Ministry of Health needs to be honest about contact tracing technology

Sunday, 27 September 2020

OPINION: Six months has now passed since that eerie Thursday in March when New Zealand entered its level 4 lockdown.

Since then, we have learnt we are dealing with a disease that is unlikely to be quickly eliminated overseas, where there are currently more than 7 million active infections.

We also now know, according to the World Health Organisation, that Covid-19 appears to kill about 0.5 per cent of the people it infects, taking into account asymptomatic cases.

Should we put cities into a level 3 alert each time there is a breach at the border, to deal with a threat of that nature?

**READ MORE:

* Coronavirus: Bluetooth technology could be in October's NZ COVID Tracer app update

* Sam Morgan gives up on CovidCard in frustration with Ministry of Health

* Coronavirus: Just one exposure notification has been sent from Covid Tracer app

**

Should we risk letting more people in and out of the country?

People will have their own views on the measures that might be justified to maintain the country’s elimination strategy in the face of the siege from Covid-19, and I’ll assume you don’t need mine.

But instead of trundling on, heads down, it must make sense to look up and take stock from time to time – even if it is just to decide we are happy with our current choices.

If we keep trying to sift the good information from the bad, those shouldn’t be impossible conversations to have, as most of our interests are pretty well aligned.

We can’t control other countries’ response to the virus, the development of a vaccine, or how Covid-19 mutates.

But one thing that is within our control is what level of sacrifice of convenience and privacy we demand from our fellow citizens to ensure outbreaks can be better contained through faster contact tracing.

Better contact tracing might allow us to carry on as we are but with less risk of clusters spreading out of control, or perhaps to make some slightly different decisions about lockdowns or the border.

A step-change in contact tracing will require new technology, but in my view the Ministry of Health is not being entirely honest with itself about what that might need to involve.

Its Contact Tracer app does provide the ministry with contact details for the 2.4 million people who have downloaded it, and acts as a useful “aide memoire” to help those people keep track of their movements.

But I’ve made the point before that the feature that lets people recieve automated alerts if they visit a venue at the same time as someone subsequently diagnosed with Covid will prove of little benefit.

That is because the feature only works if both the infected person and the people who were at the venue at the same time happened to both use the app.

The Covid Trader app is currently being used to scan into about 1.5 million venues a day.

The Ministry of Health doesn’t appear to have conducted any surveys to estimate how many venues people visit on average each day.

That means it can’t estimate how many million scans there would be if everyone was using the app religiously.

That’s unfortunate, because without that missing number, no one can really know how useful the feature actually is.

But say the country’s 4 million adults visit a total of 8 million venues each day.

That would imply people were “scanning in” about 19 per cent of the time.

Because of the exponential relationship between usage and efficacy, that means the app would only be capable of creating alerts for a meagre 3.5 per cent of contacts that might have occurred in an outbreak.

Say, we think people one visit just one venue a day on average.

That figure would still only rise to 12 per cent.

Ministry of Health deputy director-general Shayne Hunter doesn’t dispute that analysis.

“I hear what you are saying,” he responds.

“The comment I would make is we are very focused on making sure we have a good base level contact tracing technology and that we are able to contact the 80 per cent of people in 48 hours.”

The most important message the ministry is trying to get out to people, is to keep a record of where they go, he says.

“We have used the app, and the app has proven to be useful in the case of some of the contact tracing we have needed to do, but we are very much focused on optimising the contact tracing process itself.”

The ministry plans to add a Bluetooth proximity feature to the Covid Tracer app that could create similar automated alerts if people’s phones come within a certain distance of each other for a prescribed time and one of them is subsequently diagnosed with Covid.

At the same time it is continuing to press ahead with a further trial of the wearable CovidCard Bluetooth tracer token that has been developed in New Zealand.

Hedging bets in this way with multiple technology tools might give policy makers comfort, but it is not necessarily a good strategy.

Mathematics continues to dictate that if we are to deploy a technology tool to make a real difference to contact tracing, and hence to our wider pandemic options, its usage would need to be near universal to be very effective.

Take a situation where half of people used Covid Tracer QR codes for contact tracing, 40 per cent decided they would just use the Bluetooth feature instead, and the remainder were issued a CovidCard.

The upshot would be that 42 per cent of close contact events might be notifiable through those features – versus 100 per cent if any one of the three options was adopted universally and the others discarded.

“You can end up with too many solutions and the fragmentation may work against you,” Hunter acknowledges,” but begrudingly and with reservations I think.

“We don’t have too many solutions and therefore I don’t think there is a fragmentation working against us,” he goes on to say.

Overall, I’m not convinced the ministry has accepted the implication that the Government may face some tough, binary choices.

And the temptation in this ‘fog’ may be to delay making difficult calls.

Communications Minister Kris Faafoi’s statement – dutifully repeated by the Ministry of Health – that the CovidCard was shown to work in a “controlled environment” in a trial in Nelson and will now be tested in a “real world” scenario in Rotorua, suggests that’s the direction with the CovidCard at least.

The reality of the trials is pretty much the exact opposite.

That device was shown to work in a “real world” technology trial at Nelson Hospital.

Bluetooth signals do not have to obey any local bylaws set down by the local DHB when they travel through the air at Nelson Hospital.

That trial overseen by Otago University appears to have shown the Kiwi-made devices worked far more accurately than smartphone-based Bluetooth solutions developed overseas.

But the new Rotorua trial appears the opposite of “real world”.

Though the ministry has not yet released its terms of reference, that trial appears mainly designed to test how people would feel about voluntarily wearing CovidCards.

Given that there would be no significant public benefit to anyone in using any Bluetooth proximity device unless either its use was near universal nationwide or the virus was known to be circulating in their community, that is not something the Rotorua trial could realistically do a “real world” test of.

But if the ministry feels it is helpful for it to test user-acceptance of wearable tokens in such a trial, that raises the question of why it hasn’t put its own Bluetooth smartphone feature through the same process.

Of course, there have been things the ministry has done extremely well.

The early communications around numbered alert levels and “bubbles” was terrific.

But I fear that six months after level 4, it has become easier not to take stock, and to instead keep muddling on through.