Research contracted by the aviation sector models the effect of selectively relaxing our border restrictions depending
on each country’s Covid-19 risk level.
It is a model of how we could manage international arrivals by their country of origin, and suggests we could offer
quarantine-free travel or shorter stays in quarantine to arrivals from some less Covid-afflicted areas.
The study is funded by Auckland Airport, Wellington Airport, Christchurch Airport and Air New Zealand, and one of the
study’s authors is an Air New Zealand employee.
The SMC asked experts to comment on the research.
Professor Michael Plank, Te Pūnaha Matatini and University of Canterbury, comments:
“This study uses a mathematical model to evaluate the risk of international arrivals entering the New Zealand community
while infectious with COVID-19. The authors use this model to assess the risk of selectively relaxing border
restrictions for arrivals from low-risk countries, with more stringent requirements for people from high-risk countries.
They predict that this approach could allow five times more people to enter New Zealand than at present, while
increasing the risk to the New Zealand community by 20-50%.
“This might not sound like a big increase in risk, but it means that breaches like the one that occurred at the Pullman
Hotel last week will occur 20-50% more frequently. This increases the chance of a community outbreak and the possibility
that an Alert Level change would be needed to contain it.
“The authors of the study claim that the recent requirement for a pre-departure test will mitigate this risk. However,
this is far from clear because pre-departure tests are not perfect and many travellers were already required to take
pre-departure tests by their airline or country of transit.
“There are also several other factors that would need to be considered in any border relaxation plan. The number of
arrivals into New Zealand is limited by the capacity of our MIQ system. As recent events have shown, we need to make
absolutely sure the facilities used for MIQ are fit-for-purpose and do not contribute to the spread of the virus. The
majority of our border breaches have occurred when a border worker got infected by a recent arrival. If the number of
infected arrivals increases, this will also increase the risk of future infections in border workers.
“It makes sense to have a risk-based border system that is based on the current rate of COVID-19 in different countries.
We will need a framework of this type to relax border restrictions once the world begins to emerge from the pandemic.
However, COVID-19 is more prevalent now that at almost any point in the past. At the moment, we need to do everything we
can to reduce the risk of importing COVID-19 into the community, not taking on additional risk.”
Conflict of interest statement: I am partly funded by MBIE for research on mathematical modelling of COVID-19.
Professor Shaun Hendy, University of Auckland, comments:
“This paper looks at the risks of a more targeted border control policy that would allow for travellers from low-risk
countries to enter New Zealand without passing through the full MIQ process. Indeed, the prospect of travel bubbles with
a country such Australia, which has followed New Zealand’s elimination strategy, has been discussed for some time. The
analysis presented in this paper combines aspects of Te Pūnaha Matatini’s border and MIQ risk models with estimates of
prevalence in overseas countries based on each country’s reported COVID-19 fatality rate.
“Broadly speaking, the methodology described in the paper is sound, although the practicality of using reported
fatalities in order to categorise traveller risk is questionable. As the authors acknowledge, there is a lag between
infection and death of around three weeks. Looking back to March 2020, if we had been using the system proposed by the
authors, the lag between infection and death would probably have prevented us from escalating border controls fast
enough to prevent the Alert Level 4 lockdown. Any approach to assessing risk from travellers, therefore, would need to
be based on much more responsive indicators. Thus, as an exercise, the study is useful, but I don’t believe it presents
us with a workable scheme for managing travel bubbles.
“Furthermore, since the paper was accepted, the global situation has changed considerably. As the authors acknowledge in
an addendum, risks have increased considerably as global prevalence has risen and new strains have emerged. With this
change in global risk, it is far from clear that even our existing border settings are sufficient to reduce the chances
of a substantial outbreak to acceptable levels. Even at current MIQ capacity, we are still seeing infections occurring
regularly within facilities in both workers and travellers, which puts us at risk of another Auckland August type
outbreak. This scheme would increase the frequency of MIQ failures, the chances of a larger outbreak and the chances of
another regional lockdown by something like 20-50%.
“As the global vaccination drive kicks into gear and the pandemic becomes better managed, risk frameworks like this will
be needed. This particular framework is impractical, however, and the specific conclusions drawn have been overtaken by
global events.”
Conflict of interest statement: “Shaun Hendy leads Te Pūnaha Matatini’s COVID-19 modelling programme, which receives
funding from MBIE’s Covid Innovation Fund and the Tertiary Education Commission.”
Dr David Welch, Senior Lecturer, Centre for Computational Evolution and School of Computer Science, University of
Auckland, comments:
“Many individuals and organisations have called for the number of travellers entering New Zealand from different
countries to be based on the Covid risk in those countries. These calls have come from those advocating for greater
freedom to travel here from low-risk countries, as well as from advocates for increased restrictions for travellers from
high-risk countries.
“At its core, this paper creates a model to categorise countries by level of risk and quantify the risk of causing a
community outbreak in New Zealand that is posed by a traveller from a given country. The calculation is fairly
straightforward and is based on reported COVID-19 prevalence in countries (which is then inflated if mortality figures
suggest case numbers are under-reported). There are a number of assumptions that need to be made (e.g., how likely are
travellers to be infected in transit or in MIQ) and these are inherently hard to check, so a range of plausible values
are considered. Given the huge disparities in countries’ prevalence, even very conservative assumptions mean the
assessed risk of travellers from different countries will differ greatly. So, without having seen the details that
appear in the appendices, it is a reasonable model.
“The difficult question is what to do with such a model. The authors advocate the policy you would expect given their
funders (airports and Air New Zealand) that travel should be opened up to more low-risk travellers. But they estimate,
based on August 2020 figures which have since dramatically changed, that this would result in roughly 25% more cases in
the community than we have so far had. The model could equally be used to argue that the numbers from high-risk
countries be reduced to lower overall risk.
“Finally, the model does not include vaccination, which will become an increasingly important part of how and to whom
the border is opened up.”
No conflict of interest.