Published 21 October 2020, @georgebatchelor
Safely opening up the skies is a fundamental pillar of the nation’s economic recovery. Yet worryingly, the UK is being held back from reducing the 14-day quarantine and introducing a testing regime because of an over reliance on misleading modelling.
That is the conclusion of our analysis published today with Oxera on the key evidence behind the Government's reluctancy to push forward with a testing regime for air travellers. The claim from modelling by Public Health England that arrival testing would "only identify 7% of infected travellers", which underpins the inaction so far, is distorting.
The reason for such a big under-estimation is simple. The modelling used by PHE was conducted back in February and made a number of substantial assumptions that have not been changed or challenged. They assumed that all symptomatic travellers, and all asymptomatic travellers who could be detectable if given a PCR test before the flight takes off, would not fly. In reality they would board the flight and then be detected by arrival testing.
For any uninfected air traveller that has to self-quarantine inside without outdoor exercise for 14 days, this is deeply frustrating. But testing on arrival is not just about reducing quarantine time. With evidence emerging that as few as 20% of travellers adhere to quarantine requirements, test on arrival could be more effective than the 14-day quarantine rule at reducing the risk of Covid-19 spreading even more widely in the community.
There is, of course, a need to try and control the spread of the virus and to do this the Government has had to make quick and complex decisions under ever-increasing pressure. Models have been relied upon for many of these decisions - essential given the uncharted territories of Covid-19. But time and again, models and algorithms have led to poor decisions over the last few months. From exam results to missing Covid results and now airport testing, models and ‘mutant’ algorithms and their interpretations are interfering with common sense and real-world evidence.
The problem is the lack of humanity in how models and algorithms are developed, considered, and used. Too often models and algorithms are developed behind closed doors and become impervious to dissenting voices. Surely someone must have thought recalibrating grades was a bad idea, particularly given the bias against individual schools and stress on young people already caused by the crisis? And surely someone from one of the three central testing labs must have wondered why positive test cases had stopped increasing nationally in early October, despite what they were seeing locally? And surely many people think, as we do, that the 7% behind the reluctance to introduce airport testing is unrealistically low?
Models can be tweaked and improved, but humanity and real-world evidence are the critical missing factors.
Data can be elusive amid a pandemic. This is why the departure testing trial at Heathrow which started this week is so essential. These data, along with data from other countries like Jersey, will provide critical information on the actual effectiveness of airport testing. Still, we can safely say that it will be significantly higher than 7%.