View Article PDF

Historically, shipping has been involved in the global spread of pandemics, and maritime quarantine has been used as a successful control measure (eg, in the 1918 influenza pandemic).1 Maritime quarantine was even used successfully to prevent the arrival of the 2009 influenza pandemic in some island jurisdictions, such as Tokelau.2

The COVID-19 pandemic has also had an impact on maritime vessels during 2020, as well as spread to people on shore. On the Diamond Princess, 19% of the passengers and crew became positive with the pandemic virus (SARS-CoV-2) and there was spread to Japanese responders on shore.3 Similarly, on the Grand Princess, 17% of those tested had positive results.3 On a much smaller cruise ship with 217 passengers and crew onboard, 59% were reported to be test-positive.4 On a fishing vessel, 85% (104/122) of the crew were infected.5 In terms of merchant vessels, an outbreak on a container ship was reported as infecting 23% (5/22) of the crew.6 Other such outbreaks have been detailed in media reporting (referred to in a review7).

In response to the COVID-19 pandemic, border controls have been widely used to limit pandemic spread. Such border controls are particularly relevant for two types of strategy for controlling pandemics: (1) the exclusion strategy, as successfully practiced by some Pacific Island nations (eg, Tonga and the Cook Islands),8 and (2) the elimination strategy, as used by New Zealand9 and other jurisdictions (eg, Mainland China, Taiwan, and Australia).

Some of these jurisdictions have completely prohibited maritime vessels arriving at their seaports from countries that are not COVID-19-free (eg, the Marshall Islands have prohibited such incoming ships8). But time periods are also used to lower risk. For example, a minimum of 14 days at sea before being allowed to enter the Marshall Islands,8 or 14 days plus a negative PCR test to enter New Zealand.10 There is also the standard international requirement for pratique whereby any ‘illness during the voyage’ must be notified to health authorities at the destination port.11 Collectively, these control measures seem to be working fairly well, although in October 2020 New Zealand reported that a ship maintenance worker became infected with SARS-CoV-2 after spending time working on a ship that had recently arrived in the country.12 This worker then infected other workers and a household contact onshore (but with no further known subsequent spread). Genome sequencing has indicated that the source of infection was shipping crew flying into New Zealand to join their ship.13 Also in October 2020, another island nation (Australia) faced outbreaks on two cargo ships in a port in Western Australia, where (in one case) two onboard workers left a ship before the outbreak was detected.14

Given this background, we aimed to expand on previous modelling work for air transport spread of COVID-19,15 to determine the risk of merchant ships being the source of COVID-19 outbreaks in an otherwise COVID-19-free country: New Zealand.

Methods

Model design and parameters for SARS-CoV-2 and COVID-19

We used a stochastic SEIR type model with key compartments for susceptible [S], exposed [E], infected [I] and recovered/removed [R]. The model is a stochastic version of CovidSIM, which was developed specifically for COVID-19 (http://covidsim.eu; version 1.1). Work has been produced from previous versions of this model,15–17 and in two places we detail the relevant equations and their stochastic treatment.18,19 The model was built in Pascal, and the computer code is available on request from the authors.

We ran 100 million simulations for each set of parameter values. Such a large number of simulations was necessary due to the very high probability of zero infected crew members boarding a departing merchant ship, given the low assumed incidence of infection (see Table 1). The overall framework for the processes modelled is shown in Figure 1. The parameters were based on available publications and best estimates used in the published modelling work on COVID-19 (as known to us on 27 August 2020). We assumed that 71% of infected COVID-19 cases develop clearly detectable symptoms (Table 1). Another assumption was the contagiousness in terms of the effective reproduction number (Reff), which was 3.0 among crew members on board the ship and 2.5 in the destination country (Table 1).

Figure 1: Flow diagram of the assumed movements of merchant ship crews in the model including interventions (simplified and not showing all control measures (eg, the seeking of medical attention when symptomatic in the destination country and the associated isolation of identified cases and contact tracing as detailed in Table 2)).  

Shore leave in the destination country

We selected New Zealand as a case study destination country, as it has previously achieved elimination of community transmission of SARS-CoV-29 and appears to have successfully controlled subsequent cross-border incursions of the pandemic virus. Upon arrival of ships in New Zealand, we used a period of shore leave by all the crews of one day (the median time ships are in port, based on Ports of Auckland data, the port in New Zealand’s largest city).

Potential control measures

Potential control measures are detailed in Table 2 and Figure 1 and include a PCR test on all the crews on arrival and mask use by the crews during shore leave. If any crew member tested positive, then the shore leave for that particular crew was assumed to be prohibited and therefore no risk of any community outbreak from shore leave was assumed. If a crew member on shore leave developed and self-reported symptoms and then tested positive, this case would be isolated, and this could also trigger contact tracing, which was assumed to identify 80% of the infected contacts within 48 hours. Identified contacts would be isolated after a delay of one or two days.

Ongoing infection transmission in the destination country

Untraced secondary cases who were infected by crew members in the destination country, and tertiary cases who were infected by traced secondary cases before they were isolated, were assumed to roam freely for the full length of their infectious period and to potentially trigger outbreaks in the community.

Table 1: Input parameters used for modelling the potential spread of SARS-CoV-2 infections via merchant shipping with the stochastic version of CovidSIM (v1.1).

Control measures assumptions

The full details on the considered control measures are given in Table 2.

Table 2: Control measures and their estimated efficacy.

Results

The results suggest that, if no pandemic-related maritime controls were in place, the COVID-19-free destination country (New Zealand) would quickly experience an outbreak because of the arrival of ships with infected crew members taking shore leave. That is an outbreak after a median duration of 0.064 years (23 days), which is equivalent to a total of 355 port visits and 7,100 total days of shore leave (for international vessels with 20 crew members and one day of shore leave per person per port; Table 3). However, there is high uncertainty, with 95% of outbreaks likely to occur between 1 to 124 days (ie, 0.0023 to 0.34 years; Table 3).

The median time to an outbreak would increase markedly by obligatory PCR testing of crew members before shore leave is permitted (ie, up to 168 days (0.46 years), or after a total of 2,592 port visits). An even further reduction of risk would occur when requiring face mask use during shore leave (increased median time to 1.00 years). But, relatively little extra gain in risk reduction would result from any sick crew on shore leave self-reporting symptoms and the associated contact tracing (Table 3). Using the base case value of Reff=2.5 in New Zealand, a single untraced infection in the community leads to an outbreak in 88.2% of cases (78.5% for Reff=2.0). When we considered super-spreading events in the community in a scenario analysis, the outbreak probability per person was actually reduced to 57.4%. This is because allowing for super-spreading events means that a smaller proportion of infected crew members transmit infection, even though those that do will typically infect more people (assuming the same overall value of Reff).

In scenario analyses, a smaller crew size reduced the outbreak risk (eg, the median time to an outbreak would be 3.8 years for ships with a crew size of five; Table 4). The risk of outbreaks was also lower when making assumptions around lower contagiousness in the destination country (ie, Reff lowered to 2.0). The risk remained basically unchanged when contagiousness on the ship was assumed to be higher (ie, Reff increased to 4.0). Increasing the shore leave to either two or three days increased the risk of an outbreak (ie, it reduced the median time to this event). When super-spreading events were considered in the destination country, this led to the same average number of untraced infections caused by crew members in New Zealand, but as each one of them had a lower risk of leading to an outbreak, the overall outbreak risk was lower than in the baseline study.

Figure 2 shows that voyage duration is a key determinant of outbreak risk in the destination country, and this risk is especially high once voyages are longer than five or so days (ie, once the latent period is typically over and crews have become infectious). Onboard spread can maintain this risk over subsequent weeks, leading to more infected individuals on board; but this also increases the detection probability by testing on arrival in New Zealand. It can take a long time for the onboard epidemic to come close to ‘burning out’. Indeed, the outbreak risk in the destination country when there are no controls only starts to decline after a voyage time of three weeks, and even then it declines quite slowly (Figure 2). For larger crew sizes of 10 to 20 people, the risk of community outbreaks is still increasing slightly after three to four weeks of voyaging when no controls are used (Figure 3 and Figure 4). Interestingly, if PCR tests are implemented, the effect of longer travel durations generates results that are the inverse: the more the infection can spread on board, the more likely it will be detected. As none of the crew members were assumed to be allowed to go to shore if any one of them tested positive, the probability that infected people being allowed to enter New Zealand decreases as the number of infected people on board the ship increases. Adding additional interventions like wearing masks, self-reporting symptoms and contact tracing further improves the results; but the main effect is obtained by PCR testing prior to shore leave being permitted. With the full set of interventions, the median time to an outbreak increased, but this time varied widely by length of voyage and size of the crew (Figure 3 and Figure 4).

Table 3: Results of the simulations without interventions and with multi-layered interventions (with these being for a base case of 10 days at sea and 108 merchant ship visits per week, 20 crew per ship, one day of shore leave each per port visit in New Zealand, and with 100 million stochastic simulations being run for each set of parameters).

Table 4: Results of the scenario analyses for 108 merchant ship visits per week and the full set of interventions taking place (see last line of Table 3) with 100 million stochastic simulations run for each set of parameters. (For further information, see text and Table 2.)

Figure 2: For ships with five-member crews, the median duration (log-scale in years) until a COVID-19 outbreak occurs in the destination country because of merchant ship crews taking shore leave. We assumed there were 108 cargo ships arrive each week. In the country of origin, each member can become infected at a rate of 0.00038 per day. Infections spread on board with an effective reproduction number Reff of 3.0 and in New Zealand with Reff of 2.5. Note that a voyage duration of 1 day is not applicable to New Zealand. Full black curves: no interventions are taken; full grey curves: all crew members are prevented from entering the country if one of them is PCR positive upon arrival; dotted grey curves: full set of interventions as outlined in Table 3. For each combination of crew size and voyage duration, 100 million voyages were simulated.

Figure 3: For ships with 10-member crews, the median duration (log-scale in years) until a COVID-19 pandemic outbreak occurs in the destination country because of merchant ship crews taking shore leave (other details as per Figure 2).

Figure 4: For ships with 20-member crews, the median duration (log-scale in years) until a COVID-19 pandemic outbreak occurs in the destination country because of merchant ship crews taking shore leave (other details as per Figure 2).

Discussion

Main findings

In this modelling work, we found that it might only be a matter of a few weeks before crew from international trading maritime vessels would trigger COVID-19 pandemic outbreaks in the destination country, if no control measures were in place. Of particular note is that even small five-person crews appear to contribute a risk after voyages of several weeks, and this risk only declines slowly thereafter. Fortunately, however, the risk of such outbreaks can be substantially reduced with the available interventions. In particular, PCR testing before leaving the vessel appears to be a valuable intervention, though this benefit still comes with high uncertainty as indicated by the wide range for 95% of the simulation results (shown in Table 3).

The results for our case study country (New Zealand) are likely generalisable to most countries that have seaports and maritime trade. Nevertheless, the risk could be somewhat less for some nations on a per population or per gross domestic product (GDP) basis because New Zealand’s economy is particularly trade orientated and especially dependent on sea trade. That is, it has no international trade by land routes and only a small proportion of trade volume is by air cargo. With a population of five million, New Zealand has 1,120 port visits from vessels with an international origin per million population per year.

Study strengths and limitations

This is the first study (to our knowledge) to explore the risk of COVID-19 outbreaks arising from shore leave of maritime ship crews. Another strength is that the work builds on an established model that has been used to also study air transport and other aspects of SARS-CoV-2 transmission (see Methods).

But, as with all modelling, there are important limitations. Some of these relate to parameters. A particularly critical one is the daily incidence of SARS-CoV-2 infection in the source country that the ship leaves from. We used a global average for this incidence to account for the diverse maritime trading patterns that New Zealand has and also because the crews are internationally diverse (often flying in from another country just prior to the ship’s departure, which may expose them to higher risks via air terminals and on aircraft). Yet there are likely to be highly variable risks of infection between different source countries that the ship leaves from and countries that the crew come from, and these will change with the evolving global pandemic of COVID-19.

Other examples of parameter limitations are the Reff onboard such vessels and the Reff for shore leave by crew. The former is likely to vary by different designs of merchant vessels (container ships vs tankers vs bulk carriers etc) and also by size (eg, it is likely that, in vessels of under 3,000 gross tonnage, the crew are in shared sleeping rooms). However, we did not have sufficient data to model such heterogeneity. We also didn’t account for prior immunity among crew members from past exposure to the SARS-CoV-2 virus, which is likely to increase over time with global progression of the pandemic. Given the data limitations, we did not consider port calls and shore leave on route between the original departure point and the first New Zealand port of call. Such port calls may either increase the risk for New Zealand (if the visited port city on route has a higher incidence of infection than the origin country), or they may decrease the risk (by extending the time length of the voyage, if the origin country had a higher incidence of infection than the visited city). We also did not model risk of transmission to port workers who might go onto arriving ships (eg, pilots and health workers conducting PCR tests on board vessels), given the assumption that they would take appropriate precautions with physical distancing and use of personal protective equipment. Yet people don’t always follow rules and accidental events may reduce the effectiveness of preventive measures.

Possible implications for future research and policy

Future research is needed to replicate this study (eg, using simulation models with a different structure and for a wider range of destination countries). The routine collection of international shipping transponder data, which is currently underway by other New Zealand-based researchers (funded by the Ministry of Business, Innovation and Employment), may also more precisely identify ship movements, travel times and also unusual events (such as ships exchanging supplies or crew at sea). Research could also explore the acceptability of, and adherence to, mask use by crews on shore leave in different settings.

As detailed above, the results in Table 2 and Table 3 might make some health authorities decide that the risk of allowing shore leave for crew is tolerable with control interventions such as PCR testing and mask use. But for small low-income island states (eg, the nations in the Pacific that were COVID-19-free in November 2020), the risk might still be considered too high, especially if they have limited surveillance and outbreak control capacity. In these states, either all shore leave could be denied (ie, cargo movement is performed without the crew leaving the vessel), or the ships that recently visited countries where COVID-19 transmission is occurring could be completely prohibited (eg, until a vaccine against COVID-19 is in use). Other policy options for risk reduction might include:

  • Working with source countries to ensure that departing shipping crew get routinely tested for SARS-CoV-2 just prior to departure, and that any infected crew member is immediately replaced.
  • Testing the crew twice with PCR tests in the destination country. Firstly, at the initial port visited in the destination country but with no shore leave permitted at this port. Then a second test at the second port visit in the country, at which point shore leave could be permitted if all rounds of test results are negative. Also, once rapid tests are considered reliable enough and cost-effective enough, then crew could potentially be tested daily after their first port contact and until they leave the country.
  • Ensuring that any shore leave is highly supervised or otherwise constrained to specific settings. Supervision by port authorities could be used to ensure high adherence with mask use and attendance at only designated settings (eg, specific seafarer clubs). Settings where super-spreading events could potentially occur (eg, restaurants, bars and night clubs) could be prohibited as part of shore leave.
  • Limiting shore leave to just a particular port in the country in a town or city where there is particularly intensive routine PCR testing of port workers and in relevant parts of the community (to facilitate early outbreak detection). Such community testing, combined with testing all people hospitalised with respiratory symptoms, can potentially accelerate early outbreak detection.19
  • Prioritising the provision of vaccination to shipping crews once vaccines against SARS-CoV-2 infection are available in the relevant countries.

Conclusions

Using simulation modelling, we estimated the risk of COVID-19 outbreaks in COVID-19-free settings as a result of merchant ship crews infected at the source of their voyage taking shore leave. Our results can potentially inform policymaker decisions about regulations regarding shore leave for crews and the use of various control measures such as PCR testing and mask use to minimise the risks if shore leave is permitted.

Summary

Abstract

AIM: We aimed to estimate the risk of COVID-19 outbreaks in a COVID-19-free destination country (New Zealand) associated with shore leave by merchant ship crews who were infected prior to their departure or on their ship. METHODS: We used a stochastic version of the SEIR model CovidSIM v1.1 designed specifically for COVID-19. It was populated with parameters for SARS-CoV-2 transmission, shipping characteristics and plausible control measures. RESULTS: When no control interventions were in place, we estimated that an outbreak of COVID-19 in New Zealand would occur after a median time of 23 days (assuming a global average for source country incidence of 2.66 new infections per 1,000 population per week, crews of 20 with a voyage length of 10 days and 1 day of shore leave per crew member both in New Zealand and abroad, and 108 port visits by international merchant ships per week). For this example, the uncertainty around when outbreaks occur is wide (an outbreak occurs with 95% probability between 1 and 124 days). The combination of PCR testing on arrival, self-reporting of symptoms with contact tracing and mask use during shore leave increased this median time to 1.0 year (14 days to 5.4 years, or a 49% probability within a year). Scenario analyses found that onboard infection chains could persist for well over 4 weeks, even with crews of only 5 members. CONCLUSION: This modelling work suggests that the introduction of SARS-CoV-2 through shore leave from international shipping crews is likely, even after long voyages. But the risk can be substantially mitigated by control measures such as PCR testing and mask use.

Aim

Method

Results

Conclusion

Author Information

Nick Wilson: BODE3 Programme, University of Otago Wellington, New Zealand; HEIRU, University of Otago Wellington, New Zealand. Tony Blakely: Population Interventions, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Australia. Michael G. Baker: HEIRU, University of Otago Wellington, New Zealand. Martin Eichner: Epimos GmbH, Germany; University of Tübingen, Germany.

Acknowledgements

Correspondence

Professor Nick Wilson

Correspondence Email

nick.wilson@otago.ac.nz

Competing Interests

Nil.

1. McLeod MA, Baker M, Wilson N, Kelly H, Kiedrzynski T, Kool JL. Protective effect of maritime quarantine in South Pacific jurisdictions, 1918-19 influenza pandemic. Emerg Infect Dis 2008;14:468-70.

2. Kool JL, Pavlin BI, Musto J, Dawainavesi A. Influenza surveillance in the Pacific Island countries and territories during the 2009 pandemic: an observational study. BMC Infect Dis 2013;13:6.

3. Moriarty LF, Plucinski MM, Marston BJ, Kurbatova EV, Knust B, Murray EL, et al. Public Health Responses to COVID-19 Outbreaks on Cruise Ships - Worldwide, February-March 2020. MMWR 2020;69:347-52.

4. Ing AJ, Cocks C, Green JP. COVID-19: in the footsteps of Ernest Shackleton. Thorax 2020;75:693-94.

5. Addetia A, Crawford K, Dingens A, Zhu H, Roychoudhury P, Huang M, Jerome K, Bloom J, Greninger A. Neutralizing antibodies correlate with protection from SARS-CoV-2 in humans during a fishery vessel outbreak with high attack rate. medRxiv 2020;(14 August). https://www.medrxiv.org/content/10.1101/2020.08.13.20173161v1.

6. Dai Q, Hu S, Yan K, Chen Z, Chen B, Cai T, Zhang S, Zhang J, Zheng J. Reflection on SARS-CoV-2 infection of container ship seafarers. Travel Med Infect Dis 2020:101787.

7. Stannard S. COVID-19 in the maritime setting: the challenges, regulations and the international response. Int Marit Health 2020;71:85-90.

8. Amos O. Ten countries kept out Covid. But did they win? BBC News 2020;(24 August). https://www.bbc.com/news/world-asia-53831063.

9. Baker MG, Wilson N, Anglemyer A. Successful elimination of Covid-19 transmission in New Zealand. N Engl J Med 2020;(7 August) doi:101056/NEJMc2025203 https://wwwnejmorg/doi/101056/NEJMc2025203?url_ver=Z3988-2003&rfr_id=ori:rid:crossreforg&rfr_dat=cr_pub%20%200pubmed.

10. Ministry of Health. COVID-19: Epidemic notice and Orders (updated 5 September, 2020). https://www.health.govt.nz/our-work/diseases-and-conditions/covid-19-novel-coronavirus/covid-19-current-situation/covid-19-epidemic-notice-and-orders#maritime.

11. Ministry of Health. Requirements for vessels to receive pratique (updated 5 April 2020). Wellington: Ministry of Health.

12. Australian Associated Press. Queensland investigates Covid-19 strain on cargo ship after New Zealand alert. The Guardian 2020;(24 October). https://www.theguardian.com/world/2020/oct/24/queensland-investigates-covid-19-strain-on-cargo-ship-after-new-zealand-alert.

13. Ministry of Health. 1 new case of COVID-19 in managed isolation; no new community cases. Media Release 2020;(30 October). https://www.health.govt.nz/news-media/media-releases/1-new-case-covid-19-managed-isolation-no-new-community-cases.

14. Laschon E. Two crew left Al Messilah livestock ship at Fremantle before coronavirus outbreak revealed. ABC News 2020;(20 October). https://www.abc.net.au/news/2020-10-20/two-australians-allowed-to-disembark-covid-ship-fremantle/12785898.

15. Wilson N, Baker M, Eichner M. Estimating the impact of control measures to prevent outbreaks of COVID-19 associated with air travel into a COVID-19-free country: A simulation modelling study. medRxiv 2020;(17 June). https://www.medrxiv.org/content/10.1101/2020.06.10.20127977v3.

16. Wilson N, Telfar Barnard L, Kvalsig A, Verrall A, Baker M, Schwehm M. Modelling the potential health impact of the COVID-19 pandemic on a hypothetical European country. medRxiv 2020;(23 March). https://medrxiv.org/cgi/content/short/2020.03.20.20039776v1.

17. Wilson N, Parry M, Verrall A, Baker M, Schwehm M, Eichner M. When can elimination of SARS-CoV-2 infection be assumed? Simulation modelling in a case study island nation. medRxiv 2020;(20 May). https://medrxiv.org/cgi/content/short/2020.05.16.20104240v1.

18. Wilson N, Telfar Barnard L, Kvalsvig A, Baker M. Potential health impacts from the COVID-19 pandemic for New Zealand if eradication fails: Report to the NZ Ministry of Health. Wellington: University of Otago Wellington. 2020;https://www.health.govt.nz/system/files/documents/publications/report_for_moh_-_covid-19_pandemic_nz_final.pdf.

19. Wilson N, Schwehm M, Verrall AJ, Parry M, Baker MG, Eichner M. Detecting the re-emergent COVID-19 pandemic after elimination: modelling study of combined primary care and hospital surveillance. N Z Med J 2020;133:28-39.

20. World Health Organization. Coronavirus disease (COVID-19) Situation Report – 208. 2020;(15 August). https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports

21. United Nations. Population Division: World Population Prospects 2019 (data query result for 2020). https://population.un.org/wpp/DataQuery/.

22. Havers FP, Reed C, Lim T, Montgomery JM, Klena JD, Hall AJ, et al. Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States, March 23-May 12, 2020. JAMA Intern Med 2020.

23. Pollan M, Perez-Gomez B, Pastor-Barriuso R, Oteo J, Hernan MA, Perez-Olmeda M, et al. Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study. Lancet 2020.

24. Treibel T, Manisty C, Burton M, McKnight A, Lambourne J, Augusto J, Couto-Parada X, Cutino-Moguel T, Noursadeghi M, Moon J. COVID-19: PCR screening of asymptomatic healthcare workers at London hospital. Lancet 2020;395:1608-10.

25. Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 2020;(20 May). https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html.

26. WHO-China Joint Mission. Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19). 2020;(16-24 February). https://www.who.int/docs/default-source/coronaviruse/who-china-joint-mission-on-covid-19-final-report.pdf.

27. Woelfel R, Corman V, Guggemos W, Seilmaier M, Zange S, Mueller M, et al. Clinical presentation and virological assessment of hospitalized cases of coronavirus disease 2019 in a travel-associated transmission cluster. medRxiv 2020;(8 March). https://www.medrxiv.org/content/10.1101/2020.03.05.20030502v1.

28. Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 2020;(Updated 10 July). https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html.

29. Endo A, Centre for the Mathematical Modelling of Infectious Diseases COVID-19 Working Group, Abbott S, Et al. Estimating the overdispersion in COVID-19 transmission using outbreak sizes outside China. Wellcome Open Res 2020, 5:67 (https://doi.org/10.12688/wellcomeopenres.15842.1).

30. Miller D, Martin M, Harel N, Kustin T, Tirosh O, Meir M, Sorek N, Gefen-Halevi S, Amit S, Vorontsov O, Wolf D, Peretz A, Shemer-Avni Y, Roif-Kaminsky D, Kopelman N, Huppert A, Koelle K, Stern A. Full genome viral sequences inform patterns of SARS-CoV-2 spread into and within Israel. medRxiv 2020;(22 May). https://doi.org/10.1101/2020.05.21.20104521

31. Hamner L, Dubbel P, Capron I, Ross A, Jordan A, Lee J, Lynn J, Ball A, Narwal S, Russell S, Patrick D, Leibrand H. High SARS-CoV-2 attack rate following exposure at a choir practice - Skagit County, Washington, March 2020. MMWR 2020;69:606-10.

32. Ministry of Transport. Freight Information Gathering System (FIGS) - Overseas ship visits. Table 1: Visits to New Zealand Ports [Extracted data for: 2019 Q2-Q4; 2020 Q1). https://www.transport.govt.nz/mot-resources/freight-resources/figs/overseas-ship-visits/

33. Wikipedia. Maritime transport. (Accessed 20 August 2020). https://en.wikipedia.org/wiki/Maritime_transport.

34. Ports of Auckland. Operations [Data for 20 August 2020]. http://www.poal.co.nz/operations/schedules/arrivals.

35. Kucirka LM, Lauer SA, Laeyendecker O, Boon D, Lessler J. Variation in false-negative rate of reverse transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since exposure. Ann Intern Med 2020;(E-publication 19 April).

36. Vienna Airport. Coronavirus PCR tests now possible at Vienna Airport. Vienna Airport 2020. https://www.viennaairport.com/pcrtest.

37. Service RF. In ‘milestone,’ FDA OKs simple, accurate coronavirus test that could cost just $5. Science 2020;(27 August). https://www.sciencemag.org/news/2020/08/milestone-fda-oks-simple-accurate-coronavirus-test-could-cost-just-5?utm_campaign=news_daily_2020-08-27&et_rid=373134636&et_cid=3461836.

38. Chu D, Akl E, Duda S, Solo K, Yaacoub S, Schünemann H, et al. Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta-analysis. Lancet 2020;(E-publication 1 June).

39. Flutracking. NZ participant annual report 2018. https://info.flutracking.net/wp-content/uploads/2020/02/NZ-participant-annual-report-2018.pdf

40. Star News. Auckland lockdown extended, rest of NZ to stay at level 2. Star News 2020;(24 August). https://www.odt.co.nz/star-news/star-national/auckland-lockdown-extended-rest-nz-stay-level-2.

41. Ministry of Health. 7 new cases of COVID-19 (Media Release); 2020;(15 August). https://www.health.govt.nz/news-media/media-releases/7-new-cases-covid-19.

Contact diana@nzma.org.nz
for the PDF of this article

View Article PDF

Historically, shipping has been involved in the global spread of pandemics, and maritime quarantine has been used as a successful control measure (eg, in the 1918 influenza pandemic).1 Maritime quarantine was even used successfully to prevent the arrival of the 2009 influenza pandemic in some island jurisdictions, such as Tokelau.2

The COVID-19 pandemic has also had an impact on maritime vessels during 2020, as well as spread to people on shore. On the Diamond Princess, 19% of the passengers and crew became positive with the pandemic virus (SARS-CoV-2) and there was spread to Japanese responders on shore.3 Similarly, on the Grand Princess, 17% of those tested had positive results.3 On a much smaller cruise ship with 217 passengers and crew onboard, 59% were reported to be test-positive.4 On a fishing vessel, 85% (104/122) of the crew were infected.5 In terms of merchant vessels, an outbreak on a container ship was reported as infecting 23% (5/22) of the crew.6 Other such outbreaks have been detailed in media reporting (referred to in a review7).

In response to the COVID-19 pandemic, border controls have been widely used to limit pandemic spread. Such border controls are particularly relevant for two types of strategy for controlling pandemics: (1) the exclusion strategy, as successfully practiced by some Pacific Island nations (eg, Tonga and the Cook Islands),8 and (2) the elimination strategy, as used by New Zealand9 and other jurisdictions (eg, Mainland China, Taiwan, and Australia).

Some of these jurisdictions have completely prohibited maritime vessels arriving at their seaports from countries that are not COVID-19-free (eg, the Marshall Islands have prohibited such incoming ships8). But time periods are also used to lower risk. For example, a minimum of 14 days at sea before being allowed to enter the Marshall Islands,8 or 14 days plus a negative PCR test to enter New Zealand.10 There is also the standard international requirement for pratique whereby any ‘illness during the voyage’ must be notified to health authorities at the destination port.11 Collectively, these control measures seem to be working fairly well, although in October 2020 New Zealand reported that a ship maintenance worker became infected with SARS-CoV-2 after spending time working on a ship that had recently arrived in the country.12 This worker then infected other workers and a household contact onshore (but with no further known subsequent spread). Genome sequencing has indicated that the source of infection was shipping crew flying into New Zealand to join their ship.13 Also in October 2020, another island nation (Australia) faced outbreaks on two cargo ships in a port in Western Australia, where (in one case) two onboard workers left a ship before the outbreak was detected.14

Given this background, we aimed to expand on previous modelling work for air transport spread of COVID-19,15 to determine the risk of merchant ships being the source of COVID-19 outbreaks in an otherwise COVID-19-free country: New Zealand.

Methods

Model design and parameters for SARS-CoV-2 and COVID-19

We used a stochastic SEIR type model with key compartments for susceptible [S], exposed [E], infected [I] and recovered/removed [R]. The model is a stochastic version of CovidSIM, which was developed specifically for COVID-19 (http://covidsim.eu; version 1.1). Work has been produced from previous versions of this model,15–17 and in two places we detail the relevant equations and their stochastic treatment.18,19 The model was built in Pascal, and the computer code is available on request from the authors.

We ran 100 million simulations for each set of parameter values. Such a large number of simulations was necessary due to the very high probability of zero infected crew members boarding a departing merchant ship, given the low assumed incidence of infection (see Table 1). The overall framework for the processes modelled is shown in Figure 1. The parameters were based on available publications and best estimates used in the published modelling work on COVID-19 (as known to us on 27 August 2020). We assumed that 71% of infected COVID-19 cases develop clearly detectable symptoms (Table 1). Another assumption was the contagiousness in terms of the effective reproduction number (Reff), which was 3.0 among crew members on board the ship and 2.5 in the destination country (Table 1).

Figure 1: Flow diagram of the assumed movements of merchant ship crews in the model including interventions (simplified and not showing all control measures (eg, the seeking of medical attention when symptomatic in the destination country and the associated isolation of identified cases and contact tracing as detailed in Table 2)).  

Shore leave in the destination country

We selected New Zealand as a case study destination country, as it has previously achieved elimination of community transmission of SARS-CoV-29 and appears to have successfully controlled subsequent cross-border incursions of the pandemic virus. Upon arrival of ships in New Zealand, we used a period of shore leave by all the crews of one day (the median time ships are in port, based on Ports of Auckland data, the port in New Zealand’s largest city).

Potential control measures

Potential control measures are detailed in Table 2 and Figure 1 and include a PCR test on all the crews on arrival and mask use by the crews during shore leave. If any crew member tested positive, then the shore leave for that particular crew was assumed to be prohibited and therefore no risk of any community outbreak from shore leave was assumed. If a crew member on shore leave developed and self-reported symptoms and then tested positive, this case would be isolated, and this could also trigger contact tracing, which was assumed to identify 80% of the infected contacts within 48 hours. Identified contacts would be isolated after a delay of one or two days.

Ongoing infection transmission in the destination country

Untraced secondary cases who were infected by crew members in the destination country, and tertiary cases who were infected by traced secondary cases before they were isolated, were assumed to roam freely for the full length of their infectious period and to potentially trigger outbreaks in the community.

Table 1: Input parameters used for modelling the potential spread of SARS-CoV-2 infections via merchant shipping with the stochastic version of CovidSIM (v1.1).

Control measures assumptions

The full details on the considered control measures are given in Table 2.

Table 2: Control measures and their estimated efficacy.

Results

The results suggest that, if no pandemic-related maritime controls were in place, the COVID-19-free destination country (New Zealand) would quickly experience an outbreak because of the arrival of ships with infected crew members taking shore leave. That is an outbreak after a median duration of 0.064 years (23 days), which is equivalent to a total of 355 port visits and 7,100 total days of shore leave (for international vessels with 20 crew members and one day of shore leave per person per port; Table 3). However, there is high uncertainty, with 95% of outbreaks likely to occur between 1 to 124 days (ie, 0.0023 to 0.34 years; Table 3).

The median time to an outbreak would increase markedly by obligatory PCR testing of crew members before shore leave is permitted (ie, up to 168 days (0.46 years), or after a total of 2,592 port visits). An even further reduction of risk would occur when requiring face mask use during shore leave (increased median time to 1.00 years). But, relatively little extra gain in risk reduction would result from any sick crew on shore leave self-reporting symptoms and the associated contact tracing (Table 3). Using the base case value of Reff=2.5 in New Zealand, a single untraced infection in the community leads to an outbreak in 88.2% of cases (78.5% for Reff=2.0). When we considered super-spreading events in the community in a scenario analysis, the outbreak probability per person was actually reduced to 57.4%. This is because allowing for super-spreading events means that a smaller proportion of infected crew members transmit infection, even though those that do will typically infect more people (assuming the same overall value of Reff).

In scenario analyses, a smaller crew size reduced the outbreak risk (eg, the median time to an outbreak would be 3.8 years for ships with a crew size of five; Table 4). The risk of outbreaks was also lower when making assumptions around lower contagiousness in the destination country (ie, Reff lowered to 2.0). The risk remained basically unchanged when contagiousness on the ship was assumed to be higher (ie, Reff increased to 4.0). Increasing the shore leave to either two or three days increased the risk of an outbreak (ie, it reduced the median time to this event). When super-spreading events were considered in the destination country, this led to the same average number of untraced infections caused by crew members in New Zealand, but as each one of them had a lower risk of leading to an outbreak, the overall outbreak risk was lower than in the baseline study.

Figure 2 shows that voyage duration is a key determinant of outbreak risk in the destination country, and this risk is especially high once voyages are longer than five or so days (ie, once the latent period is typically over and crews have become infectious). Onboard spread can maintain this risk over subsequent weeks, leading to more infected individuals on board; but this also increases the detection probability by testing on arrival in New Zealand. It can take a long time for the onboard epidemic to come close to ‘burning out’. Indeed, the outbreak risk in the destination country when there are no controls only starts to decline after a voyage time of three weeks, and even then it declines quite slowly (Figure 2). For larger crew sizes of 10 to 20 people, the risk of community outbreaks is still increasing slightly after three to four weeks of voyaging when no controls are used (Figure 3 and Figure 4). Interestingly, if PCR tests are implemented, the effect of longer travel durations generates results that are the inverse: the more the infection can spread on board, the more likely it will be detected. As none of the crew members were assumed to be allowed to go to shore if any one of them tested positive, the probability that infected people being allowed to enter New Zealand decreases as the number of infected people on board the ship increases. Adding additional interventions like wearing masks, self-reporting symptoms and contact tracing further improves the results; but the main effect is obtained by PCR testing prior to shore leave being permitted. With the full set of interventions, the median time to an outbreak increased, but this time varied widely by length of voyage and size of the crew (Figure 3 and Figure 4).

Table 3: Results of the simulations without interventions and with multi-layered interventions (with these being for a base case of 10 days at sea and 108 merchant ship visits per week, 20 crew per ship, one day of shore leave each per port visit in New Zealand, and with 100 million stochastic simulations being run for each set of parameters).

Table 4: Results of the scenario analyses for 108 merchant ship visits per week and the full set of interventions taking place (see last line of Table 3) with 100 million stochastic simulations run for each set of parameters. (For further information, see text and Table 2.)

Figure 2: For ships with five-member crews, the median duration (log-scale in years) until a COVID-19 outbreak occurs in the destination country because of merchant ship crews taking shore leave. We assumed there were 108 cargo ships arrive each week. In the country of origin, each member can become infected at a rate of 0.00038 per day. Infections spread on board with an effective reproduction number Reff of 3.0 and in New Zealand with Reff of 2.5. Note that a voyage duration of 1 day is not applicable to New Zealand. Full black curves: no interventions are taken; full grey curves: all crew members are prevented from entering the country if one of them is PCR positive upon arrival; dotted grey curves: full set of interventions as outlined in Table 3. For each combination of crew size and voyage duration, 100 million voyages were simulated.

Figure 3: For ships with 10-member crews, the median duration (log-scale in years) until a COVID-19 pandemic outbreak occurs in the destination country because of merchant ship crews taking shore leave (other details as per Figure 2).

Figure 4: For ships with 20-member crews, the median duration (log-scale in years) until a COVID-19 pandemic outbreak occurs in the destination country because of merchant ship crews taking shore leave (other details as per Figure 2).

Discussion

Main findings

In this modelling work, we found that it might only be a matter of a few weeks before crew from international trading maritime vessels would trigger COVID-19 pandemic outbreaks in the destination country, if no control measures were in place. Of particular note is that even small five-person crews appear to contribute a risk after voyages of several weeks, and this risk only declines slowly thereafter. Fortunately, however, the risk of such outbreaks can be substantially reduced with the available interventions. In particular, PCR testing before leaving the vessel appears to be a valuable intervention, though this benefit still comes with high uncertainty as indicated by the wide range for 95% of the simulation results (shown in Table 3).

The results for our case study country (New Zealand) are likely generalisable to most countries that have seaports and maritime trade. Nevertheless, the risk could be somewhat less for some nations on a per population or per gross domestic product (GDP) basis because New Zealand’s economy is particularly trade orientated and especially dependent on sea trade. That is, it has no international trade by land routes and only a small proportion of trade volume is by air cargo. With a population of five million, New Zealand has 1,120 port visits from vessels with an international origin per million population per year.

Study strengths and limitations

This is the first study (to our knowledge) to explore the risk of COVID-19 outbreaks arising from shore leave of maritime ship crews. Another strength is that the work builds on an established model that has been used to also study air transport and other aspects of SARS-CoV-2 transmission (see Methods).

But, as with all modelling, there are important limitations. Some of these relate to parameters. A particularly critical one is the daily incidence of SARS-CoV-2 infection in the source country that the ship leaves from. We used a global average for this incidence to account for the diverse maritime trading patterns that New Zealand has and also because the crews are internationally diverse (often flying in from another country just prior to the ship’s departure, which may expose them to higher risks via air terminals and on aircraft). Yet there are likely to be highly variable risks of infection between different source countries that the ship leaves from and countries that the crew come from, and these will change with the evolving global pandemic of COVID-19.

Other examples of parameter limitations are the Reff onboard such vessels and the Reff for shore leave by crew. The former is likely to vary by different designs of merchant vessels (container ships vs tankers vs bulk carriers etc) and also by size (eg, it is likely that, in vessels of under 3,000 gross tonnage, the crew are in shared sleeping rooms). However, we did not have sufficient data to model such heterogeneity. We also didn’t account for prior immunity among crew members from past exposure to the SARS-CoV-2 virus, which is likely to increase over time with global progression of the pandemic. Given the data limitations, we did not consider port calls and shore leave on route between the original departure point and the first New Zealand port of call. Such port calls may either increase the risk for New Zealand (if the visited port city on route has a higher incidence of infection than the origin country), or they may decrease the risk (by extending the time length of the voyage, if the origin country had a higher incidence of infection than the visited city). We also did not model risk of transmission to port workers who might go onto arriving ships (eg, pilots and health workers conducting PCR tests on board vessels), given the assumption that they would take appropriate precautions with physical distancing and use of personal protective equipment. Yet people don’t always follow rules and accidental events may reduce the effectiveness of preventive measures.

Possible implications for future research and policy

Future research is needed to replicate this study (eg, using simulation models with a different structure and for a wider range of destination countries). The routine collection of international shipping transponder data, which is currently underway by other New Zealand-based researchers (funded by the Ministry of Business, Innovation and Employment), may also more precisely identify ship movements, travel times and also unusual events (such as ships exchanging supplies or crew at sea). Research could also explore the acceptability of, and adherence to, mask use by crews on shore leave in different settings.

As detailed above, the results in Table 2 and Table 3 might make some health authorities decide that the risk of allowing shore leave for crew is tolerable with control interventions such as PCR testing and mask use. But for small low-income island states (eg, the nations in the Pacific that were COVID-19-free in November 2020), the risk might still be considered too high, especially if they have limited surveillance and outbreak control capacity. In these states, either all shore leave could be denied (ie, cargo movement is performed without the crew leaving the vessel), or the ships that recently visited countries where COVID-19 transmission is occurring could be completely prohibited (eg, until a vaccine against COVID-19 is in use). Other policy options for risk reduction might include:

  • Working with source countries to ensure that departing shipping crew get routinely tested for SARS-CoV-2 just prior to departure, and that any infected crew member is immediately replaced.
  • Testing the crew twice with PCR tests in the destination country. Firstly, at the initial port visited in the destination country but with no shore leave permitted at this port. Then a second test at the second port visit in the country, at which point shore leave could be permitted if all rounds of test results are negative. Also, once rapid tests are considered reliable enough and cost-effective enough, then crew could potentially be tested daily after their first port contact and until they leave the country.
  • Ensuring that any shore leave is highly supervised or otherwise constrained to specific settings. Supervision by port authorities could be used to ensure high adherence with mask use and attendance at only designated settings (eg, specific seafarer clubs). Settings where super-spreading events could potentially occur (eg, restaurants, bars and night clubs) could be prohibited as part of shore leave.
  • Limiting shore leave to just a particular port in the country in a town or city where there is particularly intensive routine PCR testing of port workers and in relevant parts of the community (to facilitate early outbreak detection). Such community testing, combined with testing all people hospitalised with respiratory symptoms, can potentially accelerate early outbreak detection.19
  • Prioritising the provision of vaccination to shipping crews once vaccines against SARS-CoV-2 infection are available in the relevant countries.

Conclusions

Using simulation modelling, we estimated the risk of COVID-19 outbreaks in COVID-19-free settings as a result of merchant ship crews infected at the source of their voyage taking shore leave. Our results can potentially inform policymaker decisions about regulations regarding shore leave for crews and the use of various control measures such as PCR testing and mask use to minimise the risks if shore leave is permitted.

Summary

Abstract

AIM: We aimed to estimate the risk of COVID-19 outbreaks in a COVID-19-free destination country (New Zealand) associated with shore leave by merchant ship crews who were infected prior to their departure or on their ship. METHODS: We used a stochastic version of the SEIR model CovidSIM v1.1 designed specifically for COVID-19. It was populated with parameters for SARS-CoV-2 transmission, shipping characteristics and plausible control measures. RESULTS: When no control interventions were in place, we estimated that an outbreak of COVID-19 in New Zealand would occur after a median time of 23 days (assuming a global average for source country incidence of 2.66 new infections per 1,000 population per week, crews of 20 with a voyage length of 10 days and 1 day of shore leave per crew member both in New Zealand and abroad, and 108 port visits by international merchant ships per week). For this example, the uncertainty around when outbreaks occur is wide (an outbreak occurs with 95% probability between 1 and 124 days). The combination of PCR testing on arrival, self-reporting of symptoms with contact tracing and mask use during shore leave increased this median time to 1.0 year (14 days to 5.4 years, or a 49% probability within a year). Scenario analyses found that onboard infection chains could persist for well over 4 weeks, even with crews of only 5 members. CONCLUSION: This modelling work suggests that the introduction of SARS-CoV-2 through shore leave from international shipping crews is likely, even after long voyages. But the risk can be substantially mitigated by control measures such as PCR testing and mask use.

Aim

Method

Results

Conclusion

Author Information

Nick Wilson: BODE3 Programme, University of Otago Wellington, New Zealand; HEIRU, University of Otago Wellington, New Zealand. Tony Blakely: Population Interventions, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Australia. Michael G. Baker: HEIRU, University of Otago Wellington, New Zealand. Martin Eichner: Epimos GmbH, Germany; University of Tübingen, Germany.

Acknowledgements

Correspondence

Professor Nick Wilson

Correspondence Email

nick.wilson@otago.ac.nz

Competing Interests

Nil.

1. McLeod MA, Baker M, Wilson N, Kelly H, Kiedrzynski T, Kool JL. Protective effect of maritime quarantine in South Pacific jurisdictions, 1918-19 influenza pandemic. Emerg Infect Dis 2008;14:468-70.

2. Kool JL, Pavlin BI, Musto J, Dawainavesi A. Influenza surveillance in the Pacific Island countries and territories during the 2009 pandemic: an observational study. BMC Infect Dis 2013;13:6.

3. Moriarty LF, Plucinski MM, Marston BJ, Kurbatova EV, Knust B, Murray EL, et al. Public Health Responses to COVID-19 Outbreaks on Cruise Ships - Worldwide, February-March 2020. MMWR 2020;69:347-52.

4. Ing AJ, Cocks C, Green JP. COVID-19: in the footsteps of Ernest Shackleton. Thorax 2020;75:693-94.

5. Addetia A, Crawford K, Dingens A, Zhu H, Roychoudhury P, Huang M, Jerome K, Bloom J, Greninger A. Neutralizing antibodies correlate with protection from SARS-CoV-2 in humans during a fishery vessel outbreak with high attack rate. medRxiv 2020;(14 August). https://www.medrxiv.org/content/10.1101/2020.08.13.20173161v1.

6. Dai Q, Hu S, Yan K, Chen Z, Chen B, Cai T, Zhang S, Zhang J, Zheng J. Reflection on SARS-CoV-2 infection of container ship seafarers. Travel Med Infect Dis 2020:101787.

7. Stannard S. COVID-19 in the maritime setting: the challenges, regulations and the international response. Int Marit Health 2020;71:85-90.

8. Amos O. Ten countries kept out Covid. But did they win? BBC News 2020;(24 August). https://www.bbc.com/news/world-asia-53831063.

9. Baker MG, Wilson N, Anglemyer A. Successful elimination of Covid-19 transmission in New Zealand. N Engl J Med 2020;(7 August) doi:101056/NEJMc2025203 https://wwwnejmorg/doi/101056/NEJMc2025203?url_ver=Z3988-2003&rfr_id=ori:rid:crossreforg&rfr_dat=cr_pub%20%200pubmed.

10. Ministry of Health. COVID-19: Epidemic notice and Orders (updated 5 September, 2020). https://www.health.govt.nz/our-work/diseases-and-conditions/covid-19-novel-coronavirus/covid-19-current-situation/covid-19-epidemic-notice-and-orders#maritime.

11. Ministry of Health. Requirements for vessels to receive pratique (updated 5 April 2020). Wellington: Ministry of Health.

12. Australian Associated Press. Queensland investigates Covid-19 strain on cargo ship after New Zealand alert. The Guardian 2020;(24 October). https://www.theguardian.com/world/2020/oct/24/queensland-investigates-covid-19-strain-on-cargo-ship-after-new-zealand-alert.

13. Ministry of Health. 1 new case of COVID-19 in managed isolation; no new community cases. Media Release 2020;(30 October). https://www.health.govt.nz/news-media/media-releases/1-new-case-covid-19-managed-isolation-no-new-community-cases.

14. Laschon E. Two crew left Al Messilah livestock ship at Fremantle before coronavirus outbreak revealed. ABC News 2020;(20 October). https://www.abc.net.au/news/2020-10-20/two-australians-allowed-to-disembark-covid-ship-fremantle/12785898.

15. Wilson N, Baker M, Eichner M. Estimating the impact of control measures to prevent outbreaks of COVID-19 associated with air travel into a COVID-19-free country: A simulation modelling study. medRxiv 2020;(17 June). https://www.medrxiv.org/content/10.1101/2020.06.10.20127977v3.

16. Wilson N, Telfar Barnard L, Kvalsig A, Verrall A, Baker M, Schwehm M. Modelling the potential health impact of the COVID-19 pandemic on a hypothetical European country. medRxiv 2020;(23 March). https://medrxiv.org/cgi/content/short/2020.03.20.20039776v1.

17. Wilson N, Parry M, Verrall A, Baker M, Schwehm M, Eichner M. When can elimination of SARS-CoV-2 infection be assumed? Simulation modelling in a case study island nation. medRxiv 2020;(20 May). https://medrxiv.org/cgi/content/short/2020.05.16.20104240v1.

18. Wilson N, Telfar Barnard L, Kvalsvig A, Baker M. Potential health impacts from the COVID-19 pandemic for New Zealand if eradication fails: Report to the NZ Ministry of Health. Wellington: University of Otago Wellington. 2020;https://www.health.govt.nz/system/files/documents/publications/report_for_moh_-_covid-19_pandemic_nz_final.pdf.

19. Wilson N, Schwehm M, Verrall AJ, Parry M, Baker MG, Eichner M. Detecting the re-emergent COVID-19 pandemic after elimination: modelling study of combined primary care and hospital surveillance. N Z Med J 2020;133:28-39.

20. World Health Organization. Coronavirus disease (COVID-19) Situation Report – 208. 2020;(15 August). https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports

21. United Nations. Population Division: World Population Prospects 2019 (data query result for 2020). https://population.un.org/wpp/DataQuery/.

22. Havers FP, Reed C, Lim T, Montgomery JM, Klena JD, Hall AJ, et al. Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States, March 23-May 12, 2020. JAMA Intern Med 2020.

23. Pollan M, Perez-Gomez B, Pastor-Barriuso R, Oteo J, Hernan MA, Perez-Olmeda M, et al. Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study. Lancet 2020.

24. Treibel T, Manisty C, Burton M, McKnight A, Lambourne J, Augusto J, Couto-Parada X, Cutino-Moguel T, Noursadeghi M, Moon J. COVID-19: PCR screening of asymptomatic healthcare workers at London hospital. Lancet 2020;395:1608-10.

25. Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 2020;(20 May). https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html.

26. WHO-China Joint Mission. Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19). 2020;(16-24 February). https://www.who.int/docs/default-source/coronaviruse/who-china-joint-mission-on-covid-19-final-report.pdf.

27. Woelfel R, Corman V, Guggemos W, Seilmaier M, Zange S, Mueller M, et al. Clinical presentation and virological assessment of hospitalized cases of coronavirus disease 2019 in a travel-associated transmission cluster. medRxiv 2020;(8 March). https://www.medrxiv.org/content/10.1101/2020.03.05.20030502v1.

28. Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 2020;(Updated 10 July). https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html.

29. Endo A, Centre for the Mathematical Modelling of Infectious Diseases COVID-19 Working Group, Abbott S, Et al. Estimating the overdispersion in COVID-19 transmission using outbreak sizes outside China. Wellcome Open Res 2020, 5:67 (https://doi.org/10.12688/wellcomeopenres.15842.1).

30. Miller D, Martin M, Harel N, Kustin T, Tirosh O, Meir M, Sorek N, Gefen-Halevi S, Amit S, Vorontsov O, Wolf D, Peretz A, Shemer-Avni Y, Roif-Kaminsky D, Kopelman N, Huppert A, Koelle K, Stern A. Full genome viral sequences inform patterns of SARS-CoV-2 spread into and within Israel. medRxiv 2020;(22 May). https://doi.org/10.1101/2020.05.21.20104521

31. Hamner L, Dubbel P, Capron I, Ross A, Jordan A, Lee J, Lynn J, Ball A, Narwal S, Russell S, Patrick D, Leibrand H. High SARS-CoV-2 attack rate following exposure at a choir practice - Skagit County, Washington, March 2020. MMWR 2020;69:606-10.

32. Ministry of Transport. Freight Information Gathering System (FIGS) - Overseas ship visits. Table 1: Visits to New Zealand Ports [Extracted data for: 2019 Q2-Q4; 2020 Q1). https://www.transport.govt.nz/mot-resources/freight-resources/figs/overseas-ship-visits/

33. Wikipedia. Maritime transport. (Accessed 20 August 2020). https://en.wikipedia.org/wiki/Maritime_transport.

34. Ports of Auckland. Operations [Data for 20 August 2020]. http://www.poal.co.nz/operations/schedules/arrivals.

35. Kucirka LM, Lauer SA, Laeyendecker O, Boon D, Lessler J. Variation in false-negative rate of reverse transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since exposure. Ann Intern Med 2020;(E-publication 19 April).

36. Vienna Airport. Coronavirus PCR tests now possible at Vienna Airport. Vienna Airport 2020. https://www.viennaairport.com/pcrtest.

37. Service RF. In ‘milestone,’ FDA OKs simple, accurate coronavirus test that could cost just $5. Science 2020;(27 August). https://www.sciencemag.org/news/2020/08/milestone-fda-oks-simple-accurate-coronavirus-test-could-cost-just-5?utm_campaign=news_daily_2020-08-27&et_rid=373134636&et_cid=3461836.

38. Chu D, Akl E, Duda S, Solo K, Yaacoub S, Schünemann H, et al. Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta-analysis. Lancet 2020;(E-publication 1 June).

39. Flutracking. NZ participant annual report 2018. https://info.flutracking.net/wp-content/uploads/2020/02/NZ-participant-annual-report-2018.pdf

40. Star News. Auckland lockdown extended, rest of NZ to stay at level 2. Star News 2020;(24 August). https://www.odt.co.nz/star-news/star-national/auckland-lockdown-extended-rest-nz-stay-level-2.

41. Ministry of Health. 7 new cases of COVID-19 (Media Release); 2020;(15 August). https://www.health.govt.nz/news-media/media-releases/7-new-cases-covid-19.

Contact diana@nzma.org.nz
for the PDF of this article

View Article PDF

Historically, shipping has been involved in the global spread of pandemics, and maritime quarantine has been used as a successful control measure (eg, in the 1918 influenza pandemic).1 Maritime quarantine was even used successfully to prevent the arrival of the 2009 influenza pandemic in some island jurisdictions, such as Tokelau.2

The COVID-19 pandemic has also had an impact on maritime vessels during 2020, as well as spread to people on shore. On the Diamond Princess, 19% of the passengers and crew became positive with the pandemic virus (SARS-CoV-2) and there was spread to Japanese responders on shore.3 Similarly, on the Grand Princess, 17% of those tested had positive results.3 On a much smaller cruise ship with 217 passengers and crew onboard, 59% were reported to be test-positive.4 On a fishing vessel, 85% (104/122) of the crew were infected.5 In terms of merchant vessels, an outbreak on a container ship was reported as infecting 23% (5/22) of the crew.6 Other such outbreaks have been detailed in media reporting (referred to in a review7).

In response to the COVID-19 pandemic, border controls have been widely used to limit pandemic spread. Such border controls are particularly relevant for two types of strategy for controlling pandemics: (1) the exclusion strategy, as successfully practiced by some Pacific Island nations (eg, Tonga and the Cook Islands),8 and (2) the elimination strategy, as used by New Zealand9 and other jurisdictions (eg, Mainland China, Taiwan, and Australia).

Some of these jurisdictions have completely prohibited maritime vessels arriving at their seaports from countries that are not COVID-19-free (eg, the Marshall Islands have prohibited such incoming ships8). But time periods are also used to lower risk. For example, a minimum of 14 days at sea before being allowed to enter the Marshall Islands,8 or 14 days plus a negative PCR test to enter New Zealand.10 There is also the standard international requirement for pratique whereby any ‘illness during the voyage’ must be notified to health authorities at the destination port.11 Collectively, these control measures seem to be working fairly well, although in October 2020 New Zealand reported that a ship maintenance worker became infected with SARS-CoV-2 after spending time working on a ship that had recently arrived in the country.12 This worker then infected other workers and a household contact onshore (but with no further known subsequent spread). Genome sequencing has indicated that the source of infection was shipping crew flying into New Zealand to join their ship.13 Also in October 2020, another island nation (Australia) faced outbreaks on two cargo ships in a port in Western Australia, where (in one case) two onboard workers left a ship before the outbreak was detected.14

Given this background, we aimed to expand on previous modelling work for air transport spread of COVID-19,15 to determine the risk of merchant ships being the source of COVID-19 outbreaks in an otherwise COVID-19-free country: New Zealand.

Methods

Model design and parameters for SARS-CoV-2 and COVID-19

We used a stochastic SEIR type model with key compartments for susceptible [S], exposed [E], infected [I] and recovered/removed [R]. The model is a stochastic version of CovidSIM, which was developed specifically for COVID-19 (http://covidsim.eu; version 1.1). Work has been produced from previous versions of this model,15–17 and in two places we detail the relevant equations and their stochastic treatment.18,19 The model was built in Pascal, and the computer code is available on request from the authors.

We ran 100 million simulations for each set of parameter values. Such a large number of simulations was necessary due to the very high probability of zero infected crew members boarding a departing merchant ship, given the low assumed incidence of infection (see Table 1). The overall framework for the processes modelled is shown in Figure 1. The parameters were based on available publications and best estimates used in the published modelling work on COVID-19 (as known to us on 27 August 2020). We assumed that 71% of infected COVID-19 cases develop clearly detectable symptoms (Table 1). Another assumption was the contagiousness in terms of the effective reproduction number (Reff), which was 3.0 among crew members on board the ship and 2.5 in the destination country (Table 1).

Figure 1: Flow diagram of the assumed movements of merchant ship crews in the model including interventions (simplified and not showing all control measures (eg, the seeking of medical attention when symptomatic in the destination country and the associated isolation of identified cases and contact tracing as detailed in Table 2)).  

Shore leave in the destination country

We selected New Zealand as a case study destination country, as it has previously achieved elimination of community transmission of SARS-CoV-29 and appears to have successfully controlled subsequent cross-border incursions of the pandemic virus. Upon arrival of ships in New Zealand, we used a period of shore leave by all the crews of one day (the median time ships are in port, based on Ports of Auckland data, the port in New Zealand’s largest city).

Potential control measures

Potential control measures are detailed in Table 2 and Figure 1 and include a PCR test on all the crews on arrival and mask use by the crews during shore leave. If any crew member tested positive, then the shore leave for that particular crew was assumed to be prohibited and therefore no risk of any community outbreak from shore leave was assumed. If a crew member on shore leave developed and self-reported symptoms and then tested positive, this case would be isolated, and this could also trigger contact tracing, which was assumed to identify 80% of the infected contacts within 48 hours. Identified contacts would be isolated after a delay of one or two days.

Ongoing infection transmission in the destination country

Untraced secondary cases who were infected by crew members in the destination country, and tertiary cases who were infected by traced secondary cases before they were isolated, were assumed to roam freely for the full length of their infectious period and to potentially trigger outbreaks in the community.

Table 1: Input parameters used for modelling the potential spread of SARS-CoV-2 infections via merchant shipping with the stochastic version of CovidSIM (v1.1).

Control measures assumptions

The full details on the considered control measures are given in Table 2.

Table 2: Control measures and their estimated efficacy.

Results

The results suggest that, if no pandemic-related maritime controls were in place, the COVID-19-free destination country (New Zealand) would quickly experience an outbreak because of the arrival of ships with infected crew members taking shore leave. That is an outbreak after a median duration of 0.064 years (23 days), which is equivalent to a total of 355 port visits and 7,100 total days of shore leave (for international vessels with 20 crew members and one day of shore leave per person per port; Table 3). However, there is high uncertainty, with 95% of outbreaks likely to occur between 1 to 124 days (ie, 0.0023 to 0.34 years; Table 3).

The median time to an outbreak would increase markedly by obligatory PCR testing of crew members before shore leave is permitted (ie, up to 168 days (0.46 years), or after a total of 2,592 port visits). An even further reduction of risk would occur when requiring face mask use during shore leave (increased median time to 1.00 years). But, relatively little extra gain in risk reduction would result from any sick crew on shore leave self-reporting symptoms and the associated contact tracing (Table 3). Using the base case value of Reff=2.5 in New Zealand, a single untraced infection in the community leads to an outbreak in 88.2% of cases (78.5% for Reff=2.0). When we considered super-spreading events in the community in a scenario analysis, the outbreak probability per person was actually reduced to 57.4%. This is because allowing for super-spreading events means that a smaller proportion of infected crew members transmit infection, even though those that do will typically infect more people (assuming the same overall value of Reff).

In scenario analyses, a smaller crew size reduced the outbreak risk (eg, the median time to an outbreak would be 3.8 years for ships with a crew size of five; Table 4). The risk of outbreaks was also lower when making assumptions around lower contagiousness in the destination country (ie, Reff lowered to 2.0). The risk remained basically unchanged when contagiousness on the ship was assumed to be higher (ie, Reff increased to 4.0). Increasing the shore leave to either two or three days increased the risk of an outbreak (ie, it reduced the median time to this event). When super-spreading events were considered in the destination country, this led to the same average number of untraced infections caused by crew members in New Zealand, but as each one of them had a lower risk of leading to an outbreak, the overall outbreak risk was lower than in the baseline study.

Figure 2 shows that voyage duration is a key determinant of outbreak risk in the destination country, and this risk is especially high once voyages are longer than five or so days (ie, once the latent period is typically over and crews have become infectious). Onboard spread can maintain this risk over subsequent weeks, leading to more infected individuals on board; but this also increases the detection probability by testing on arrival in New Zealand. It can take a long time for the onboard epidemic to come close to ‘burning out’. Indeed, the outbreak risk in the destination country when there are no controls only starts to decline after a voyage time of three weeks, and even then it declines quite slowly (Figure 2). For larger crew sizes of 10 to 20 people, the risk of community outbreaks is still increasing slightly after three to four weeks of voyaging when no controls are used (Figure 3 and Figure 4). Interestingly, if PCR tests are implemented, the effect of longer travel durations generates results that are the inverse: the more the infection can spread on board, the more likely it will be detected. As none of the crew members were assumed to be allowed to go to shore if any one of them tested positive, the probability that infected people being allowed to enter New Zealand decreases as the number of infected people on board the ship increases. Adding additional interventions like wearing masks, self-reporting symptoms and contact tracing further improves the results; but the main effect is obtained by PCR testing prior to shore leave being permitted. With the full set of interventions, the median time to an outbreak increased, but this time varied widely by length of voyage and size of the crew (Figure 3 and Figure 4).

Table 3: Results of the simulations without interventions and with multi-layered interventions (with these being for a base case of 10 days at sea and 108 merchant ship visits per week, 20 crew per ship, one day of shore leave each per port visit in New Zealand, and with 100 million stochastic simulations being run for each set of parameters).

Table 4: Results of the scenario analyses for 108 merchant ship visits per week and the full set of interventions taking place (see last line of Table 3) with 100 million stochastic simulations run for each set of parameters. (For further information, see text and Table 2.)

Figure 2: For ships with five-member crews, the median duration (log-scale in years) until a COVID-19 outbreak occurs in the destination country because of merchant ship crews taking shore leave. We assumed there were 108 cargo ships arrive each week. In the country of origin, each member can become infected at a rate of 0.00038 per day. Infections spread on board with an effective reproduction number Reff of 3.0 and in New Zealand with Reff of 2.5. Note that a voyage duration of 1 day is not applicable to New Zealand. Full black curves: no interventions are taken; full grey curves: all crew members are prevented from entering the country if one of them is PCR positive upon arrival; dotted grey curves: full set of interventions as outlined in Table 3. For each combination of crew size and voyage duration, 100 million voyages were simulated.

Figure 3: For ships with 10-member crews, the median duration (log-scale in years) until a COVID-19 pandemic outbreak occurs in the destination country because of merchant ship crews taking shore leave (other details as per Figure 2).

Figure 4: For ships with 20-member crews, the median duration (log-scale in years) until a COVID-19 pandemic outbreak occurs in the destination country because of merchant ship crews taking shore leave (other details as per Figure 2).

Discussion

Main findings

In this modelling work, we found that it might only be a matter of a few weeks before crew from international trading maritime vessels would trigger COVID-19 pandemic outbreaks in the destination country, if no control measures were in place. Of particular note is that even small five-person crews appear to contribute a risk after voyages of several weeks, and this risk only declines slowly thereafter. Fortunately, however, the risk of such outbreaks can be substantially reduced with the available interventions. In particular, PCR testing before leaving the vessel appears to be a valuable intervention, though this benefit still comes with high uncertainty as indicated by the wide range for 95% of the simulation results (shown in Table 3).

The results for our case study country (New Zealand) are likely generalisable to most countries that have seaports and maritime trade. Nevertheless, the risk could be somewhat less for some nations on a per population or per gross domestic product (GDP) basis because New Zealand’s economy is particularly trade orientated and especially dependent on sea trade. That is, it has no international trade by land routes and only a small proportion of trade volume is by air cargo. With a population of five million, New Zealand has 1,120 port visits from vessels with an international origin per million population per year.

Study strengths and limitations

This is the first study (to our knowledge) to explore the risk of COVID-19 outbreaks arising from shore leave of maritime ship crews. Another strength is that the work builds on an established model that has been used to also study air transport and other aspects of SARS-CoV-2 transmission (see Methods).

But, as with all modelling, there are important limitations. Some of these relate to parameters. A particularly critical one is the daily incidence of SARS-CoV-2 infection in the source country that the ship leaves from. We used a global average for this incidence to account for the diverse maritime trading patterns that New Zealand has and also because the crews are internationally diverse (often flying in from another country just prior to the ship’s departure, which may expose them to higher risks via air terminals and on aircraft). Yet there are likely to be highly variable risks of infection between different source countries that the ship leaves from and countries that the crew come from, and these will change with the evolving global pandemic of COVID-19.

Other examples of parameter limitations are the Reff onboard such vessels and the Reff for shore leave by crew. The former is likely to vary by different designs of merchant vessels (container ships vs tankers vs bulk carriers etc) and also by size (eg, it is likely that, in vessels of under 3,000 gross tonnage, the crew are in shared sleeping rooms). However, we did not have sufficient data to model such heterogeneity. We also didn’t account for prior immunity among crew members from past exposure to the SARS-CoV-2 virus, which is likely to increase over time with global progression of the pandemic. Given the data limitations, we did not consider port calls and shore leave on route between the original departure point and the first New Zealand port of call. Such port calls may either increase the risk for New Zealand (if the visited port city on route has a higher incidence of infection than the origin country), or they may decrease the risk (by extending the time length of the voyage, if the origin country had a higher incidence of infection than the visited city). We also did not model risk of transmission to port workers who might go onto arriving ships (eg, pilots and health workers conducting PCR tests on board vessels), given the assumption that they would take appropriate precautions with physical distancing and use of personal protective equipment. Yet people don’t always follow rules and accidental events may reduce the effectiveness of preventive measures.

Possible implications for future research and policy

Future research is needed to replicate this study (eg, using simulation models with a different structure and for a wider range of destination countries). The routine collection of international shipping transponder data, which is currently underway by other New Zealand-based researchers (funded by the Ministry of Business, Innovation and Employment), may also more precisely identify ship movements, travel times and also unusual events (such as ships exchanging supplies or crew at sea). Research could also explore the acceptability of, and adherence to, mask use by crews on shore leave in different settings.

As detailed above, the results in Table 2 and Table 3 might make some health authorities decide that the risk of allowing shore leave for crew is tolerable with control interventions such as PCR testing and mask use. But for small low-income island states (eg, the nations in the Pacific that were COVID-19-free in November 2020), the risk might still be considered too high, especially if they have limited surveillance and outbreak control capacity. In these states, either all shore leave could be denied (ie, cargo movement is performed without the crew leaving the vessel), or the ships that recently visited countries where COVID-19 transmission is occurring could be completely prohibited (eg, until a vaccine against COVID-19 is in use). Other policy options for risk reduction might include:

  • Working with source countries to ensure that departing shipping crew get routinely tested for SARS-CoV-2 just prior to departure, and that any infected crew member is immediately replaced.
  • Testing the crew twice with PCR tests in the destination country. Firstly, at the initial port visited in the destination country but with no shore leave permitted at this port. Then a second test at the second port visit in the country, at which point shore leave could be permitted if all rounds of test results are negative. Also, once rapid tests are considered reliable enough and cost-effective enough, then crew could potentially be tested daily after their first port contact and until they leave the country.
  • Ensuring that any shore leave is highly supervised or otherwise constrained to specific settings. Supervision by port authorities could be used to ensure high adherence with mask use and attendance at only designated settings (eg, specific seafarer clubs). Settings where super-spreading events could potentially occur (eg, restaurants, bars and night clubs) could be prohibited as part of shore leave.
  • Limiting shore leave to just a particular port in the country in a town or city where there is particularly intensive routine PCR testing of port workers and in relevant parts of the community (to facilitate early outbreak detection). Such community testing, combined with testing all people hospitalised with respiratory symptoms, can potentially accelerate early outbreak detection.19
  • Prioritising the provision of vaccination to shipping crews once vaccines against SARS-CoV-2 infection are available in the relevant countries.

Conclusions

Using simulation modelling, we estimated the risk of COVID-19 outbreaks in COVID-19-free settings as a result of merchant ship crews infected at the source of their voyage taking shore leave. Our results can potentially inform policymaker decisions about regulations regarding shore leave for crews and the use of various control measures such as PCR testing and mask use to minimise the risks if shore leave is permitted.

Summary

Abstract

AIM: We aimed to estimate the risk of COVID-19 outbreaks in a COVID-19-free destination country (New Zealand) associated with shore leave by merchant ship crews who were infected prior to their departure or on their ship. METHODS: We used a stochastic version of the SEIR model CovidSIM v1.1 designed specifically for COVID-19. It was populated with parameters for SARS-CoV-2 transmission, shipping characteristics and plausible control measures. RESULTS: When no control interventions were in place, we estimated that an outbreak of COVID-19 in New Zealand would occur after a median time of 23 days (assuming a global average for source country incidence of 2.66 new infections per 1,000 population per week, crews of 20 with a voyage length of 10 days and 1 day of shore leave per crew member both in New Zealand and abroad, and 108 port visits by international merchant ships per week). For this example, the uncertainty around when outbreaks occur is wide (an outbreak occurs with 95% probability between 1 and 124 days). The combination of PCR testing on arrival, self-reporting of symptoms with contact tracing and mask use during shore leave increased this median time to 1.0 year (14 days to 5.4 years, or a 49% probability within a year). Scenario analyses found that onboard infection chains could persist for well over 4 weeks, even with crews of only 5 members. CONCLUSION: This modelling work suggests that the introduction of SARS-CoV-2 through shore leave from international shipping crews is likely, even after long voyages. But the risk can be substantially mitigated by control measures such as PCR testing and mask use.

Aim

Method

Results

Conclusion

Author Information

Nick Wilson: BODE3 Programme, University of Otago Wellington, New Zealand; HEIRU, University of Otago Wellington, New Zealand. Tony Blakely: Population Interventions, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Australia. Michael G. Baker: HEIRU, University of Otago Wellington, New Zealand. Martin Eichner: Epimos GmbH, Germany; University of Tübingen, Germany.

Acknowledgements

Correspondence

Professor Nick Wilson

Correspondence Email

nick.wilson@otago.ac.nz

Competing Interests

Nil.

1. McLeod MA, Baker M, Wilson N, Kelly H, Kiedrzynski T, Kool JL. Protective effect of maritime quarantine in South Pacific jurisdictions, 1918-19 influenza pandemic. Emerg Infect Dis 2008;14:468-70.

2. Kool JL, Pavlin BI, Musto J, Dawainavesi A. Influenza surveillance in the Pacific Island countries and territories during the 2009 pandemic: an observational study. BMC Infect Dis 2013;13:6.

3. Moriarty LF, Plucinski MM, Marston BJ, Kurbatova EV, Knust B, Murray EL, et al. Public Health Responses to COVID-19 Outbreaks on Cruise Ships - Worldwide, February-March 2020. MMWR 2020;69:347-52.

4. Ing AJ, Cocks C, Green JP. COVID-19: in the footsteps of Ernest Shackleton. Thorax 2020;75:693-94.

5. Addetia A, Crawford K, Dingens A, Zhu H, Roychoudhury P, Huang M, Jerome K, Bloom J, Greninger A. Neutralizing antibodies correlate with protection from SARS-CoV-2 in humans during a fishery vessel outbreak with high attack rate. medRxiv 2020;(14 August). https://www.medrxiv.org/content/10.1101/2020.08.13.20173161v1.

6. Dai Q, Hu S, Yan K, Chen Z, Chen B, Cai T, Zhang S, Zhang J, Zheng J. Reflection on SARS-CoV-2 infection of container ship seafarers. Travel Med Infect Dis 2020:101787.

7. Stannard S. COVID-19 in the maritime setting: the challenges, regulations and the international response. Int Marit Health 2020;71:85-90.

8. Amos O. Ten countries kept out Covid. But did they win? BBC News 2020;(24 August). https://www.bbc.com/news/world-asia-53831063.

9. Baker MG, Wilson N, Anglemyer A. Successful elimination of Covid-19 transmission in New Zealand. N Engl J Med 2020;(7 August) doi:101056/NEJMc2025203 https://wwwnejmorg/doi/101056/NEJMc2025203?url_ver=Z3988-2003&rfr_id=ori:rid:crossreforg&rfr_dat=cr_pub%20%200pubmed.

10. Ministry of Health. COVID-19: Epidemic notice and Orders (updated 5 September, 2020). https://www.health.govt.nz/our-work/diseases-and-conditions/covid-19-novel-coronavirus/covid-19-current-situation/covid-19-epidemic-notice-and-orders#maritime.

11. Ministry of Health. Requirements for vessels to receive pratique (updated 5 April 2020). Wellington: Ministry of Health.

12. Australian Associated Press. Queensland investigates Covid-19 strain on cargo ship after New Zealand alert. The Guardian 2020;(24 October). https://www.theguardian.com/world/2020/oct/24/queensland-investigates-covid-19-strain-on-cargo-ship-after-new-zealand-alert.

13. Ministry of Health. 1 new case of COVID-19 in managed isolation; no new community cases. Media Release 2020;(30 October). https://www.health.govt.nz/news-media/media-releases/1-new-case-covid-19-managed-isolation-no-new-community-cases.

14. Laschon E. Two crew left Al Messilah livestock ship at Fremantle before coronavirus outbreak revealed. ABC News 2020;(20 October). https://www.abc.net.au/news/2020-10-20/two-australians-allowed-to-disembark-covid-ship-fremantle/12785898.

15. Wilson N, Baker M, Eichner M. Estimating the impact of control measures to prevent outbreaks of COVID-19 associated with air travel into a COVID-19-free country: A simulation modelling study. medRxiv 2020;(17 June). https://www.medrxiv.org/content/10.1101/2020.06.10.20127977v3.

16. Wilson N, Telfar Barnard L, Kvalsig A, Verrall A, Baker M, Schwehm M. Modelling the potential health impact of the COVID-19 pandemic on a hypothetical European country. medRxiv 2020;(23 March). https://medrxiv.org/cgi/content/short/2020.03.20.20039776v1.

17. Wilson N, Parry M, Verrall A, Baker M, Schwehm M, Eichner M. When can elimination of SARS-CoV-2 infection be assumed? Simulation modelling in a case study island nation. medRxiv 2020;(20 May). https://medrxiv.org/cgi/content/short/2020.05.16.20104240v1.

18. Wilson N, Telfar Barnard L, Kvalsvig A, Baker M. Potential health impacts from the COVID-19 pandemic for New Zealand if eradication fails: Report to the NZ Ministry of Health. Wellington: University of Otago Wellington. 2020;https://www.health.govt.nz/system/files/documents/publications/report_for_moh_-_covid-19_pandemic_nz_final.pdf.

19. Wilson N, Schwehm M, Verrall AJ, Parry M, Baker MG, Eichner M. Detecting the re-emergent COVID-19 pandemic after elimination: modelling study of combined primary care and hospital surveillance. N Z Med J 2020;133:28-39.

20. World Health Organization. Coronavirus disease (COVID-19) Situation Report – 208. 2020;(15 August). https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports

21. United Nations. Population Division: World Population Prospects 2019 (data query result for 2020). https://population.un.org/wpp/DataQuery/.

22. Havers FP, Reed C, Lim T, Montgomery JM, Klena JD, Hall AJ, et al. Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States, March 23-May 12, 2020. JAMA Intern Med 2020.

23. Pollan M, Perez-Gomez B, Pastor-Barriuso R, Oteo J, Hernan MA, Perez-Olmeda M, et al. Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study. Lancet 2020.

24. Treibel T, Manisty C, Burton M, McKnight A, Lambourne J, Augusto J, Couto-Parada X, Cutino-Moguel T, Noursadeghi M, Moon J. COVID-19: PCR screening of asymptomatic healthcare workers at London hospital. Lancet 2020;395:1608-10.

25. Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 2020;(20 May). https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html.

26. WHO-China Joint Mission. Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19). 2020;(16-24 February). https://www.who.int/docs/default-source/coronaviruse/who-china-joint-mission-on-covid-19-final-report.pdf.

27. Woelfel R, Corman V, Guggemos W, Seilmaier M, Zange S, Mueller M, et al. Clinical presentation and virological assessment of hospitalized cases of coronavirus disease 2019 in a travel-associated transmission cluster. medRxiv 2020;(8 March). https://www.medrxiv.org/content/10.1101/2020.03.05.20030502v1.

28. Centers for Disease Control and Prevention. COVID-19 pandemic planning scenarios. 2020;(Updated 10 July). https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html.

29. Endo A, Centre for the Mathematical Modelling of Infectious Diseases COVID-19 Working Group, Abbott S, Et al. Estimating the overdispersion in COVID-19 transmission using outbreak sizes outside China. Wellcome Open Res 2020, 5:67 (https://doi.org/10.12688/wellcomeopenres.15842.1).

30. Miller D, Martin M, Harel N, Kustin T, Tirosh O, Meir M, Sorek N, Gefen-Halevi S, Amit S, Vorontsov O, Wolf D, Peretz A, Shemer-Avni Y, Roif-Kaminsky D, Kopelman N, Huppert A, Koelle K, Stern A. Full genome viral sequences inform patterns of SARS-CoV-2 spread into and within Israel. medRxiv 2020;(22 May). https://doi.org/10.1101/2020.05.21.20104521

31. Hamner L, Dubbel P, Capron I, Ross A, Jordan A, Lee J, Lynn J, Ball A, Narwal S, Russell S, Patrick D, Leibrand H. High SARS-CoV-2 attack rate following exposure at a choir practice - Skagit County, Washington, March 2020. MMWR 2020;69:606-10.

32. Ministry of Transport. Freight Information Gathering System (FIGS) - Overseas ship visits. Table 1: Visits to New Zealand Ports [Extracted data for: 2019 Q2-Q4; 2020 Q1). https://www.transport.govt.nz/mot-resources/freight-resources/figs/overseas-ship-visits/

33. Wikipedia. Maritime transport. (Accessed 20 August 2020). https://en.wikipedia.org/wiki/Maritime_transport.

34. Ports of Auckland. Operations [Data for 20 August 2020]. http://www.poal.co.nz/operations/schedules/arrivals.

35. Kucirka LM, Lauer SA, Laeyendecker O, Boon D, Lessler J. Variation in false-negative rate of reverse transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since exposure. Ann Intern Med 2020;(E-publication 19 April).

36. Vienna Airport. Coronavirus PCR tests now possible at Vienna Airport. Vienna Airport 2020. https://www.viennaairport.com/pcrtest.

37. Service RF. In ‘milestone,’ FDA OKs simple, accurate coronavirus test that could cost just $5. Science 2020;(27 August). https://www.sciencemag.org/news/2020/08/milestone-fda-oks-simple-accurate-coronavirus-test-could-cost-just-5?utm_campaign=news_daily_2020-08-27&et_rid=373134636&et_cid=3461836.

38. Chu D, Akl E, Duda S, Solo K, Yaacoub S, Schünemann H, et al. Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta-analysis. Lancet 2020;(E-publication 1 June).

39. Flutracking. NZ participant annual report 2018. https://info.flutracking.net/wp-content/uploads/2020/02/NZ-participant-annual-report-2018.pdf

40. Star News. Auckland lockdown extended, rest of NZ to stay at level 2. Star News 2020;(24 August). https://www.odt.co.nz/star-news/star-national/auckland-lockdown-extended-rest-nz-stay-level-2.

41. Ministry of Health. 7 new cases of COVID-19 (Media Release); 2020;(15 August). https://www.health.govt.nz/news-media/media-releases/7-new-cases-covid-19.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE