Hospitals are not hiding Covid cases. There really isn't much Covid around right now. Be wary of those who say otherwise.

A somewhat angry post.

The problem

Throughout the pandemic I have reported on what the data are telling us about the state of the Covid pandemic. I’ve done so in good faith and with due care to the inherent biases in the data - Kit Yates and I even wrote an article about how to interpret Covid data for Science magazine.

So, while I alert people when there are signs of a new surge (most recently the JN.1 wave in December), I am also clear on when things are getting better or flat and low, like right now (see below).

Increasingly, however, when I post better news such as the above chart, I get a chorus of responses “but there’s no testing”, the implication being that both prevalence and hospital admissions with Covid are actually high and that the data are not useful. This is wrong.

Worse still are charts being shared purporting to prove that cases and admissions are high, that this is somehow being covered up, and that anyone saying cases are low is, for instance, a “douchebag” (see below). This is also wrong. The authors of such tweets are either being deliberately deceptive or they do not understand the data - either way, their data interpretations should not be trusted.

In this post I am going to take you through what the chart above actually represents (nothing scary) and then pull together all the current data to show that there really is not much Covid around right now. Insisting otherwise is, quite simply, a non-evidenced based belief.

Reported cases, Covid hospital admissions, hospital testing - what is going on?

Reported cases for people who are not in hospital

People talk about cases in England, but what they are actually talking about are reported cases. To make it onto the UK Health Security Agency (UKHSA) dashboard as a non-hospitalised case, someone has to take a test, test positive, and then report that positive result to the UKHSA. Test availability and reporting requirements have changed a lot over the pandemic (e.g. see the helpful timeline below up to March 2022). Even during peak testing times, most people did not report rapid LFD test results to the UKHSA. Things then changed dramatically in March 2022.

The legal requirement to isolate if positive stopped in February 2022, and both the requirement to report test results and access to free LFD tests stopped for almost everyone in March 2022. Results from shop-bought tests (for a reason I will never understand) simply could not be reported to UKHSA. As people’s stocks of free NHS tests got used up after March 2022, their ability to report positive tests ended (even if they wanted to report). With no real expectation to test, let alone report, results, reporting of test results has continued to fall over the past year. The final nail in the coffin for reporting of LFD tests came in February this year, when the UKHSA closed its reporting portal. Hospitals or care homes can still batch report staff test results, but this is not giving a high number of reported test results.

Hospital testing

So since February 2024 , almost the only way to make it onto the UKHSA case dashboard is if you are a Covid hospital patient. But what has happened to Covid testing in hospital over time? To see this we can use UKHSA data on the number of weekly PCR tests (almost all PCR tests are on hospital patients).

Firstly, hospitals are doing much less testing of patients than they used to. In April 2023, they stopped testing most asymptomatic patients and switched to testing mainly patients who had symptoms of Covid or where a positive Covid test would change a patient’s treatment plan. This led to a large drop in PCR tests conducted in hospitals after 1 April 2023 (very clear in the chart below). However, since then, the number of tests within hospital have been stable and in fact increased last autumn and winter, due to the fact that more people were presenting with potential Covid symptoms given the Covid waves in October and December 2023.

This means that while there will be more (asymptomatic) people in hospital with Covid than recorded, trends in hospital admissions since April 2023 should reflect trends in Covid prevalence in the community. Overall, we might expect hospital admissions with Covid to go down over time for the same community prevalence, as the most vulnerable to severe acute disease add to their imunity through the vaccine booster programme. So, when hospital admissions are low compared to a year ago (as now, see red line on chart below), it is a combination of the fact that Covid levels have fallen and that the vaccine is protecting our more vulnerable people.

The below chart also shows reported cases (in blue) - while the peaks during autumn and winter peaks last year are clear, the absolute numbers are far below the actual number of new cases (e.g. peak reported daily cases was 1,500 in December 2023, whereas from the ONS Winter infection survey, we know over 200,000 people were getting infected each day at the December peak). We can also see that reported cases dropped a lot in April 2023 (after most asymptomatic testing in hospitals stopped). Finally - because LFD reporting was switched off in February 2024 - reported cases now match hospital admissions almost exactly. This is not surprising.

The misleading chart

You can plot the number of hospital admissions divided by the number of reported cases (below), which is what the author of the original tweet did. But this does NOT tell you anything anymore about the chance of being admitted to hospital if you have Covid. Nor does it tell you anything about how much hospitals are testing - the ratio has come close to 100% in recent weeks because of changes in how cases are being reported, not in how hospitals are testing. It is neither surprising nor scary. Despite the author of the original tweet asserting “It means that loads more people are being hospitalised by it without being tested”, this is not true.

In fact, UKHSA report the positivity rate in hospital Covid tests - that is, the percentage of people given a PCR test who test positive (below). As you can see, this is currently around 4% and as low as it’s been since last summer. The peaks in autumn and December suggest, to me at least, that hospital testing policy hasn’t changed that much (they test if a patient is sick with flu-like symptoms) but that during Covid waves, people with flu-like symptoms are more likely to have Covid (instead of say flu or RSV). There is absolutely no evidence to say that there are loads of symptomatic people sick with Covid in hospital who are being missed.

I believe that the kind of chart that prompted me to write this post is deceptive. The aim is to find a way of presenting available data to convince people that Covid is still raging or that there are far more people very sick with Covid in hospital than are being reported. In fact, the data used in that chart (as I hope I have convinced you) does not support that. It could be that the author simply didn’t understand the data they were plotting (in which case, they shouldn’t be plotting it). However, there is a lot of other relevant data showing a more positive view that they could also have shown, but did not. Why not?

If you see posts like these ask yourself:

• has the person explained how they got that chart and provided the context for it?

• If showing a ratio, do they also show/discuss its constituent parts (i.e. the numerator and denominator?)

• Do they ever acknoweldge data sources or new research that presents a more positive picture?

Incidentally, these sorts of questions are also valid for those denying Covid was ever a problem or those who deny the efficacy of the vaccines.

ALL the available evidence points to low levels of Covid in the community

When I say that there is not much Covid around right now, I am not basing it just on hospital admission data. To make that judgement, you have to look at a range of evidence. While available data is getting less, there is nonetheless enough to corroborate the hospital data.

First - the ONS winter infection survey which tested a random, representative, sample every week from November 2023 to last month. The last estimate they gave was from the beginning of March and that showed that the number of new cases was low and flat.

Monitoring wastewater is another way of getting a handle on how much Covid is around which does not rely on people testing or reporting test results. England does not monitor wastewater (it absolutely should by the way) but other countries do - and we can use them as a proxy for England. The most sensible proxy is Scotland and that also shows very low levels up to 4 March.

Our European neighbours’ wastewater monitoring is also showing low levels of virus after a winter peak, including Denmark, Finland, Austria, Netherlands, and Germany (shown below to 11 March). The US is showing exactly the same pattern.

A final check is whether there are any new variants growing that might drive a new wave. And the answer at the moment in England is “no” - as seen from the UKHSA dashboard below, the JN.1 variant is still dominant.

Four years on from the first wave and things are better

We are now four years on from the first devastating wave of Covid in England. And things are better. They aren’t perfect - I would much much prefer a world where Covid did not exist - but they are, measurably and meaningfully, better.

This does not mean that we got to this place with a good strategy (we mostly didn’t, apart from the initial vaccination roll out). It does not mean that hundreds and thousands of people aren’t still living with debilitating Long Covid, which I wrote about a few months ago with Prof Danny Altmann. It does not mean that Covid has lost its teeth, particularly for the clinically vulnerable or those with Long Covid who are at risk of worsening with reinfection. It does not mean we shouldn’t worry about longer term consequences of Covid infection. It does not mean that I don’t wish that we were testing more (especially wastewater and the ONS infection survey). It does not mean we shouldn’t be trying to reduce transmission of all airborne pathogens through cleaner indoor air and masks in healthcare settings. It does not meant we won’t see any more Covid waves (we definitely will).

But it does mean that for most people, their chance of either getting very sick or dying from Covid, or developing new Long Covid, is now much much lower than it was four years ago. And I, for one, am glad of it.

PS To be clear, I have deliberately not published names or tweet handles because I don’t want to cause any pile-ons.