When the UK became the European country with the highest number of COVID-19 deaths earlier this month, there was renewed criticism of how it had handled the crisis. A common complaint was that it had entered lockdown too late.
The UK closed down all non-essential businesses and severely restricted public movement on March 23, nearly 10 days after the World Health Organization had declared the coronavirus a pandemic. This was two full weeks after Italy – then the worst affected country in the world – had imposed its own lockdown.
Sir Patrick Vallance, chief scientific adviser to the UK government, said that this delay was necessary because people would get “fed up” of following rules. Delaying the start of the lockdown, the theory went, would make sure the public wouldn’t run out of patience with the restrictions when the outbreak was at its worst.
The idea that the public would be susceptible to this “behavioural fatigue” prompted critique from some scientists as well as support from others. Was the government correct to think that adherence would fall over time?
Traffic data and location information from people’s phones suggest that compliance to the lockdown did wane, as predicted. Adherence to public health measures has decreased over time in past pandemics too. But this isn’t because of fatigue.
Rather, the adoption of health-protective behaviours depends on our beliefs about the risks we run when we don’t comply. For people to comply, they need to believe that the risk of not doing so is high – especially with measures that demand a high degree of effort.
So far, so good. But there is a problem here. It’s “optimism bias”: the idea that we predict the probability of us facing negative life events (such as getting cancer) as being much lower than the probability of others facing the same event.
Such thinking is seen in many different situations, and researchers have documented the phenomenon during the current coronavirus crisis. In a survey conducted in four European countries – France, Italy, the UK and Switzerland – in late February 2020 (just around the time of the Italian lockdown), researchers asked participants to estimate the probability of themselves, and the general population, getting COVID-19 in the next few months. Just over 30% of the sample thought they had a 0% chance of catching the virus, but only 6.5% reported a 0% chance of others catching it.
In general, optimism bias is quite useful, yielding better life outcomes in certain situations. People with higher levels of optimism work harder, save more and are more likely to remarry after divorce. But it is problematic for compliance to guidelines over time. This is because our optimism causes good news to change our beliefs more rapidly than bad news. Essentially, this means that we are predisposed to thinking the virus will not affect us, and the more successful a virus containment strategy is, the more likely we are to believe that we are immune.
Focus on risk
Once we understand that compliance drops not due to fatigue, but because of a lowering of perceived risk, it’s clear that any strategy should have focused on how to maintain high perceptions of risk.
The UK government also needed to think about trust, because trust in authorities influences how risk is perceived. This in turn can lead to an effect on compliance with health measures. For example, a study from the 2009 swine flu pandemic found that having trust in authorities influenced people into complying with control measures such as quarantine and avoidance of crowds.
The authorities therefore should have done all they could to maintain high levels of trust. One key area they could have focused on is consistency. It’s been shown theoretically that inconsistent information reduces trust levels over time, with people ultimately ignoring inconsistent information altogether. In practice this happened in Toronto during its SARS outbreak in 2003. Inconsistent information from the Canadian authorities affected people’s compliance with quarantine measures.
Overall, the UK government was right in thinking that lockdown compliance would reduce over time. But the critical error it made was thinking that this would be because of people getting tired of the rules. This prompted the government to delay the lockdown, most likely further lowering the perceived risk and so making it less likely that people would stick with the guidelines once they were implemented, as well as yielding a further erosion of trust.
Perhaps this decision was taken due to trust levels already being low. According to the World Governance Indicators, perceptions of UK government effectiveness have been declining since 2015, and since 2017 have been at the lowest levels ever since the first year of reporting – 1996. But whatever the reason, it seems that an incomplete understanding of human behaviour has informed the UK’s pandemic response.