COVID-19 reporting, or why it is stupid to stop testing

As part of a transition to “living with the virus”, many countries are abandoning mass testing. COVID-19 is losing the status of a “notifiable disease” and so it is becoming increasingly more difficult to actually draw any conclusions on what is actually going on.

Looking first at the number of reported cases in countries like the UK, Denmark, Sweden, Israel and the US, the severity of the COVID-19 outbreak seems to be rapidly declining across the board:

Even in Denmark, which experienced a large Omicron wave earlier this year and was one of the first countries to abandon all non-pharmaceutical measures, the numbers start to go down. Sweden and Israel both exhibit a huge decline in reported cases.

The decline in the UK and the US is much slower, but noticeable – and the UK is world-leading in the removal of all restrictions.

While the cases are declining, so is the number of tests:

The resulting proportion of positive tests seems to tell a very different story to the one told by the number of cases:

Norway and Sweden still experience a rapid increase in the share of positive tests, reaching the levels of 60%-70% which basically suggest that only those who know are ill and are willing to test – and report the test results – are bothered to do so.

Denmark sees a slightly slower growth and Israel appears to have plateaued. The UK has been seeing constant levels since the beginning of the year and certainly does not see any decline, so trumped up in the news. It is only the US that shows consistency between the two measures of disease severity.

The share of daily COVID-19 tests is a less intuitive but more robust measure of the severity of the outbreak, as the – more intuitive – reported number of cases depends on testing intensity.

So, is the epidemic still rising in these countries? What can we learn from studying the number of deaths reported?

WHO recommends that when the share of positive tests exceeds 5%, this means not enough is being done to detect and report the disease levels. At 70%, the reporting results become completely unreliable.

The deaths time series is delayed compared to the number of cases, sometimes by many weeks. The rapid increase in reported cases seen in Israel peaked around January 24th. In contrast, the wave of deaths was only starting then and peaked around 3rd February – a 2 weeks delay.

So, we will have to wait a bit to see what the trends are. Nevertheless, deaths seem to be still on the rise in Sweden and Denmark, plateauing in Israel in Norway, and slowly declining in the US and the UK.

Why do we bother? With ridiculously low testing levels the reported number of cases simple becomes completely irrelevant. Soon in countries like Israel, Sweden, Norway and the UK, we will have no idea what the virus is doing.

When a new variant emerges, we will have abandoned one of the key elements of our strategy to fight it off. What a stupid thing to do.