On 10 March Scotland’s Deputy Chief Medical Officer, Gregor Smith, published a statement setting out the way that contact tracing was being used to counter the spread of coronavirus. He described contact tracing as “a fundamental part of outbreak control, used by public health professionals around the world.”

But at the First Minister’s press conference on 2 April, Chief Medical Officer Catherine Calderwood said: “the thought that the testing in some way slows the virus or is a part of our strategy to prevent transmission is a fallacy.”

What happened between these two statements? I sent the Scottish Government a freedom of information (FOI) request to try to find some of the answers. What I found was shocking.

How does contact tracing work?

Those who test positive for coronavirus will speak to a clinician who gathers detailed information on places they visited and people they came into close contact with since they became unwell or, in the case of international travellers, since they arrived in the UK.

This builds a very specific picture of the people who need to be contacted, such as family members, colleagues or fellow travellers.

Gregor Smith, Deputy Chief Medical Officer, 10 March 2020

Asked when the strategy set out by Gregor Smith was first implemented, the Scottish Government said (in a 6-page response that appears at the end of this post):

“It is an established process delivered by health protection professionals, dating back to the early 20th century. Contact tracing was therefore carried out for all cases in Scotland up until 13 March.

It was appropriate to put on hold contact tracing when it was announced the UK was moving from the ‘contain’ phase to the ‘delay’ phase on 12 March. The policy was changed to focus on providing the best care to help those who developed Covid-19, delaying and reducing the spread as much as possible, and protecting those who are particularly vulnerable.”

Asked whether contact tracing was still (31 March) in operation and, if not, when it was terminated, the Scottish Government said:

“On 12 March it was announced the UK was moving to the ‘delay’ phase, at which point those who displayed symptoms of a new continuous cough and/or a high temperature, regardless of their travel history or contact with confirmed cases, were asked to self-isolate.”

The Scottish Government added that a contact tracing programme known as Test and Protect was initiated in pilot form in Fife, Highland and Lanarkshire on 18 May and then rolled out across Scotland as a full programme on 28 May.

Astonishingly, the FOI statement confirms that no serious steps to counter the spread of coronavirus were being taken over the crucial period from 13 March to 26 March, when health protection regulations came into effect. Contact tracing had stopped, but the legal framework for lockdown was not yet in place. The number of COVID-19 patents in hospital in Scotland grew four-fold over this period, from 149 patients on 18 March (the first available data) to 575 patients on 26 March (SPICe spotlight).

No contact tracing at all was carried out in Scotland between 13 March and 18 May. And from 18-28 May limited contact-tracing was being done on a purely experimental basis, rather than in a prioritised way designed to have maximum impact on the spread of coronavirus.

No need for other colleagues to isolate as well

Asked about contact tracing in hospitals, the Scottish Government said:

“A staff member who has been in contact with anybody who has tested positive for COVID-19 whether at work (most likely a colleague in communal areas) or in the community will be required to isolate for 14 days in line with advice to the general population. There is no need for other colleagues to isolate as well, although it is possible that if a member of staff tests positive, they may have had contact with multiple colleagues or have been exposed to the same source of infection. It is, therefore, important that staff wear appropriate PPE when caring for or working with patients, and use it with appropriate hygiene measures for donning and doffing and disposing of it safely. It is also important to adhere rigorously to social distancing with colleagues and other staff when not providing care, for example, when having coffee breaks.

A staff member who has been caring for a person who has tested positive for COVID-19, or who has symptoms of COVID-19 while the staff member was wearing appropriate PPE, in the majority of cases, will not need to isolate. These cases will be referred to the local health protection team (HPT) to advise on, but unless there are very specific circumstances around the contact, it is likely that the staff member will be advised that they can continue to work as normal.”

A referral to the local HPT team does not mean that a contact tracing process has been initiated, unless the HPT team is in any case operating such a process. So contact tracing was apparently carried out amongst hospital staff only insofar as it was covered by community contact tracing. For most hospitals in Scotland, that means no contact tracing was carried between 13 March and 28 May.

The risk to staff and patients was supposedly mitigated by staff wearing PPE when working with patients, and observing social distancing with other staff.

A different strategy was adopted in England after 13 March. A Sunday Telegraph editorial on 31 May referring to a decision to “discontinue” contact tracing resulted in a sharp rebuttal from Public Health England (PHE). PHE said:

“The article claims that contact tracing was abandoned. Again, this is not the case. Once there was clear evidence of widespread, sustained community transmission and the Prime Minister announced the move to delay phase, contact tracing was unlikely to control the outbreak alone. At this point contact tracing was targeted where it could be most effective during this phase – focusing on the most vulnerable, for instance, carrying out contact tracing in care homes, hospitals and institutional environments.”

PHE does not dispute that community contact tracing was abandoned, but claims that contact tracing was carried out in hospitals and other institutions. In this respect, the shut-down of contact tracing was less complete in England than in Scotland.

Contact tracing has been called for repeatedly by the WHO, for example in interim guidance published on 13 March, and in comments by the WHO’s Director General, Tedros Adhanom Ghebreyesus, on 13 April. The WHO’s interim guidance for contact tracing in hospitals is more rigorous than for contact tracing in the community, and involves testing of all staff and patient contacts, whether they show symptoms or not.

Asked about constraints and challenges to contact tracing arising from costs, limited availability of equipment, staff or facilities, data protection considerations or any other considerations, the only challenge the Scottish Government noted was the “ability to predict the number of index cases that are likely to need their contacts traced, and the number of contacts that each index case will declare, as the pandemic progresses.”

The only explanation that the Scottish Government has given for its decision to put contact tracing on hold on 13 March is a remark that this step was “appropriate” when the UK moved from the “contain” to the “delay” phase.

Health Secretary Jeane Freeman MSP was equally unforthcoming in a statement issued on 1 June in response to an email passed to her by Alison Johnstone MSP from a constituent, Julia Davidson. The email said:

“It is of course true that testing, literally speaking, does not in itself slow the virus. Testing as part of a policy to test, trace contacts, and isolate them, is another matter and was the focus of my email to you and of much of the debate on this topic. Presumably the CMO must be correct in saying that it is a fallacy to think that testing is part of the Scottish Government’s strategy to prevent transmission. But if so, it is a disturbing feature of the strategy and sets it apart from what appears to be understood around the world as best practice.

Can you please confirm that the Scottish Government’s view is that it is now beyond the time when testing can assist in halting the spread of the virus, and let me know what evidence and analysis the Scottish Government has based that assessment on?

Can you please also ask the Scottish Government what part the availability of test kits has played in its assessment of the utility of a testing and contact tracing strategy?”

Jeane Freeman’s response focussed mainly on testing, rather than the contact tracing programme of which testing is a part. She says (in a 2-page response that appears at the end of this post):

“In order to make best use of available testing capacity we have identified three priorities for testing, these are: to direct our testing capacity effectively to save lives and protect the vulnerable; to ensure that key workers can return to work as soon as possible and to monitor and report on the spread and prevalence of the virus in the population and the impact of public health measures.”

This perhaps implies that testing capacity was an issue, though it was not raised as such in the Scottish Government’s later FOI response.

The first of Jeane Freeman’s three points – “to save lives and protect the vulnerable”is rather unspecific. Her second point emphasises returning to work “as soon as possible” instead, as might have been hoped, as safely as possible. Her third point relates to surveillance testing rather than to testing as part of a prevention strategy.

But she does in the end refer to the Test and Protect contact tracing strategy, saying:

“Now we have moved to ease current lockdown measures, our work increases with the introduction of NHS Test and Protect. The strategy for that was published on 4th of May.”

She simply takes it for granted, without explanation, that contact tracing was not carried out until lockdown eased. How did this assumption take root, counter to the advice of the WHO and other experts?

The Scottish Government referred me, at the end of its FOI response, to the minutes of the SAGE meeting on 20 February, which included a discussion item on trigger points for halting contact tracing. The minutes refer (without providing a link) to a PHE paper on monitoring and contact tracing. It appears to be this paper (the pdf file was created on 19 February).

The two-page paper outlines mathematical modelling by PHE and the universities of Manchester and Cambridge, proposes some additional modelling, and sets out three triggers. It is headed “When to stop contact tracing: Developing triggers from PHE systems.” The heading is odd, because nothing in the paper is directed towards establishing a point at which contact tracing would have to be stopped. Instead, it predicts that, as the number of cases grows, a point will be reached where “the CCI [Case and Contact Isolation] approach is no longer likely to result in successful containment” – in other words, contact tracing alone will be insufficient to contain the epidemic. The triggers are intended to give warning that this point has been reached.

The minutes of the SAGE meeting on 20 February record that SAGE thought the proposed triggers were “sensible” and that “SAGE should offer further advice should those triggers be met.” The next point recorded in the minutes is surprising. It says:

“Any decision to discontinue contact tracing will generate a public reaction – which requires consideration with input from behavioural scientists.”

Obfuscation

Concern that a halt to contact tracing would by itself generate a significant public reaction seems rather overblown. The point makes more sense if a halt to contact tracing is taken to imply the start of lockdown-type measures. There was indeed a public reaction to the move towards lockdown. In the two week before the lockdown regulations came into effect SAGE noted later that there was a surge in spending, a drop in the numbers of people using transport hubs, and a rise in the number of people using parks.

It looks as if the phrase “discontinue contact tracing” should be read as an obfuscation – a code if you prefer – more or less meaning “start lockdown.” In February many people still saw lockdown as the stuff of disaster movies, so it would not be surprising if SAGE was rather coy about it. But through February and March SAGE examined the components of what we now loosely call “lockdown” – school closures, a halt to gatherings, a reduction in non-household contacts. SAGE referred to these sort of measures at first as “non-pharmaceutical interventions” and later as “behavioural and social interventions.”

On 5 March SAGE minuted that surveillance of ICUs suggested that “sustained community transmission is underway in the UK” and that the Government should “plan for the introduction of behavioural and social interventions within 1-2 weeks.”

The evidence of sustained community transmission implies that PHE’s proposed triggers for what it had confusingly labelled as “when to stop contact tracing” had been met, but nothing in SAGE minutes spells this out or otherwise refers to the triggers.

On 13 March SAGE minuted:

“Community testing is ending today – which will increase the pace of testing (and delivery of results) for intensive care units, hospital admissions, targeted contact tracing for suspected clusters of cases and healthcare workers. This includes faster confirmation of negative results.”

This is as close as SAGE gets to acknowledging that community contact tracing stopped on 13 March. It strongly implies that the decision was driven by a shortage of test capacity. The change was intended to facilitate a focus on contact tracing for suspected clusters and amongst healthcare workers. But nothing like that happened in Scotland, if the Scottish Government’s FOI response is correct. There was no contact tracing around suspected clusters (perhaps there wasn’t in England either, since the PHE rebuttal of the 31 May Sunday Telegraph article doesn’t mention it). And there was no contact tracing amongst health workers in Scotland. Instead, testing of health workers focussed on getting them back to work.

SAGE renewed its interest in contact tracing at the end of April. A SAGE meeting on 19 May concluded that “an effective Test, Trace, and Isolate system will be necessary (but not sufficient on its own) to allow further substantive adjustments to distancing measures without pushing R above 1” – in other words, contact tracing was now seen as a necessary element in any move to ease lockdown restrictions. This was to be introduced as a new “Test, Trace, Isolate” strategy, apparently distinct from the well-established contact tracing strategy that operated until 13 March.

The Scottish Government’s current Test and Protect programme is in line with this new strategy, but nothing in SAGE’s documented discussions – notwithstanding the puzzling discussion about triggers to discontinue contact tracing – provides any grounds for believing that it was beneficial for Scotland’s contact tracing programme to be put on hold between 13 March and 28 May.

It must surely be obvious that contact tracing over this period would have reduced the number of people infected, saved lives and provided a firmer base for subsequent moves to ease restrictions. The Scottish Government has not claimed, even when more or less invited to do so in my FOI request, that its decision was driven by resource limitations. But the minutes of the SAGE meeting on 13 March suggest very strongly that the decision was in fact driven by a UK-wide shortage of testing capacity. In these circumstances, contact tracing up to the limit of capacity, prioritised appropriately, would surely have been beneficial. Instead:

  • No contact tracing at all was carried out anywhere in Scotland between 13 March and 18 May.
  • In most of Scotland (including Glasgow and Edinburgh), no contact tracing was carried out between 13 March and 28 May.
  • Contact tracing was halted two weeks before “lockdown” regulations came into effect, creating a hiatus in which no serious steps were being taken to counter the spread of the virus.
  • No exception to the suspension of contact tracing was made for hospitals, care homes etc, though this was done in England, according to Public Health England.

Systemic failure of governance

The decision to suspend contact tracing because of a shortfall in testing capacity was hidden behind an unexplained claim that contact tracing was automatically inappropriate after the UK moved from the “contain” to the “delay” phase of its coronavirus strategy. The basis for this claim seems to be a PHE and SAGE discussion in February that was aimed at determining the point at which contact tracing alone would no longer be sufficient to contain the epidemic and further measures would be needed. A foolish obfuscation used to describe the discussion has been foolishly – and I think wilfully – misunderstood ever since. The effect was to create a dogma that strangled effective decision-making.

Was testing capacity sufficient in mid-March to maintain targeted contact tracing even if community contact tracing had to be scaled back? Probably. That was what SAGE expected would happen. But the question disappeared beneath the dogma that contact tracing was no longer appropriate.

In the absence of adequate testing capacity would it have been useful to trace contacts of unconfirmed cases? That might have involved asking substantial numbers of people to self-isolate needlessly, but perhaps that would not have been too great a burden at a time when everyone was sharply reducing their social contacts. The question disappeared beneath dogma.

And what of the ramping up of testing capacity during April and May? How should the new capacity have been used? Media coverage left a troubling impression that the Scottish Government was simultaneously struggling to meet its targets for testing capacity and struggling to find a use for the tests. At a press briefing on 1 May Nicola Sturgeon said:

“This expansion of testing that I set out today is separate and distinct from our move to establish a test, trace, isolate system as part of our approach to changing and hopefully alleviating the lockdown measures.”

Dogma made it impossible to use the expanded testing capacity to immediately reinstate contact tracing. It was simply not “appropriate”. Contact tracing was positioned instead as a future project, part of a good news story aimed at easing lockdown and getting people back to work. And so an opportunity to save lives was missed.

A narrative that was in the first place perhaps intended to deflect public attention from an embarrassing shortage of testing capacity had become a policy straitjacket. If anyone in or around government understood what had happened or felt inclined to question it, they took care not to say so. Everyone apparently believed that it had somehow been scientifically established that contact tracing was not “appropriate” during the delay phase.

People who prefer cockup theories to conspiracy theories will perhaps find this story gratifying. They shouldn’t. If the stakes were less high it might be an endearing tale of how humans behave in bureaucracies. But many thousands of lives were at stake. At least since the beginning of March – very belatedly, but let that pass – coronavirus strategy has commanded the attention of senior politicians and has been played out under the glare of the brightest lights our media can muster. It has displaced much of the ordinary business of society and government. Under these circumstances a misunderstanding should not persist for more than a day or two.

The fact that it has done so is symptomatic of a systemic failure of governance, across all its branches and through all its layers. As the stakes get higher and we approach the problems that the future holds, we should think carefully about that.

Photo: © No-Mad/Shutterstock

FOI response from the Scottish Government, 18 June 2020

Letter from Health Secretary Jeane Freeman MSP, 1 June 2020