how te pūnaha matatini became part of our covid-19 response
28 october 2020
the faces of shaun hendy and siouxsie wiles have become very familiar in 2020. but they're just two of a multi-disciplinary team of researchers at te pūnaha matatini, who have contributed to new zealand's covid-19 response. jonathan burgess meets some of the team.
the written proposal to set up te pūnaha matatini (tpm) has an eerie prescience to it. the national centre of research excellence, hosted by the university of auckland, is a transdisciplinary centre whose name means ‘the place where many faces meet’. it was founded to apply complexity science ‘to the critical issues of our time’, with a focus on communication and connection to government and the private sector, and launched in 2015.
“you could have added ‘and then we’ll tackle covid-19’ to the end of the first paragraph,” jokes tpm’s director, professor shaun hendy. shaun and his colleague, associate professor siouxsie wiles, have become household names thanks to their science communication in the media during the covid-19 pandemic.
but they haven’t been working alone. around two dozen tpm researchers from this university and around new zealand have worked across multiple areas in response to covid-19. that includes phd students who suspended their studies to be involved in the project and postgraduate students whose employment had been put on hold.
their models predicting the spread of the virus contributed to the swift initial lockdown in march and informed the alert levels used in response to the outbreak in august. they have also contributed to modelling hospital capacity, genome sequencing and tracking the spread of disinformation.
“this has felt very, very much like living in history and having our team be a really key part of it,” says tpm executive director kate hannah.
tpm’s involvement with covid-19 began early in the year.
“siouxsie wiles is one of our superstars, and she’d been talking about covid-19 in the media since january,” explains shaun.
siouxsie and professor michael baker of the university of otago had been fronting the government’s communication about the pandemic, and shaun wanted to ensure they had good information. he put siouxsie in touch with the prime minister’s chief science advisor, professor juliet gerrard (faculty of science), and was pulled into the middle of the whole response. listening to siouxsie talk about the epidemic, shaun started to wonder whether tpm’s annual hui planned for april would go ahead.
“i put some data from italy into a basic epidemic model,” says shaun. “i went ‘okay … we’re not going to be holding our hui in april’.”
it was this one moment where you bring all your knowledge and expertise to bear.
tpm’s catchphrase of “data, knowledge, insight” couldn’t have been more relevant.
“at that point i realised this was going to be something really serious. what i’d calculated – if it was anywhere near right – was an important thing to start communicating.”
peter-lucas jones from northern iwi te aupōuri was present at the tpm board meeting at which shaun presented his modelling and communicated the decision for tpm to focus its work on covid-19.
“peter-lucas told the story of te aupōuri’s experience in the 1918 influenza pandemic and the mass graves,” remembers kate. “it was really profound to be reminded of it. we came away from hearing that, knowing this was the only thing we needed to be working on.”
statistician andrew sporle (ngāti apa, rangitāne, te rarawa) was brought in to co-lead work focusing on at-risk communities.
“this research directly informed the iwi-led pandemic response that kept covid-19 out of at-risk communities during the first outbreak,” says kate. “because as soon as shaun said we could have tens of thousands of people die in new zealand, i knew it would be vastly more impactful on māori and pacific peoples.”
this prediction also had big implications for the healthcare system. associate professor in statistics ilze ziedins worked with dr mike o’sullivan and associate professor cameron walker from the faculty of engineering to model the effect on hospitals if the virus spread through new zealand. “it was this one moment where you bring all your knowledge and expertise to bear,” ilze says.
“we looked at what the loads would be on intensive care units and in wards.”
ilze is thankful that we haven’t reached the point where their work has been needed. intensive care specialists advised the team that the surge capacity that they had modelled was not sustainable for extended periods.
during lockdown, questions were arriving from the government with one-hour deadlines.
“the first question was around breaking things into waves,” says shaun.
“there had been some work done in the united kingdom about mitigation versus suppression. suppression is basically going hard to try to contain things. mitigation is going just hard enough to keep things within your hospital capacity.
“we got asked if we could produce the new zealand version of that in an hour. it turns out we could, with about a minute to spare.”
in order to respond to more nuanced questions, a team led by dr dion o’neale, a lecturer in physics in the faculty of science, has built a statistical network of everyone in new zealand, linking people by dwelling, workplace and school, and with attributes such as age, ethnicity and sex. their data was drawn mainly from the integrated data infrastructure (idi) research database created by stats nz.
“individuals are different,” says dr emily harvey, an honorary academic in physics who is also part of tpm. “they have quite different ways that they’re getting exposed to things, and different people they’d interact with.”
dion, emily, dr oliver maclaren and steven turnbull have used the computing power of the new zealand escience infrastructure (nesi) to run contagion processes on their
“Network-based models help you answer questions that you can’t address with simple models,” says Oliver, a lecturer in engineering science and the James & Hazel Lord Emerging Faculty Fellow.
“for rapid-response stuff, simple models are very useful, but more complex models allow you to answer specific policy questions, like what if we shut down ponsonby?”
more complex models allow you to answer specif ic policy questions, like what if we shut down ponsonby?
the call that came on tuesday 11 august was the one they had been standing by for: a positive case with no known link to the border had been found. tpm was back on the job by 6pm, with associate professor alex james of the university of canterbury getting the first modelling results back to officials by 7.30pm. alex’s initial estimates of the size of the outbreak were used directly in the cabinet meeting that evening.
“we were prepared for this,” says shaun. “we had recently held a workshop in wellington between the key government agencies in the covid-19 response and our modelling teams. there was a period where both alex and i were hoping it was a drill, because this was one of the worst-case scenarios that we’d considered.”
tpm’s work had previously predicted a much higher infection fatality rate for māori and pacific peoples, and a new infection in south auckland was cause for concern.
alongside getting prompt projections to the government, their more complex models were quickly fired up again, and work began on genomic sequencing to see whether community transmission was from the same or different strains of the virus.
shaun says tpm’s involvement in the covid-19 response demonstrated the value of having a rapid-response team of scientists ready to go at short notice. as a new zealand-wide centre, it was equipped to handle a complex and fast-moving national challenge.
“on the one hand, you can tell the sort of forrest gump-like story of how we just had the right conversations at the right time and then found ourselves in this position,” reflects shaun.
“on the other hand, we’ve been building a national community at te pūnaha matatini to do exactly this and we’ve put emphasis on communicating our work and working with policymakers.”
shaun says the covid-19 crisis has shown the government and its agencies need to have access to trusted expertise.
“some of us will be working on covid-19 until march 2021 and beyond, so we’ll keep the models active and going.
“i’m hoping that’ll be in perpetuity so that if in the next 20 years we have another pandemic, there is a set of new zealand-specific modelling tools. we didn’t have that when we started this time round.”
this article first appeared in the spring 2020 edition of ingenio magazine.