Oxford muddled a good vaccine with bad PR

A vaccine technician works in the Covid lab at the University of Oxford’s Jenner Institute. Picture: AFP

Statistics can be depressing. Back in the spring, the government asked vaccine experts what they thought the chances of success were. How likely was it that any vaccine in clinical trials would protect us against Covid-19? Their answer? 15 per cent.

Here’s another statistic that was, at the time, moderately depressing. As recently as November 2, the Tony Blair Institute was preparing us for the possibility that a vaccine may only stop 50 per cent of infections, if that. Right until the middle of last month, more pessimistic scientists were warning that we might never get a vaccine.

Now, a matter of weeks later, the Oxford-AstraZeneca vaccine has greatly exceeded all their expectations – in speed alone it’s one of the greatest scientific achievements of the century. The response? It’s being treated like it’s the recipient of a mediocre 2:2, compared with the starred firsts of Pfizer and Moderna.

This is, once again, because of statistics – or, in this case, because of the way Oxford communicated them and how we understood them. Somehow – and it’s difficult not to feel a twinge of wounded national pride – Britain’s vaccine, which is set to become the workhorse of Covid immunisation around the world, looks a little less shiny. And it needn’t have been this way: it’s still pretty shiny.

UK Prime Minister Boris Johnson with a vial of the AstraZeneca/Oxford University COVID-19 candidate vaccine on November 30. Picture: Getty Images

We have, in this pandemic, got used to statistics and sometimes quite complicated ones at that. There has been R, herd immunity and exponentials. By contrast, the latest statistic in which armchair virologists gained expertise seemed relatively simple. Pfizer’s vaccine was 95 per cent effective, Moderna’s was 94. This was, pleasingly, as easy to understand as an exam grade.

Then along came the Oxford vaccine and, with all the clarity of an All Souls exam, it was 62 per cent, 70 per cent and 90 per cent. How could we compare it with the others? It got more confusing. The overall 70 per cent efficacy rate was impossible because it was an average of two mutually exclusive clinical trials. Journalists asked repeatedly about the data in the group who got 90 per cent protection. How many cases led to that figure? We weren’t told. Statisticians were left to reverse-engineer the results to find out.

Clarity came only slowly. It emerged that this apparently winning group, which started with a half dose followed by a boost with a full one, had come about by chance, or, depending on your interpretation, by “error”. Later still, it emerged that the group didn’t include old people.

So what are we to make of it? Privately, many scientists were bemused; others were incandescent. Why not lay out all the data at once? Why not quote the 62 per cent figure and then flag the interesting 90 per cent result as just that – an interesting result? The vaccine would still be a triumph at 62 per cent efficacy.

In a pandemic, clarity matters. Antivaxers used the confusion to sow more confusion about a vaccine that, let us not forget, is safe and appears to stop all serious illness from Covid. And it’s not just the antivaxers this matters to. The US regulator, the Food and Drug Administration, looks likely to want more data amid signs of what might, diplomatically, be called “tensions” between it and AstraZeneca.

Anti-vaxxers protesting the pandemic in October. Picture: Damian Shaw

Here, though, is the counter argument. Yes, AstraZeneca could undoubtedly have communicated better, but its vaccine data sounds complex because it is complex.

The problem is that no one really expected these results. Back in May, a batch of the vaccine was produced that appeared to be too strong. It wasn’t, but to be safe and to keep the human trials going, it was given at half dose. The regulators were well aware of this, and they agreed with Oxford they would report all the data together in a single figure – the presumption being there wouldn’t be much difference between the two approaches. But, it seems pretty clear, there was.

As obliged, Oxford published the combined data – the unobtainable 70 per cent figure. They also published a breakdown of how it was arrived at. And now here we are. As Adrian Hill, head of Oxford’s Jenner vaccine institute, puts it, the findings are complex, but “we are happy with complexity”. Even if others aren’t. So what now? We await the regulators’ verdicts, and we await more data.

The truth is, if these really are the exam results of the vaccines, then the way we have used them so far is like judging an entire degree on a half glimpse of the first mark from the first paper. Many other tests are still to come.

AstraZeneca company and University of Oxford’s Covid vaccine. Picture: AFP

We don’t know how long immunity lasts, how good the jabs are at stopping transmission, or whether some might even work better if we mixed and matched doses between different types. We don’t actually know if these figures are correct on their own terms.

But we do know that this is a vaccine that works and is safe, and that we are still in a pandemic. In a normal year, the world makes 600 million doses of the flu vaccine. Next year it needs to make billions of a coronavirus vaccine, and it still won’t be enough. This is not a buyer’s market.

In 2022, when this vaccine has gone into a billion arms, it feels unlikely we are going to say, “Thanks for the vaccine, shame about the communication strategy.” In our obsession to mark vaccines like test papers, we (and perhaps some scientists too) have forgotten that the real competition is not between drug companies but between humanity and a virus.

Tom Whipple is Science Editor of The Times

The Times