COVID-19 and Dot2Dot

We would like to emphasise that we’re cautious about the use of Dot2Dot in the COVID-19 epidemic. Dot2Dot has been designed primarily as a tool to help with manual contact tracing for TB, whereas a lot of the COVID 19 discussion is about population-level tracing.

Having said that, we’re keen to help if we can!

Contact Tracing at Scale

With COVID 19, the world is experiencing the transmission of an infectious disease on a vast scale –billions of people may eventually be infected. In this scenario, the conventional view was that contact tracing isn’t useful – because there are just too many people to be traced.

But this view has been challenged by China and other East Asian countries. China deployed a combination of strong state authority, huge investment in contact tracing, proactive testing and innovative technology to successfully contain the epidemic in Wuhan – much to the surprise of epidemiologists.

According to WHO, the main tool used by Asia was testing – if a person had a suspected infection, they tested quickly and told the patient whether they had the illness. The benefit of this approach is a huge improvement in compliance – people just don’t stay isolated if they don’t think they have the illness. But China also deployed huge resources very quickly: they had 1800 teams of 5 people doing contact tracing in Wuhan.

Technological Approaches

There is a lot of discussion about technology in COVID-19 Contact Tracing. In theory, if a whole population could be tracked using GPS, it would be simple to ask the computer to find all people who had been within 2 metres of an infected person. In practice, the data volumes are vast, and it’s not easy to get a whole population to install a tracking app.

China has a strong head-start in this area – in many cities, a sizeable proportion of the population use the WeChat or AliPay apps. There are stories that these apps already have GPS location tracking included, or that this was enabled during the epidemic – and it’s quite possible this was used by China in contact tracing. By comparison, trying to do the same thing in a Western country would trigger a significant privacy debate and it would be hard to achieve a reasonable percentage of the population.

Other options are discussed: the mobile phone network already tracks the phone mast we’re nearest too, and this data could be used – although it’s not as precise as GPS data. Some countries are reportedly passing laws to harvest bank transaction data – because this can tell us where someone ate or bought petrol, giving an implied location.

At present, it’s not clear how practical these approaches will be. Our understanding is that China’s plan is to double-down on these approaches and hope to return to normal life – but with huge vigilance and rapid response to any new outbreaks. By comparison, it seems that the same approach is considered unlikely to work in the West.

Contact Tracing Scenarios

We see three levels or scenarios for contact tracing:

Level 1

Small scale / manual contact tracing, to track an outbreak of e.g. TB within the UK, or perhaps a new COVID-19 outbreak within a small defined geographical area. Augmented with location tracks for selected contacts.

Level 2

Large scale manual contact tracing, with import of location tracks for any infected people from other data sources.  Potentially thousands of users working on a single outbreak.

Level 3

Population level automated contact tracing – harvest GPS data on every individual and use it to directly find who was in proximity of a contact – less need for manual tracing if it really worked.

Dot2Dot is aimed mostly at L1 but may be scalable to L2. Level 3 is completely different, requires vast state resources and may even then be impractical.

At L1 and L2, contact tracing is still substantially a manual effort. We identify a person who has the illness, and we may obtain a location trace of that person from Mobile Phone, Bank or Google data. But then we still need to do a lot of legwork to find the people who might have met the contact.

There is also a halfway house to L3: a state publishes an App that tracks a users’ location and tells that user when they have been near an infected person by checking a central database. The obligation is that the user should then report to the authorities, get tested and – if positive – hand over their location trace to the central database. This approach might overcome privacy concerns, but we think getting widespread adoption might be a challenge.

Conclusions

Dot2Dot was inspired by a contact tracing failure that occurred years before COVID 19. Even after the pandemic is over, there will still be small scale outbreaks occurring all over the world regularly – and these will not have the benefit of huge state intervention to resolve.

Dot2Dot is aimed at saving lives in the long term – if it can be useful in the current crisis, we’ll be delighted – but really, we’re looking beyond COVID-19.

In particular, there has been a huge drive recently to bring down infection rates in TB due to the high mortality rates and financial burden it imposes. We are hoping that Dot2Dot can address this issue.