Data: Ministry of Health https://www.health.govt.nz/system/files/documents/pages/covid-cases-30april2020.xlsx ; own calculations
Author: Bronwyn Howell, May 1 2020
Much has been made in New Zealand and internationally by politicians and media commentary of the effect of placing the country into unprecedented near-total social and economic lockdown on March 25 to combat the spread of the coronavirus responsible for Covid-19. Media messaging has portrayed the lockdown as a collective nationwide fight. Prime Minister Jacinda Ardern has phrased the lockdown and associated social distancing, hand-washing and surface-disinfecting activities as “being kind to each other” in the effort to stop widespread community-based transmission of the virus.
Overwhelmingly, the population of nearly 5 million strongly supports the government’s policy which has required the whole country remain at a single designated home location except for trips to purchase from a handful of “essential” businesses (mainly supermarkets, pharmacies and technology suppliers, but not butchers, bakers, hardware stores or “big box” retailers) and personal exercise (walking or cycling in one’s immediate vicinity). Even online retail sales have been restricted to a list of “essential” items.
The economic effect, however, has been devastating, with over 60 percent of the workforce reliant upon a wage subsidy scheme. The tourism industry, responsible for 21 percent of the country’s foreign exchange earnings, has totally collapsed in the face of an effective travel ban arising from stringent new visa conditions.
Yet the curve of new and probable cases has been well and truly “flattened” (Figure 1). The Director-General of Health has confirmed that the virus had been “eliminated” (i.e. not zero cases but health officials know where all new cases come from). The Prime Minister’s statement that the country has “stopped the widespread community transmission” of the virus has interpreted by some to mean that New Zealand has won the battle with Covid-19. However, while thanking the country as the stringent lockdown provisions were loosened slightly on April 29, she warned that the fight is not yet over. New Zealanders are still largely confined to their homes for at least another two weeks, but workers have returned (with strictly-enforced social distancing) to a number of industries (notably forestry and construction). Retail, however (with the exception of limited provision of take-away coffee and foods ordered remotely and collected contactlessly), remains closed.
One might conclude from officials’ statements, media hype and the “evidence” of Figure 1 that New Zealand’s stringent lockdown policy has been instrumental in bringing about the current state of affairs. When the lockdown was implemented on March 25, the exponential curve of cumulative cases had just begun its upward surge and does not appear to begin tapering off until around April 5. The point of negligible new cases is reached around April 15. By the time the lockdown relaxation was announced on April 20, it was essentially flat. When the relaxation was implemented eight days later, there had been negligible change.
But is it plausible to credit the lockdown with “flattening the curve”? The lockdown was only one of a range of policy interventions deployed, albeit the most economically and socially costly. Earlier interventions included:
testing of suspicious symptomatic individuals beginning on January 21, with tracing of contacts for all positive (confirmed) or likely (probable) test subjects;
active case management of confirmed or probable cases instigated from early February (varied across District Health Board areas);
closing the borders to travelers originating in or travelling through China (the original epicenter of the virus) on February 3;
quarantining New Zealand citizens evacuated from Wuhan on February 5 for fourteen days;
self-isolation for fourteen days requested for all travelers returning to New Zealand after March 4, regardless of whether or not they were presenting symptoms; and
self-isolation for at least 14 days made a requirement for issuing visas to foreign travelers from March 18.
While the resources for contact tracing and case management were initially stretched, and somewhat fragmented across the country, when case numbers started increasing in mid-March, additional resources were deployed. National co-ordination was instigated from March 24, initially with manual processes and, from April 6, using a computerized system.
With such a large range of policies interacting, it is difficult to isolate the effects of any specific one in order to credit it with changing the trajectory of the contagion curve. However, It may be possible, with hindsight, to identify which additional interventions could not have plausibly influenced it, given the characteristics and infection patterns of the virus and underlying population exposure.
At this point, it is apposite to note that an individual exposed to the novel coronavirus responsible for Covid-19 takes several days (generally at least five to ten) to develop the infection. If no symptoms develop after fourteen days after contact with an infectious individual or contaminated surface, then it is unlikely that an exposed individual will go on to develop the illness. This reasoning underpins the fourteen-day isolation period in the above policies.
It is also noted that the generally-agreed optimal strategy for managing an epidemic is dependent upon the proportion of the population infected. If only a small proportion of the population has contracted or is suspected of having contracted the virus, then isolation of those individuals away from healthy individuals is the most effective strategy. This is the strategy adopted in countries such as Singapore, Taiwan and South Korea. It is only when the underlying level of contagion in the population is high that it is optimal to quarantine healthy individuals to prevent them from becoming ill. This strategy was adopted in densely-populated cities in Northern Italy, and in New York, when it became clear that a large number of people were infected. To understand the underlying level of infection in the population, the World Health Organization head has urged countries to “test, test, test”: sound policy-making requires this information.
Returning to New Zealand’s strategy to impose isolation on its entire healthy population. If this strategy was to have an effect on the contagion rate, then given the five-to ten day incubation period for the virus, the effect of a lockdown in curbing spread of the infection would not be evident in the statistics for at least five days following implementation. The number of new cases identified would be expected to increase day on day until at least a window between March 30 and April 4, as those who developed the infection after the March 25 lockdown would have been exposed to contamination prior to that date.
Figure 2 shows that the “flattening” of New Zealand’s contagion curve occurred between the 21st and 25th of March. The highest number of new cases (83 – counting both confirmed and probable cases) was recorded on March 25th, the day the lockdown began. This suggests it is most unlikely that the lockdown had any material effect on curbing New Zealand’s contagion rate. The flattening of the curve up until April 4 was most likely driven by the variety of other policies in place prior to the lockdown commencing.
The effect is even more pronounced when separating out “confirmed” and “probable” cases. “Confirmed” cases are those returning a positive genetic (PCR or molecular) assay test. This test has a sensitivity of only around 45 to 70 percent, meaning of infected persons, between 30% and 55% will return a negative result on the first test. As only symptomatic individuals were tested in New Zealand, at least up until mid-April, clinicians made judgements as to whether those returning negative tests actually had the illness. These are recorded as “probable” cases, and were managed as if they were positive cases. The number of positive tests per day (confirmed cases) begins its decline after the March 25 peak. The total cases hovers at a high plateau until April 5 largely because of the “probable” cases – determined from clinical judgement rather than test results. As the number of tests undertaken was increasing significantly over this time, but the actual number of positives was falling, It is quite likely that the “probable” statistic reflects a considerable erring on the side of caution amongst clinicians who classified a individuals presenting with other respiratory illnesses as Covid-19 cases. The total cases per day falls back to be driven mainly by confirmed cases only around April 8. “Confirmed” cases only are recorded in the international statistics (e.g. those prepared by Johns Hopkins University) to address this possible human bias.
New Zealand is a small, isolated island country far removed from the rest of the world. It is comparatively easy to contain new infection risks from abroad at the border, provided steps are taken quickly enough. This principle governs the management of much of New Zealand’s primary industry risk, as befits a country making most of its overseas income from this sector. Occasionally, things slip in (e.g. mycoplasma bovis in the cattle industry and Psa-V in the kiwifruit industry), but for the most part, prudent border management suffices. In this case, it appears that moving early to require self-isolation of individuals arriving from overseas, combined with imperfect, but sufficiently effective contract tracing and subsequent case management) has been the main reason why the curve flattens around March 25. The last self-isolation provision was implemented on March 18, consistent with the curve flattening and beginning its descent on March 25.
In sum, therefore, there was no widespread level of infection in the New Zealand population when the lockdown was implemented. Healthy people were put into quarantine, but the risk of an individual encountering the infection when “at large” was negligible. This is easy to observe in hindsight, but could also have been ascertained by undertaking random testing amongst the population in addition to testing symptomatic people, especially around early to mid-March, before a large number of cases had emerged and in conjunction with the obligations placed on travelers coming into the country to self-isolate. It would seem to be largely the efforts of these early self-isolators and the herculean efforts of public health officials identifying the contacts around those individuals who may not have been as vigilant in their isolation efforts that appear to have been responsible for New Zealand’s low infection rate, and early (apparent) effective elimination of the virus, and not the nationwide lockdown.
Data: Ministry of Health https://www.health.govt.nz/system/files/documents/pages/covid-cases-30april2020.xlsx ; own calculations
Data: Ministry of Health https://www.health.govt.nz/system/files/documents/pages/covid-cases-30april2020.xlsx ; own calculations