time required, and perhaps an even greater proportional improvement in the time taken to reach herd immunity.
That’s a back-of-the-envelope calculation, and there are a lot of ‘ifs’ here, but if we could bring forward a given (low) death rate and/or an end to some restrictions by even a few days, by simply merging a coronavirus test dataset with a vaccination list dataset, that really seems worth exploring. Even from a narrow fiscal perspective, the furlough scheme for example is costing HM Treasury at least hundreds of millions of pounds each week.
I think there are at least four potential counterarguments to this idea.
Is it technically feasible?
Specifically, how easy would it be to link the UK’s coronavirus test records with the records of who needs to be vaccinated and use that information when scheduling vaccinations? Is it naive to assume there are two straightforward national datasets out there that include NHS numbers and/or names and addresses? (Probably.) Do let me know if you know the answer, but it would clearly be sad if (more) people died as a result of poor data structures.
For how long does infection confer immunity?
The case for deprioritising those who’ve been infected would clearly be weakened if past infection did not confer much immunity and/or prevent transmission in future. But my impression is that it’s reasonable to think that people infected in October/November/December/January (i.e. the majority of the recorded cases) will have some natural protection for the next few months, at least.
Could this send misleading messages?
It is possible that a policy of moving previously-infected people to the back of the queue may A) discourage people from being tested, if doing so could delay their eventual vaccination; and B) send a message to many people that they are no longer at risk, which might promote virus-spreading behaviour. Those are not great outcomes, but they would have to be large effects to outweigh the plausible benefits of prioritisation.
Is this a distraction from vaccinating as many people as possible?
The priority should – of course – be maximising the supply and use of vaccines; doing every possible to ramp up to perhaps even 5 million vaccinations per week in the UK. But there’s no reason we couldn’t also consider some additional prioritisation. And vaccination could easily take (even) longer than we currently expect, particularly if current vaccines prove ineffective against some new variant. So we should use every tool available to speed things up.
To stress again, I am not an expert in any element of this discussion, but it does seem like this policy option deserves more debate – with potentially significant benefit for relatively little effort. If you have any thoughts, or think I’ve missed any discussions, let me know in the comments or on Twitter.