On Friday, Feb. 21, Duncan Selbie, chief executive of Public Health England, was in a cheerful mood. It was near the end of the school half-term holiday.
He wrote on an official blog that there had been no new positive cases of the new coronavirus that week in the United Kingdom. It was a “testament,” he said, “to the robust infection control measures” and the “diagnostic and testing work” at laboratories nationwide.
Selbie, who joked when he took his then £185,000-a-year job in 2013 that his public health credentials could be fitted “on a postage stamp,” headed a government agency with a mission to prepare for and respond to public health emergencies. He was now facing an epochal one.
Public Health England (PHE) is a cornerstone of the UK’s state-run health system; Selbie reports directly to the government’s health minister, Matt Hancock.
The agency is much smaller and has a lower profile than the vast National Health Service (NHS) that oversees Britain’s hospitals, clinics, and general practitioners, providing healthcare to all. Public Health England’s 5,500 staff have an important supporting role running a laboratory, developing tests for new infections, and managing outbreaks in England, home to 56 million of the UK’s 67 million people.
ven as Selbie posted his blog, all was not well. Reuters estimates that across the UK between 1,500 and 5,000 people may already have been infected with the coronavirus. The estimate is based on current scientific assessments of the fatality rate from COVID-19, the disease caused by the virus, and an average interval from infection to death of between 18 and 23 days.
Within five weeks of Selbie’s blog, Britain had hit the peak of an outbreak that has caused at least 43,000 deaths so far, according to the official toll. This compares to 35,000 in Italy, the first country in Europe to be severely affected, or nearly 9,000 in Germany, where, in the early days, the spread of infection appeared similar to Britain. With the exception only of Belgium, Britain has the highest per-capita mortality rate from COVID-19 among major economies, Reuters data show.
As Reuters has reported previously, Britain was slow to impose lockdown measures. That delay was costly. Professor Neil Ferguson, a disease modeling expert from Imperial College, London, has said that introducing lockdown measures a week earlier “would have reduced the final death toll by at least half.”
Now, a Reuters investigation reveals further missteps and failures by officials and government agencies, including Selbie’s Public Health England, in testing, tracking and tracing. Among decisions that doctors and epidemiologists say cost lives were:
• Failure to build up capacity to perform mass tests for COVID-19.
• Deciding on a narrower definition of COVID-19 than used by the World Health Organization and other countries.
• A decision to abandon testing of most people who didn’t require hospitalization, and failure, early on, to create any way to track infection.
• A decision to abandon a programme of widespread “contact tracing,” in which people in contact with an infected person were traced and told to isolate to stop the outbreak spreading.
• Deciding to share almost no details about the location of infections with local public health officials or the public.
• Fragmenting local responsibility for public health.
“Every mistake that was made did, unfortunately, cost lives,” said Professor Tim Spector, an epidemiologist at King’s College London.
Prime Minister Boris Johnson has said his government followed scientific advice at all times. Ministers point out that from mid-April, the UK dramatically increased the number of tests carried out and recruited thousands of contact tracers to control outbreaks.
In a statement, Public Health England chief executive Selbie told Reuters his agency’s work had helped to save lives. “Beating this virus was always going to be a marathon not a sprint,” Selbie said. “It remains our sole focus and we have done our absolute best with all our available resources.”
Still, as the country began emerging from its lockdown this month, there was another blow to the government. It was forced to delay until at least autumn a contact-tracing app that it had hoped would be a game changer in tracking the virus and averting a second wave of infection. It emerged the app didn’t work properly with iPhones.
By Feb. 23, when most British schools had ended their half-term holidays, Reuters estimates there were thousands of COVID-19 infections across the country. In areas such as Birmingham and nearby communities, there were already hundreds of infections, Reuters calculates*. Within a few weeks, the virus had taken hold and was rapidly spreading across the country.
For every COVID-19-related death, Reuters estimated there were at least 100 new infections on average.
Once it surfaced, the virus spread rapidly in Britain. Infections peaked weeks earlier than the authorities expected not just in England, but in all parts of the UK.
England’s Chief Medical Officer Chris Whitty, the government’s top health adviser, recently conceded, “Many of the problems we had came because we were unable to actually work out exactly where we were, and we were trying to see our way through the fog.”
Several senior doctors, who led the response to COVID-19 at major hospitals, told Reuters the fog was partly a result of early decisions by Whitty and other advisers that restricted testing for the virus. The doctors asked not to be named because they were not authorised by the NHS to speak publicly.
blic Health England had two main ways of monitoring and controlling the virus: a programme of testing and one of contact tracing to track down the contacts of anyone with the infection. These methods depended on what doctors call a “case definition,” the description of what a COVID-19 case looks like, to determine who should be tested.
In early February, the World Health Organization listed eight symptoms for COVID-19, including nausea and vomiting. In Germany, the government’s disease-control agency, the Robert Koch Institute, said doctors should use their judgment in ordering tests and it encouraged tests even for those with “unspecified general symptoms,” meaning mild symptoms.
Britain took a different course. Whitty and his colleagues adopted a narrower view, defining COVID-19 symptoms as only a sharp new cough, fever or shortness of breath. Britain clung to this definition through the peak of the outbreak. It was not updated until May 18 – after 34,796 deaths – when a loss of taste and smell was added to the symptom list.
Spector, the King’s College professor, said Britain’s refusal to recognize more symptoms created space for the virus to spread. Because testing was restricted to people exhibiting the UK’s narrow range of symptoms, many cases likely went unnoticed and the virus spread unchecked.
“The country that invented epidemiology was the only one that had no surveillance programme and no idea what was going on in their own country,” he told Reuters.
Several senior doctors said they had pleaded with the authorities to expand the case definition. But they got nowhere.
One London consultant said a Public Health England laboratory refused to accept test samples from patients who didn’t meet the UK definition. Asked if this was true, a Public Health England spokesman said: “That is right, but there is nothing unusual in this. Lab testing capacity was used to establish whether people that met the case definition had the virus.”
Britain also restricted testing in other ways. At first it would test only those who had come into contact with confirmed cases and those with symptoms who had travelled from Wuhan, China, the origin of the outbreak, then from February 1 the rest of China, and six days later from other parts of Asia. Parts of Italy, the first European country to be badly hit by the virus, and Iran were added on February 25. Others were excluded.
“We were testing all these people with a cough from China and parts of Asia,” said one senior doctor at a London teaching hospital, but meanwhile cases from other countries passed unnoticed.
Community testing in London, using the narrow guidelines, found only roughly 50 positive cases among some 5,000 tested by mid-March, another infectious disease specialist doctor in the capital said. Reuters has estimated that more than 18,000 Londoners were or had been infected with the virus by the end of February.
“A lot of us felt it was looking for a needle in a haystack,” this doctor said.
A large genetic study would later show that most UK infections had come from Europe, not Asia. Beginning in early February, the virus strain could be traced back to Italy, then Spain and France, a team from the University of Oxford and University of Edinburgh found.
Many frontline doctors believe the disease was carried home to Britain from Europe’s ski resorts. The biggest influx was when families returned from their half-term break, shortly after Selbie’s February 21 blog post that celebrated recording no further cases.
Public Health England told Reuters it didn’t set policy. The Department of Health and Social Care, Britain’s health ministry, said the case definition was set by the government’s top medical advisers, based on scientific advice. The definition was continually reviewed, the ministry said. “The symptoms selected were those that captured those most likely to have COVID-19, while not capturing a great number of people who did not,” the department said in a response to Reuters.
In Scotland, which followed the same UK-wide case definition, there were also far more early cases than realised. In the Glasgow region, for instance, Reuters calculated there were at least 200 cases by the end of February. But, according to official data, there were no cases.
Who’s in charge?
The government would later concede the UK hadn’t tested enough. Matt Hancock, the health minister, blamed a shortage of lab capacity. “We have the best scientific labs in the world, but we did not have the scale,” he said in April. Germany, in contrast, “could call upon a hundred test labs, ready and waiting when the crisis struck.” He did not specify how many labs the UK had.
But according to Mark Reed, general manager of Pro-Lab Diagnostics, a Canadian-owned company, that wasn’t the case. “Every maker of diagnostics was saying ‘We are here.’ I don’t think the government engaged enough; they underestimated what was required.”
Reed and officials from two other diagnostic firms – Novacyt and GeneFirst Ltd – told Reuters that Public Health England evaluations of their testing kits were not completed until mid-to-late March. As a result, NHS labs were reluctant to use tests made by the three companies, they said. Public Health England told Reuters that although it carried out evaluations, it wasn’t its job to approve or recommend tests. It didn’t comment about these specific cases.
By contrast, laboratories in Germany began using commercially made tests from Jan. 27 and then validated new ones as they became available, said Evangelos Kotsopoulos, CEO at laboratory firm Sonic Healthcare Germany. Into February, “we simply grabbed every single instrument we could find,” while hiring new lab staff and stockpiling crucial chemicals needed for testing.
Of all countries, South Korea had best demonstrated the value of mass testing early in the pandemic, making use of commercially supplied kits. One, from South Korea’s Seegene Inc, was not evaluated until mid-March by Public Health England. Based on that evaluation, NHS labs began purchasing and using it in early April. German labs had been using that test since mid-February.
A spokesman for Public Health England said that, within the UK’s overall strategy, the agency’s role “was to provide an assured testing capability” for the virus “while commercial partners worked to develop their own capability.” NHS labs were free to order and validate their own private test kits or develop their own, the agency said.
The heads of three NHS labs told Reuters otherwise: They said they had waited for Public Health England’s evaluation before placing such orders.
The NHS didn’t directly address a question from Reuters about the evaluation of tests by Public Health England. It said in a statement that overall policy on testing is set by the government, informed by advice from Public Health England. The statement added that “as more reagent supplies became available, NHS-run laboratories were able to increase their capacity to test for COVID-19.”
The bulk of tests for the coronavirus in the early stage of the outbreak was done within Public Health England’s own labs. A clinical adviser to the government said the agency had been “very centralised” and “made quite clear they are controlling all of the testing.”
At the beginning, “the only game in town was the PHE test,” agreed Allan Wilson, president of the Institute of Biomedical Science, which represents NHS laboratory staff.
The system wasn’t designed for fast results.
Ravi Gupta, a Cambridge University professor of clinical microbiology, said patients at Cambridge’s Addenbrooke’s Hospital had likely infected others with the coronavirus while kept on wards waiting for test results. In the early weeks, he said, they had to wait two to four days. Gupta helped develop a 90-minute test.
Members of parliament’s science committee criticised Public Health England in May for initially restricting testing to a small number of labs, calling it “one of the most consequential” decisions of the crisis. Selbie responded by stating the responsibility for testing policy lay with the government’s health ministry, the Department of Health. “PHE operates reference laboratories for novel and dangerous pathogens and would not be expected to operate mass testing programmes,” he said.
In Birmingham, the virus arrived unnoticed.
Zia Choudhry remembers his father, Mohamed Aslam Wassan, a 67-year-old retired businessman “full of love and full of energy” coming home with a cough on March 15. Choudhry lived with his dad and his mother, together with his wife and children, in a house in Lozells, a district of Birmingham.
In the next few weeks, Wassan, remembered by some for bringing a famous snack, Bombay Mix, to the city, would die of COVID-19. So would four of his close friends, all well-known faces in the local Pakistani community. They were among at least 1,000 who would die across the city.
Lozells, where many people are of Pakistani or Bangladeshi origin, was particularly hard hit – with at least 25 people dying. Yet, for many weeks into the outbreak, local health officials were kept in the dark by central government and Public Health England.
In interviews, seven directors of public health at large local authorities told Reuters they received little more information than the government released to the public on its website. The directors, who are employed by local authorities, not by Public Health England, said this lack of communication hampered their efforts to prevent the virus spreading.
“We were flying blind,” said Lisa McNally, director of public health in Sandwell, a district neighbouring Birmingham.
McNally said the government refused to share models used by its scientific advisers and the data that fed them. They wouldn’t reveal, for example, how many people from her area were calling the NHS helpline with COVID-19 symptoms. If local officials had been kept informed there could have been special efforts to work out the source of infection, or an information campaign aimed at a particular area to reinforce the lockdown. “We basically would have done contact tracing,” she said.
Louise Smith, the director of health in the rural county of Norfolk, in eastern England, was angry that mortality and case data weren’t shared by local hospitals. From March to mid-April, she said, “I was finding out about the deaths by reading the local newspaper.”
In Birmingham, the city’s health director Justin Varney said when the first deaths occurred in early March, he used to get a personal call from Chris Whitty or his deputy. But he was given neither a name nor postcode of the victim. It was just the “barest minimum of details” and he was told “not do anything with that until the NHS announced it.”
Whitty did not respond to a request for comment.
And the information blackout continued. Throughout most of the outbreak, Birmingham’s Varney, like other public health directors, was given no more detail about deaths and infections than was published on national websites, with the exception of information about outbreaks in nursing homes or other institutions. A breakdown of deaths by neighbourhood, which would have identified clusters like in Lozells, was not released until May 1.
Secrecy was pervasive. Dr Helen Carter, Public Health England’s deputy director in the Midlands, revealed in a May 15 memo to Birmingham councillors that the government had initially ordered the agency not to share with local councils its surveillance reports containing data about notified cases from local hospital emergency rooms and general practitioners.
“Until April 2020 PHE was not permitted to share the surveillance reports with local partners by the Cabinet Office,” she wrote in the memo, which is reported here for the first time. After April, information was shared with public health directors, but the reports remained official secrets and were “marked as official sensitive and not in the public domain,” said the memo, which was reviewed by Reuters.
ublic Health England declined requests by Reuters to interview Carter or any member of its staff involved in contact tracing. A spokesman said some of its surveillance reports were classified as “official sensitive” which, the spokesman said, was standard practice for non-public information. The agency had a “constant dialogue” with local health directors, with “routine exchanges of information and data,” the spokesman added.
The government also restricted the power of local authorities to inform their communities about the spread of the virus. Health director Varney, who works for the city of Birmingham, said in the beginning there had been “a complete embargo on local government doing proactive (communications) about COVID. At this stage, everything was being centrally controlled.”
A spokesman for the health ministry declined to respond to questions about its relationship with local health directors or its media strategy.
In Varney’s view, communication was urgently needed to prevent misunderstandings about the virus taking hold and to target communities where many elderly people have poor English. He used a regulatory loophole to get the message out: He found he was allowed to talk about issues affecting “community cohesion.”
Calling a halt
Sophie Grenville and her husband, Mark, who manage a country estate in the county of Hampshire, took matters into their own hands.
It was early March when the couple tested positive for the coronavirus. She described her experiences later in a journal she shared with Reuters called “A Diary of a Super Spreader.” Her case illustrates the complexity and the slow pace of the early UK version of contact tracing.
The couple was tested for the virus after returning from a trip to Indonesia on Feb. 23. They developed what grew into a very distinctive, dry cough – like a fox or a deer barking – and reported it to the NHS helpline. They waited three days for a test. All the while, she wrote in the diary, she and her husband were thinking of the hundreds of people they recently had come into contact with at social and business events, from “High Court judges to football chiefs.”
As Britain’s COVID-19 infections soared in the spring, the government reached for what it hoped could be a game-changer – a smartphone app that could automate some of the work of human contact tracers.
The origin of the NHS COVID-19 App goes back to a meeting on March 7 when three Oxford scientists met experts at NHSX, the technical arm of the UK’s health service. The scientists presented an analysis that concluded manual contact tracing alone couldn’t control the epidemic.
“Given the infectiousness of SARS-CoV-2 and the high proportion of transmissions from presymptomatic individuals, controlling the epidemic by manual contact tracing is infeasible,” concluded the Oxford scientists’ paper, which was published in the journal Science two months later.
The Oxford researchers believed that a smartphone app could help locate individuals who didn’t know they were infected – and by alerting them quickly could reduce and even halt the epidemic if enough people used it. Within days of the meeting, NHSX began the process of awarding millions of dollars worth of no-bid contracts to develop such an app, government procurement records show.
In the weeks that followed, ministers seized on the technology as a route out of Britain’s lockdown that began on March 23. At a Downing Street coronavirus briefing on April 12, health secretary Matt Hancock announced that testing had begun on what he called the government’s “next step – a new NHS app for contact tracing.”
He explained that people could use the app to report feeling unwell and it would anonymously alert other app users who recently had been in close contact with them. On April 28, he said he expected the app to be ready by mid-May.
Privately, some researchers who had proposed the app were dismayed that the government had stopped widespread testing on March 12, a decision they believed undermined the app’s effectiveness and public health in general. “We were very clear from the start that this thing needed to work with testing,” David Bonsall, a clinical scientist at Oxford who attended the March 7 meeting, told Reuters.
By early May, transport secretary Grant Shapps was heralding a test of the app on England’s Isle of Wight. “Later in the month, that app will be rolled out and deployed, assuming the tests are successful, of course, to the population at large,” he said. “This is a fantastic way to ensure that we are able to really keep a lid on this going forward.”
Pat Gelsinger, chief executive of VMware Inc, a Silicon Valley tech firm hired to develop the app, told a Fox Business television interviewer on May 8, “I tell you, we think this is the best one in the world and we’re really thrilled to be working with the NHS in the UK to help bring it about.”
But by the end of May, government officials were downplaying the app. In an interview with Sky News, Hancock called the app “helpful” but said traditional contact tracing needed to be rolled out first. Quoting another official, he said, “It puts the cherry on the cake but isn’t the cake.”
Behind the scenes, NHSX testers were discovering serious technical problems.
The agency had opted to develop an app that collected and stored data on central servers that could be used by health authorities and epidemiologists to study the disease. It relied on a technology called Bluetooth to determine who recently had been near someone displaying symptoms and for how long.
NHSX testers were finding that while the app could detect three-quarters of nearby smartphones using Google’s Android operating system, it sometimes could only identify four percent of Apple iPhones, according to government officials. The problem was that, on Apple devices, the app often couldn’t utilize Bluetooth because of a design choice by Apple to preserve user privacy and prolong battery life.
The issue was no secret. Apple and Google had jointly announced in April that they would release a toolkit to better enable Bluetooth on contact-tracing apps. But to protect user privacy, it would only work on apps that stored data on phones, not central servers. The NHSX app didn’t work that way.
The government insisted it had developed a successful work-around to overcome the Apple issue. But not everyone was convinced. The advocacy group Privacy International, which had tested the app in early May, “found it wasn’t working properly on iPhones,” Gus Hosein, the group’s executive director, told Reuters. But because of the government’s assurances, he said, “We just assumed we were doing something wrong.”
Other countries, including Germany, decided they would change their apps to work with the Apple-Google toolkit. That raised another problem with the UK app — it likely wouldn’t be compatible with many other contact-tracing apps so British travellers wouldn’t be notified if they were exposed to the virus.
On June 18, weeks after the UK app was supposed to be rolled out, government officials announced a dramatic U-turn — they would abandon the app being tested on the Isle of Wight and try to create one that worked with the Apple-Google technology. Work had already begun on it and they had learned lessons from the test, they said.
NHSX referred questions about the app to the health department, which said, “Developing effective contract-tracing technology is a challenge facing countries around the world and there is currently no solution that is accurate enough on estimating distance, identifying other users and calculating duration which are all required for contract tracing.”
A spokesman for VMware said it “is proud of the work we have done and continue to do to rapidly develop an application to support the UK’s contact tracing and testing efforts.”
A government official expressed confidence the app would be ready by the autumn or winter — although initially, the official said, it might not contain contact tracing at all, but offer other services that are yet to be determined. Reuters