The birth of a new category of travel restrictions

On July 8, the UK announced that from July 19, people who were fully vaccinated could come into the UK from amber list countries as if they were coming from green list countries. That means skipping the requirement to home quarantine and avoiding the need to take a second test at day 8. The vaccinations had to have been carried out in the UK, at least initially, as the verification method was the NHS app. The exemption is promised to be extended to vaccinations received in other countries, as soon as practical issues around verifying documentation get solved. All good news for vaccinated travellers and for the travel industry.

However, on July 16, only 8 days later, Grant Shapps dropped another bombshell. He announced that the exemption would not come into effect as planned in the case of France, due to concerns about the Beta variant. That’s the one originally discovered in South Africa, which studies have shown to be one of the more vaccine resistant variants. A brand-new risk category “amber plus” was created as a container for the new rules, which are the same rules as used to apply to the amber list until July 19.

What triggered the change?

Apparently, the change was triggered by what the government thought was evidence of rapidly rising cases of the Beta variant in France. There is speculation in the media that the report that experts were looking at included France’s overseas territories, notably Reunion which has been having an outbreak of the Beta variant. That wouldn’t make much sense in terms of deciding policy for travel from mainland France, as Reunion is 5,800 miles away in the Indian Ocean and France has its own restrictions in place for travel to France from there.

Here are the figures for Beta variant submissions to GISAID, by day of submission. I’ve shown both the mainland France figures and the Reunion numbers. Could the French figures alone have triggered such a reaction by themselves? It seems unlikely, so maybe there is something to the media speculation.

Even if the Reunion figures were included, it still seems like a very “trigger happy” response. But after the level of criticism levelled at the government about being slow to respond to the Delta variant, you can see why they would have wanted to err on the side of responding quickly to any signals which were even potentially worrying.

The reaction to the increase in Beta variant cases seems even harder to understand if I show you the same graph as before, but “zoomed out” to give you the benefit of a longer data history (below). You can see that I chose my start date for the previous chart very carefully, didn’t I? But that doesn’t mean that a chart just like the last one wasn’t what prompted the policy change. Perhaps prepared by people who felt the relaxation of rules for fully immunised people was too risky? The “spreadsheet error” which included overseas territories might not have been an error either. We’ll probably never know.

Should we be worried about Beta cases in mainland France?

Let us rewind a bit and examine the question about whether the Beta variant in France should in fact be a concern for the UK.

The Beta variant has always been the one that seemed to show the most vaccine resistance. I’ll come back to the evidence on that in a moment. So far, it hasn’t really taken off anywhere in the world, generally being outcompeted by the more infectious Alpha and Delta variants.

But the worry is that we haven’t seen what happens in highly vaccinated populations. Maybe the superior ability of the Beta variant to “escape the vaccines” will start to compensate for the greater infectiousness of Alpha and Delta?

Whilst France is behind the UK in terms of vaccinations, it has been catching up rapidly with 56% of the population having received at least one dose. At this point, it is a country where vaccine resistant variants might be developing a selection advantage. France also has the Delta variant now. So, I can see why policy makers would be hypersensitive to any signs that Beta might be showing signs of outcompeting Delta in France.

In the chart below, I’ve shown the GISAID data on the Beta variant on a timeline based on the date the sample was taken, rather than the date it was submitted to GISAID. You can see that Beta prevalence has been declining steadily in recent weeks, from a high point of over 9% of cases to below 4% in the latest data.

*Source: GISAID, outbreak.info, GridPoint analysis*

The reason for this is the growth of the Delta variant, which is rapidly becoming the dominant strain in France, as it has in the UK. It certainly doesn’t look like Beta is outcompeting Delta in France based on this data.

How resistant is Beta to the vaccines anyway?

There have been a lot of scare stories about Beta, in particular its ability to “evade” the Astra Zeneca vaccine. These were triggered by South Africa’s decision to halt the roll out of Astra Zeneca in the light of apparently poor effectiveness figures, in a population where Beta was the predominant variant. I don’t have space to go into all the issues with the South Africa effectiveness results, but very small sample sizes and a very short duration were just two of them.

We now have much better data from Canada. On July 3, Canada published the results of a much bigger study based on real world experience in the country. The paper was updated on July 16, the very day that Grant Shapps was announcing the new amber plus list. Canada has all the major variants and uses all the “big three” Western vaccines. The results are very reassuring about Beta.

Firstly, the effectiveness estimates against symptomatic infection after one dose show great results for Pfizer and Moderna. There are confidence intervals around all these figures of course, so don’t pay too much attention to a few percentage points either way, but effectiveness of these two against Beta seems undiminished. If anything, the protection is higher than against Delta.

It is only the Astra Zeneca vaccine that looks to have a somewhat reduced effectiveness against symptomatic infection after one dose. But the results are far better than those claimed in the much smaller and very short duration South African study, approximately halving the risk of symptomatic infection by Beta.

The figures for protection against hospitalisation and death are even better. All three of the vaccines provide very high levels of protection (around 80% or better), and the protection appears to be undiminished in the face of the Beta variant.

We don’t yet have results from Canada for Astra Zeneca or Moderna after two doses. But the results for Pfizer are very reassuring, with 85% protection against symptomatic infection and 98% protection against hospitalisation or death against the Beta variant.

Is Beta really a threat to the UK?

What do these results mean for the threat posed to the UK by the Beta variant, were it to get established here? The protection against hospitalisation and death provided by any of the vaccines remains just as solid as it is against the Delta variant. The only question mark is whether the virus would create more mild and asymptomatic cases amongst those vaccinated with Astra Zeneca.

We know there is a diminished protection against infection after only one dose but following the decision not to use it for under 40-year-olds, almost everyone who had AZ will have had two doses by now. Unfortunately, we don’t have data for two-dose protection for AZ, but even if two doses offers reduced protection against infection for the Beta variant, there will be an offset because the Beta variant is less infectious than Delta.

The original strain in China had an R0 of about 2.6. D614G, the variant that nobody has heard about, was the one that caused the first wave in Europe and had an R0 of about 3. The Alpha variant has an R0 of around 4.5 and Beta is similar, whilst Delta’s R0 is around 7, 60% higher than Beta. That’s why Delta is taking over everywhere.

As long as the vaccines provide protection against infection by Beta that is at least 60% as good as they do against Delta, Beta will not spread any faster, even in a fully vaccinated population. Astra Zeneca’s protection after one dose is 67% as good as it is against Delta, so even AZ should clear that hurdle. And at least half those vaccinated in the UK population got Pfizer or Moderna, which appear to be better against Beta than they are against Delta.

So, I can’t see how Beta would ever be able to get a sustainable foothold in the UK. It is outcompeted by Delta, with or without vaccines.

The Spanish question

There has been much speculation about whether Spain might also get added to the Amber Plus list, since about 8% of its cases are now Beta variant. Let’s test our theory about whether Beta can get established where Delta is also prevalent by looking at the geographical distribution of the variants. Firstly, let’s start with the more infectious Delta. Data from early July showed that it had already become the dominant strain in the Balearic Islands, in the Valencia region and a strip north of there. It is rapidly taking over in Madrid, Galicia and the Canaries too.

Source: Centro de Coordinación de Alertas y Emergencias Sanitarias, Spain. GridPoint analysis. — *Source: Centro de Coordinación de Alertas y Emergencias Sanitarias, Spain. GridPoint analysis.*

For the same period, the pattern of places where Beta was present was very different. The following chart show the geographical distribution of Beta cases. Note the scale differences between this chart and the previous one. The regions with the highest proportions of Delta reached over 85% of all cases. For Beta, the highest is under 10%. Comparing the two maps, you can also see that Beta is only really a thing where Delta hasn’t yet arrived. Just as we would expect, given the higher infectiousness of Delta.

It is also worth noting that there are no cases in the Balearic Islands or the big tourist destinations in the South. There is some in the Canaries, but fewer than 4% of cases.

What should the government do?

Whatever it was that triggered the government’s decision to create the amber plus list and put France on it, the stated reason of rising Beta cases doesn’t hold up to further inspection. It would of course be embarrassing to acknowledge that they got it wrong because of a “spreadsheet error”, but they could easily point to further reassuring data on declining Beta cases in France and to the newly published evidence from Canada about vaccine effectiveness against Beta as the reason for the change.

What they certainly shouldn’t do in my opinion is to “double down” on the amber list by adding Spain and Greece to it, as some have been speculating that they might.

The government is promising “no further changes” until August 2. By that time, more than two weeks will have passed since the amber plus list got announced. Hopefully, that will be long enough for the government to quietly bury it and put France back to amber.

The whys and wherefores of the UK’s new amber plus list