Right now, everyone on the news in the UK is talking about testing. The health minister has just said he wants to get the number of tests per day up to 100,000 and the Prime Minister and health officials have said that testing is the key to unlocking the virus. The world health organisation has said "test, test, test". People have described an antibody test, which will say if you have had the virus and are immune, as a game changer.
But why? At first, it seems like we get much more information if we do lots of testing. What we certainly get is lots more data. But do we actually get more useful information? I'm not sure that there is much that mass testing would give us that would actually change policy decisions or save huge numbers of people getting infected.
Firstly, though, if you are an NHS worker, then before you slam your phone down in anger at me - I am not talking about targeted testing. I strongly believe that sick patients, NHS worker and social carers should be tested. It is obvious that sick patients with covid-19 like symptoms need testing. And if NHS staff and social carers are self isolating when they actually don't have the disease, then this is a severe problem that is reducing health and social capacity and leading to worse outcomes for sick people. All these people should be getting tested regularly and routinely.
What I'm talking about here is mass testing for the public. Something like a drive through centre where you can turn up and get tested to either see if you have covid-19 or if you have had it previously and are now immune.
Antibody test
I'm going to talk first about the "game changing" test to check if you have had the disease and are now immune - the so-called antibody test. Let's say we test could test 700,000 people over the course of a week or two. If we did that this starting now(when the UK has approximately 38,000 confirmed cases, ~0.06% of the population). Then we'd probably find that between 0.5% and 2.5% of the population has had the disease (10-50 times the confirmed cases). It seems incredibly unlikely that the numbers would be higher than this. So let's explore what we would do differently if the number came out at 0.5% vs. 2.5%, different by a factor of 5!
Maybe you would say that we could model the virus better, if we knew the number more accurately. Two numbers that feed into the computer simulations are how fast it spreads and what percentage of people end up critically ill. At the moment the scientists have estimated these numbers to fit the data they have, primarily the number of people ending up critically ill. If we had these numbers wrong by a factor of 5, you would think it makes a big difference. But the actual difference in the simulation, would be that scientists change the estimated rate of spread in one direction and change the estimated percentage of people who get critically ill by a factor of 5 in the opposite direction. The result would be basically no change in the number of people we expect to get critically ill at any given time! Our simulations don't really need this number. You can try out the simulator on this page to see this for yourself. This means that regardless of this number, the policy decisions will stay the same.
What about checking how strong we need to lockdown? Well this is basically the same as for the simulations. We know how strong our lockdown needs to be, because we are doing one and the hospitals are just about coping. The only thing that really matters for lockdown strength is the number of people that get critically ill. We can see that, from the information we already have, so we are gaining nothing here. Again, the policy decisions stay the same.
Perhaps this test will tell us how close we are to herd immunity. Unfortunately, the answer to that is not very close. We need around 60% of the population to be immune for us to have herd immunity. Whether we have 0.5% immunity or 2.5% immunity or even whether we have 5% or 10% is irrelevant. The next wave will still grow exponentially, and it will spread at 99.5 %, 97.5%, 95% or 90% the speed that the first wave spread, almost no noticeable difference at all. We can tell this, by the fact that our cases are currently still growing exponentially, meaning that the infection rate just before lockdown had not really dropped at all due to herd immunity. You've got it, the policy decisions stay the same.
Perhaps testing will tell us how many lockdowns we will need - how far are we along the path before we get back to normal. As mentioned before, we need to get to 60% immune for herd immunity. If each wave of infection and lockdown results in 0.5% immunity, it will take 120 waves. If each results in 2.5% immunity then it will take 24 waves. Although this seems like a big difference, is it really that important in the end? Both are big numbers which will require lockdowns to continue for years. It seems unlikely that it is possible for a country to maintain the life changing impacts of repeated lockdowns for so long. So whether it is two years or 10 years, both these numbers are probably so large that, yet again, the important policy decisions will be the same.
Maybe it would mean some people who are immune could go back to normal life. Perhaps this is true. But I'm not sure how politically palatable it would be. It's not clear that people would be as happy to stay in lockdown, if they knew their friends and colleagues could go about their business as normal. Especially if it was just a privileged few who could go out as normal. Even worse it could lead to things like "covid parties" where people deliberately try and catch it from infected friends just so they can get back to normal. It seems unlikely that something like this would be backed by the government.
The last item I can think of is the one that I think may make a difference. Knowing what fraction of people have had the virus will tell us something about the actual mortality rate and the percentage of people that become critical. This will inform us how many people may be affected if we let the virus run it's course and allow the health service to be overwhelmed. Basically the only thing this kind of testing does, is tell us how many people will die if we stop the lockdowns and do nothing. Don't get me wrong though, that may be an important thing to know. If chances of a vaccine turn out to be slim or a long way off, then choices have to be made about treating covid-19 vs. other patients. Currently many cancer treatments are being postponed and 1.4 million people in the UK look likely to lose their jobs. People are having reduced quality of life - you may know a person who, even if they don't catch covid-19, is unlikely to live more than another year - they are looking at spending their final summer in lockdown and isolation. Lockdown cannot go on indefinitely.
The currently infected test
Maybe it is more useful to do lots of testing to check if people have the disease right now. Again, I'm not so sure. We are way past the situation of tracking individual cases and working out who they could have infected. The disease is out and we are not going to stop that. So, it's not clear what we really get from knowing we have it.
One strategy may be to test people with covid-19-like symptoms. However, no test is 100% correct. The latest fast test is called Samba II and claims catch 98.7% of cases. If you have the symptoms, but get a negative result from this test, would you go back to work and go visit your elderly relatives? Or would you think, there's more than a 1% chance the test was wrong, I better stay at home just in case? Your employer may even say that, irrespective of test results, they do not want you to come to work anyway, just to be safe. You may actually have flu or another virus and not be well enough to go to work regardless. Of course if you tested positive then you would still need to isolate, so you have gained nothing.
If instead, you just tested random people, then perhaps you would catch people who are asymptomatic. But would you catch enough to make it worthwhile. If you tested 100,000 people in a day you would only catch 0.15% of the asymptomatic people in the population. In the time it takes an asymptomatic person to recover (around 14 days), you would catch only around 2% of them, so would have close to zero impact. Even worse, again because no test is perfect, there will be some false positives. This is where someone who does not have the disease incorrectly tests positive. There seem to be no published figures on the number of false positives for the Samba II test, but a test would need to be better than 99% accurate in this sense - otherwise it would cause us to quarantine as many healthy people as infected people! A chinese study was withdrawn by the publisher after it found that around half to three quarters of people who tested positive, probably did not have the disease. So if we do mass testing many people may be unnecessarily quarantined.
There are some real disadvantages to increasing testing too. We will be taking NHS resources away from where it may be needed - i.e. dealing with sick patients. If people get a negative test, they may get infected the next day, in fact they may get a sense of invincibility and might actually take more risks.
Also, do you remember your science teacher at school telling you that if you want to repeat an experiment, you must keep all the things the same or you will get a different result? If we change the testing strategy significantly, then it actually makes it harder to interpret the data. If next week we do ten times as much testing and we see that cases have gone up, is that because we have done more testing or because our lockdown is not working and we need stricter controls? It may mean we are actually less prepared for the second wave, because we cannot analyse the first wave properly, nor compare it to the second.
Conclusions
It seems that the only reason to do mass testing to see if the public have the disease, is to see what the impact of halting all constraints will be, or "just so you know". Are these good enough reasons to take NHS staff away from patients, skew data analysis or have imperfect tests quarantine healthy people and give false confidence to carriers? Personally I'm not sure it is. My gut feeling is that it would be best to concentrate the testing on those who need it most - ill patients and frontline staff. If you think differently and have any comments, you can add them below. I will try to update this article if people think of some things that I haven't.
Comments
Post a Comment