Coronavirus hospital admissions inflated at height of pandemic
Analysis for Sage concludes people were being counted as Covid-19 admissions if they had ever had the disease
By
Sarah Knapton,
SCIENCE EDITOR
21 August 2020 • 7:18am
Hospital admissions for Covid-19 were over-reported at the peak of the pandemic, with patients who were taken in for other illnesses being included in outbreak statistics, it has emerged.
An investigation for the Government's Scientific Advisory Group for Emergencies (Sage) found that people were being counted as Covid hospital admissions if they had ever had the virus, and were added to those being admitted directly due to it.
Government figures show that, at the peak of the pandemic in early April, nearly 20,000 people a week were being admitted to hospital with coronavirus (see graph below), but the true figure is unknown because of the problem with over-counting.
The oversight echoes recent problems with the data for Covid-19 deaths, in which it emerged that thousands of people who died of other causes were being included in coronavirus statistics if they had once tested positive.
Professor Graham Medley, of the London School of Hygiene and Tropical Medicine, asked by Sage to look into the situation, told The Telegraph: "By June, it was becoming clear that people were being admitted to hospital for non-Covid reasons who had tested positive many weeks before.
"Consequently, the NHS revised its situation report to accommodate this."
The investigation led to a readjustment of how the figures were compiled at the beginning of July.
On Thursday night, experts warned that the miscalculation was particularly concerning because the number had been used to reflect the current state of the virus epidemic.
Professor Carl Heneghan, the director of the Centre for Evidence-Based Medicine at the University of Oxford, said: "The admissions data is a crucial point. I'd say it is more important than the death data because it is the best marker of the impact of the disease."
The issue came to light on June 18, when Sage minutes record that Government scientists registered concern that non-Covid patients were being included in the outbreak data.
Public Health England, the NHS Medical Director and Prof Medley were asked to convene a group to find the "ground truth". Prof Medley said: "With any epidemic of a new disease, there is always a period during which the clinical manifestation of the disease is being developed.
"The initial aim is to be as complete as possible, even though there will be overestimation.
"So, we saw with the death statistics, that initially a 'Covid death' was defined as any death in which the person had previously tested positive.
"The same is true, but even more so, for hospital admissions. In order to capture all the potential manifestations of the disease at the start of the epidemic, all admissions of people who had tested positive were counted as 'Covid-related admissions'."
From July 1, the NHS began counting virus admissions as those who tested positive within a short period of entering hospital, and NHS England said there had not been a significant difference in the admissions trend line since the change. However, it is unclear how many people were misclassified as Covid-19 patients before July 1.
The over-count may have included patients admitted with an unrelated ailment but who tested positive for coronavirus on arrival, or who had previously been diagnosed with the virus. It may also include patients who were tested while already in hospital for a different condition.
Experts said knowing the correct admissions figures was crucial, not only for determining how many people are really being hospitalised with the virus but also for working out death rates. The table below shows how UK death rates compared with the rest of the world.
Prof Heneghan added: "If admissions are going up, then that should drive the lockdown. But currently you have people with active infections, those who have tested positive but have been discharged, and those who have contracted it in hospital, so it isn't helpful.
"This really does need sorting out as we go into the winter, otherwise we get into this noisy position where we can't understand what's going on. There will be loads of people coming in with different conditions who have survived this, so it's a huge problem. It's clouding our judgment as to whether the disease is having a significant impact."
There are also wide discrepancies between NHS England admissions data and that reported daily by the Government.
For example, on April 8, around the peak of the virus, NHS data suggests there were 480 admissions with Covid-19 and a further 2,264 people diagnosed with the virus in hospital – a total of 2,758. However the Government's figures suggest there were 2,340 overall admissions.
For March, NHS England recorded 15,810 admissions, while the Government recorded more than 20,000. Even as recently as August 9, the NHS recorded 33 admissions with Covid-19 and a further 40 diagnosis in hospital – a total of 73. Yet Government figures record 50 patients overall.
It is also still unclear how many people included in current admissions data are asymptomatic and were tested when in hospital for a different condition, or who had caught the disease once they were in hospital.
Prof Heneghan said it was possible that in hotspot areas, such as Oldham (see video below), there was now nobody in hospital with an active infection, but it was impossible to tell from the way the data was being recorded.
"The key, to me, is the admissions data," he said. "You could have 1,000 people with the infection, but if you have no admissions then you're actually building up a population more able to withstand the virus."
Scientists said compiling data at the start of an epidemic was tricky, but warned that the figures could be preventing proper understanding of the epidemic.
Professor Robert Dingwall, of Nottingham Trent University, who sits on the Government's New and Emerging Respiratory Virus Threats Advisory Group, which feeds into Sage, said: "It may be very difficult to go back now and recalculate the early data.
"But we should certainly be aware that it may have led to some exaggeration of the number of Covid hospital admissions and related deaths early on and it may partially explain the lower numbers that we are now seeing – although other factors are probably contributing more to the recent decline in both deaths and admissions."
Professor Stephen Powis, the National Medical Director of NHS England said: "It is completely untrue to claim there has been any ambiguity in the data on Covid positive hospital inpatients.
"For the safety of both patients and staff it is obviously sensible and important that NHS trusts monitor and track data on patients in their hospitals that have Covid, whether admitted as a Covid patient or for another condition.
"The instructions to hospitals in England have not fundamentally changed since the start of the pandemic, with updated advice to hospitals being published as scientific evidence has developed, as hospitals and Sage requested further advice, and as the availability of testing has become more widespread.
"Data on hospital admissions, people in hospital with Covid and people on mechanical ventilation in England, were published and presented daily at the Downing Street press conferences, along with clear and transparent explanations online, including, for example, clarification that some admissions in devolved administrations had to be reclassified and included suspected and not just confirmed cases.
"Throughout the pandemic, we have worked with Sage and follow their advice to ensure data is collected in a manner that supports their important work, including through mutually agreed changes to existing guidance."[\quote]
