How’s the smartphone contact tracing going then, or is it dead?

It’s now over three weeks since I installed the UK’s much-vaunted and -debated smartphone app to trace contacts of those with symptoms of Covid-19. Although we’re the first substantial trial of this new app, I’m surprised to report that, after those three weeks, we’re still using its initial release, version 1.0.2. This is despite repeated promises of an update to address changes in diagnostic criteria, to fix a batch of security issues which have been detailed openly, and the launch of a national test and contact tracing service – which doesn’t integrate with the app.

The latest feedback we’ve been given about this trial comes via our MP and the local newspaper, who wishfully describe it as “a big success”.

Looking at the figures reported I’m not convinced. Adoption hasn’t been as great as might have been anticipated, and I still don’t think that accurate figures have been released. The MP is quoted as saying that “around 56,000 Islanders now had the app on their phones”, which coincidentally happens to be exactly 70% of the original predicted total number of smartphones on the Island, 80,000. It also doesn’t reveal how many have actually registered the app with NHSX, which is the important number.

The only figures given for performance of the app are the total who have been tested for Covid-19 in the local centre, which is given as “around 500”, or about the same 25 cases per day claimed shortly after the release of the app. As the test centre will have an exact figure, and “no number could be put on the number of positive results”, it’s impossible to tell whether the app is accomplishing anything useful at all.

The UK government, which originally claimed the app was at the centre of its strategy to control the transmission of Covid-19, also seems to have lost interest. The Health Minister Nadine Dorries is quoted as saying that the app “would be used”, but wouldn’t give a date for its national roll-out. A glance at the app’s GitHub confirms that there has been almost no activity there since the trial started here on 7 May.

The inception of the English national test and contact tracing service on 28 May came as a surprise to those of us in the trial. For a start, criteria for contact tracing seem to differ from those used by the app, although even those aren’t particularly clear.

This new national service, which operates only from web and call centre interfaces, takes those with symptoms through a different diagnostic algorithm to determine whether they should be sent a test kit. If they do meet the criteria for testing, a kit is sent out which they use and return. Those who are positive on testing then have details of their contacts taken, and those are traced by one of many newly-trained contact tracers.

The criteria being used for national contact tracing are that the index case tests positive for Covid-19, and that each contact either:

  • remained within 2 metres of the index case for a period of 15 minutes or more, or
  • came into direct physical contact or “face to face” within 1 metre of the index case at any time,

during the period from 2 days before the onset of their symptoms, to 7 days afterwards, assuming that they make a full and prompt recovery.

In one of my earlier articles here, I used a simple example of someone who travels by public transport to work, there completes a normal shift, then returns home by public transport. During each journey, they come into close contact with two other people; at work they come close to a further two, giving a total of six close contacts in the day. The following morning, they feel rotten, and have the early symptoms of Covid-19.

Conventional contact tracing of the type being used by the national service is only going to be able to identify the two contacts made at work, and any that are made in the case’s household, such as their partner and children. Those are, in any case, the easy contacts to identify and put into isolation: colleagues at work should be warned the moment that the index case reports their illness to line management, and their household will be even quicker to know.

Activities which are most likely to result in high risk contacts are also some of those in which conventional contact tracing is least likely to be successful, such as:

  • journeys on public transport taking longer than 15 minutes,
  • indoor environments such as bars, restaurants, and social gatherings,

unless distancing is enforced effectively.

In my simple example, the index case could have infected four contacts who are inaccessible to conventional tracing methods. If only half of those catch Covid-19, that one case would have infected two others, giving an R value of 2. Scale that up over the population of a large city, and you can understand how important tracing ‘anonymous’ contacts becomes. Without the sort of invasive access to personal data which has been practised in South Korea, and without an effective smartphone app, it’s hard to see how the UK is going to control infection rates in the months before immunisation becomes generally available.

Surprisingly, much debate about contact tracing in the UK has been centred on the problems of mistaken identity, and the risk of malicious misidentification of contacts. These are actually the least of its problems, particularly given the criteria being used for high risk contacts, who almost inevitably are either people that we know and therefore readily traceable, or those who remain anonymous. Misidentification is not normally much of a problem when tracers have been properly trained.

Another significant problem with the new system which hasn’t elicited much debate is its reliance on positive test results before initiating tracing of contacts. As I’ve explained, time is of the essence. Typically, home testing adds a delay of 2-4 days, which is long enough for contacts to be spreading coronavirus to others, pushing up the R value further. This increases the efficiency of tracing, and reduces numbers who might needlessly be put into quarantine, but it ignores the fact that Covid-19 tests have high rates of false negatives, index cases who will incorrectly be assumed not to have the infection when they are highly infectious.

Like many countries, England is rushing to relax lockdown measures before it has even introduced, let alone evaluated, effective measures to detect and address subsequent increases in infection. If this trial of its smartphone app has been “a big success”, it begs the question as to why the app isn’t part of those measures when there’s been ample time and opportunity to enhance and improve it. I think the answer is that, for whatever reasons, the UK smartphone app as we have been trialling it is dead.

One final word on the phrase being widely used for contact tracing, track and trace. This is appropriate for parcels, where we want to know where they are and when they’re going to arrive. It is completely inappropriate for contact tracing among people, for whom where is largely irrelevant, but who is absolutely critical. Look the word track up in any English dictionary and you’ll see that it is all about location. In this context, obtaining location data about individuals is rightly a very sensitive issue. Misusing the phrase track and trace is an excellent way of confusing people into believing that contact tracing is a threat to their privacy. Properly conducted, it should never be.