Where is the evidence that early pub closures will lower infection rates?

In imposing new rules and restrictions on public behaviour to try to stem the accelerating spread of coronavirus, is the British government still “following the science”? Some of the latest measures—which prime minister Boris Johnson has warned may be in place for up to six months—are puzzling, and many scientists, including some linked to the Scientific Advisory Group for Emergencies (Sage), have been critical of them. One of the most perplexing is the closure of pubs and bars at 10pm. Where did this come from?

It’s easy enough to understand the thinking. As the evening wears on, people drink more and lose their inhibitions, and social distancing goes out the window. What’s more, the hubbub rises and people talk more loudly, expelling more droplets that could be laden with virus. Best, then, to curtail proceedings before they reach that point, right?

Yet the reasoning falls apart quickly. Every student (at least) knows that if you’re going out for some serious alcohol-fueled socialising within a limited time frame, you just drink faster. I know of no evidence that pubs and bars get any louder after 10pm. And this relatively early closing increases the chance that everyone leaves at the same time, cramming onto buses and trains.

So it’s not at all obvious whether the early closures will have any benefit, nor that they won’t in fact make matters worse. Yet in response to a question in parliament, Johnson said: “What we’ve seen from the evidence is that, alas, the spread of the disease does tend to happen later at night after more alcohol has been consumed.”

That’s a claim of a different order from “We’re doing our best and we think this might help.” It’s an assertion that the pub curfew is science-led and evidence-based. Does such evidence exist? If so, behavioural scientists don’t seem to know of it.

“I am not aware of any evidence of the kind that Johnson claims exists,” says health psychologist Robert West of University College London, a member of the SPI-B group that advises Sage on behavioural science. And he emphasises that the new rule, or the motivation for it, “have not come before [SPI-B], which is something you would have thought would be important if science is going to have any say in the matter.”

Some researchers acknowledge that alcohol consumption increases risk-taking behaviour. But whether this gets significantly worse after 10pm, and whether it does actually increase viral transmission in pubs and bars after that time, seem to remain matters of speculation.

Susan Michie, also a professor of health psychology at UCL, says that the pub curfew may not make much difference to alcohol intake anyway. She draws an analogy with smoking: “there is evidence that when the nicotine in cigarettes is reduced, people inhale more deeply,” she explains. She suspects that bringing the traditional closing time of 11pm forward an hour “may result in people starting drinking earlier, drinking more rapidly and/or going somewhere else to carry on drinking at 10pm.”

West agrees: “the 10pm curfew may well push the drinking to earlier in the evening,” he says. “There is evidence, for example, that extending opening hours [as the UK government did in 1988, for instance] did not result in an increase in alcohol consumption in bars, but rather, spread it out more.” Another member of SPI-B, behavioural scientist Stephen Reicher of the University of St Andrews, shares these concerns, adding that if people just drink faster to cram it in before closing time, the curfew might even make matters worse by increasing inebriation. At any rate, he says, “I know of no empirical studies on this.”

The 10pm curfew looks, then, worryingly like a cosmetic attempt to be seen to be “taking action” while in fact making a political calculation. Johnson’s unwillingness to court public disapproval by appearing to be a killjoy—the exact opposite of the ebullient, libertarian persona that got him into power—probably played a part in landing us in this predicament in the first place. Notoriously, his inordinate concern with “the inalienable right of free-born people of the United Kingdom to go to the pub” made him drag his feet over the lockdown in March, a delay that is now known to have cost thousands of lives. It’s not hard to see how Johnson might think a “pint in hand” image would chime with the demographic that supported his election on a ticket of “getting Brexit done,” and the government reinforced that image emphatically in post-lockdown photo opportunities.

Michael Head, an expert in global health at the University of Southampton, has said that a far better approach “would be to close all pubs and restaurants and properly compensate businesses and employees for the loss of income.” But that, of course, would place financial obligations on the government that can be evaded if the bars stay open, even though their revenues (and staff earnings) are likely to take a hit from the curfew.

The Independent Sage group formed by former government chief scientific adviser David King has recommended instead that pubs and bars be restricted to serving drinks outside—a solution that of course will not be available to all of them, and which would severely limit clientele numbers and become increasingly problematic as the autumn progresses.

There’s some suggestion that the pub curfew might be inspired by measures introduced in Belgium, where bars and restaurants in Antwerp were ordered to close at 11pm (rather than continuing into the small hours) as part of a general nighttime curfew, one of a suite of measures to curb an outbreak of infection in the late summer. That, however, in itself offers no support for the prime minister’s claim about “evidence.” The measures helped, but one can’t single out the effects of bar closures (other late-operating businesses were closed too) or changes in alcohol consumption—and the 11pm closure was simply a measure to give people time to get home before the blanket overnight curfew of 11.30pm-6am began. In a city like Antwerp with a bustling nightlife, you’d expect to see changes in the virus spreading if you cleared the streets of people entirely and closed everything down for any substantial period of day or night.

No one can expect behavioural measures—short of a complete lockdown—to have the internal logic and consistency of a theorem in mathematics. There must inevitably be rather arbitrary thresholds—the “rule of six,” the limit of 15 guests at a wedding and 30 at a funeral. No precise calculus delivers these figures; they can only be estimates that seek a compromise between safety and efficacy and the wish to permit some degree of social contact. There is nothing to be lost by admitting as much. But if necessarily speculative or uncertain measures are introduced with a false veneer of scientific credibility, that will seriously undermine public trust.

This is now the real issue. Whether the new rules will be effective, says Michie, “depends on how much people agree with the guidance—that is, whether they believe that it will have the desired effects and that the desired effects are important. Sadly, a large majority of the population don’t have confidence in the government’s handling of the pandemic and are likely to have been further alienated by being blamed for transmission rates rising [and] testing shortages. In this context, I fear that the buy-in of the public to this restriction may not be high.”

The biggest millstone around Johnson’s neck is painfully obvious. Back in July, Reicher told me presciently that the consequences of Dominic Cummings’s breach of lockdown rules (and apparently common sense) in his “eye-test” trip to Barnard Castle would be felt later, when the government encountered again a need for compliance rather than when, as at that point, it was relaxing lockdown. Now, sure enough, Johnson has hardly finished saying things like “if people don’t follow the rules we have set out, then we must reserve the right to go further” before everyone responds “as you did with Cummings, then?”

Compliance is indeed critical at this highly dangerous point in the pandemic—but the government has squandered all authority in calling for it. “The basis of trust and compliance,” says Reicher, “is to show you are on the side of your audience. If you want the public to abide by the rules, to act for the common good, and to take their responsibilities seriously, then don’t lecture; set an example.”

Social psychologist John Drury of the University of Sussex, another behavioural adviser to Sage, agrees that the problem is not late-night drinking—he is equally puzzled where the 10pm curfew came from—but a lack of trust and clarity in the government’s strategy. “It was a pity” that its messaging “laid so much blame on the public for ‘not sticking to the rules’ and so on, and is resorting to threats of punishment,” he says. “The main reasons for lack of adherence to distancing, I would suggest, are a decline in trust of the government, decline in sense of unity, and decline in understanding/knowledge of the rules.” The answer is “good engagement”—and “threats of coercion do not fit well with this, as research on other public health emergencies shows.”

Whether or not the pub curfew is indeed backed by “evidence” therefore seems a question with deeper ramifications. If indeed such evidence exists, it seems likely to be weak, circumstantial or cherry-picked. That would be bad enough for rebuilding any trust.

But if it does not even exist at all, Johnson’s claims would appear to be made either without an understanding of what “evidence” entails or—worse still—in full Trumpian mode: a declaration of convenience that simply does not care about the science or the truth. If the government has entered that phase, we’re heading for real trouble.