When the coronavirus disease 2019 (COVID-19) pandemic was first declared, several cruise ships were found to be carrying infected passengers or crew at various points. Having such a situation at sea is a vexing problem. A new study published on the preprint server medRxiv* in November 2020 discusses such a scenario in terms of various possible approaches. It suggests that a viable strategy could be evacuating everyone from the ship and extensive surveillance testing.
*Important notice: medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.
Of course, any setting where people are gathered in close proximity, such as schools and care homes, pose a risk of pathogen transmission during pandemics. However, outbreaks on a ship are incredibly challenging to handle for several reasons.
For one, testing facilities are typically limited on board and may be very difficult to upsize, short of air-dropping testing kits and trained staff.
Secondly, isolation facilities are also likely to be limited, and local authorities are all too often likely to forbid offloading infected passengers at their ports in the face of a spreading pandemic.
On a military carrier, these problems are exacerbated by the limited space within which the crew typically lives. For respiratory pathogens such as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the use of fans to ventilate the ship means that air carrying infectious particles is circulated throughout the quarters without being filtered. This high-contact situation also makes contact tracing non-feasible.
The mass infection aboard the aircraft carrier USS Theodore Roosevelt attracted a great deal of attention. However, the current study seeks to garner valuable lessons from this outbreak, using the Navy staff's detailed notes. Work of this sort helps to assess the efficacy of various strategies used to control the disease.
The Roosevelt outbreak caused several hospitalizations but only one death among the sailors. However, the need for isolation on-shore led to high costs, while the aircraft carrier was off duty for about two months. This was compounded by the inadequate availability of FDA-approved testing kits.
The advantages the Roosevelt enjoyed included a highly compliant standard of medical care, the ability to command an airbridge to shore whenever required because of its own aircraft, as well as having access to a mobile laboratory equipped for COVID-19 testing at sea.
The approach
The first case aboard was identified on March 22, 2020, in the course of its voyage through the Philippine Sea. The Navy used the approach of testing suspected cases followed by isolation of confirmed positives aboard or by airlifting them to land.
However, Guam, which is two days away, at the edge of the Sea, during a typical sail, was reached on March 27, and the crew were not allowed at first to step off the pier. Subsequently, from March 31 until May 5, a mass evacuation was carried out to a Guam-based quarantine center.
After this time, there were around 800 sailors on board, and around 1,200 cases had been confirmed among crew members – indicating that about a quarter of them had been infected. On the other hand, serological testing carried out by the Centers for Disease Control and Prevention showed that 60% of the crew had seroconverted. This study was, however, obviously skewed in favor of those who had received a positive viral RNA test previously, with 61% of the volunteers having tested positive, and therefore indicates that case detection onboard was accurate.
SEIR model for early outbreak parameters
The SEIR epidemiological model was fitted to the outbreak on the aircraft carrier, which allowed the contact period and the initial population exposed to the virus to be identified. The contact period of around 2.2 days was found to correspond to a basic reproduction number R0 of 4.5. Compared to R0 = 15 of another cruise ship that suffered an outbreak while at sea, the Diamond Princess, this was a low value.
Testing failed to identify infected sailors for evacuation
The model was used to predict the number of cases had the sailors been randomly evacuated. This was found to vary very little from the actual number of cases that resulted from targeted evacuation of suspected positives. In other words, even though 500 cases were being confirmed per day in April, compared to the 100 per day in the beginning, the evacuation measures failed to effectively segregate infected and healthy sailors.
The reasons for this may have been the delay in obtaining test results, the lag in the appearance of symptoms following infection, and the high percentage of asymptomatic infection. Thus, one lesson becomes obvious: testing must yield rapid results to produce rapid isolation of infectious cases and reduce viral spread.
Evacuation should protect the health of most evacuees
The researchers also explored what might have happened in the absence of testing. The results indicate that evacuation can reduce the size of the exposed population and thus limit the outbreak. In other words, while most evacuees are not infected, most of those who are not eventually become infected. The researchers found that the predicted size of the outbreak without any naval intervention would have reached around 98% of the ship's crew.
The conclusion of this part of the simulation is: "Above a certain threshold for the rate of evacuation, the outbreak size is only dependent on the final number of sailors remaining on the ship." In this situation, evacuating 6% of the sailors daily worked well to limit the size of the outbreak.
If a ship carrying infection is at sea, a mild delay in evacuation may be compensated for by increasing the rate of evacuation once the ship docks. This is no longer possible after a delay of 3 weeks from the detection of the first case, even if 20% of the crew is evacuated every day after the ship reaches port.
If the initially recommended strategy of evacuating 200 sailors a day until only a tenth were left on board for essential functions had been carried out, the model predicts a final outbreak size of less than half the final actual count of around 1,200 by May 5, at 733. The issue with this plan is that all the sailors left on board finally do become infected, which may not allow the basic functions of the ship to be carried out properly due to the need for physical fitness and alertness. The fate of the sailors left on the ship in real life is not clear since data was available only up to May 5.
The importance of testing
With a 5% rate of daily testing and a 5% evacuation rate per day, from the beginning, the plan above could have been safely carried out, because, say the researchers, "Implementing an effective testing-based isolation strategy significantly improves outcomes for crewmembers remaining on board the ship." For evacuation above this level, the higher the testing goes above 4%, the more marked the drop in outbreak size.
Again, high testing rates are essential in safely retaining larger crews on board, corresponding to daily testing of 25% or more for retention of 50% of the original crew, to below 5% with less than 10% being retained. Below a certain rate, testing fails to improve containment. Such simulations dealing with the possibility of retaining the whole crew are important when the crew cannot be evacuated, and only infected cases can be isolated.
Implications
The occurrence of outbreaks on ships at sea can allow the infectious agent to spread internationally if not controlled appropriately. This has been observed throughout recent history. Several strategies must be considered before determining the best one, such as mass evacuation, surveillance testing and isolation, or quarantine and full crew testing. The choice will depend on the parameters, particularly the R0, the exposed population at baseline, the availability of testing and of isolation facilities and the possibility of mass evacuation. The Navy learned from this experience to nip the outbreak on the destroyer USS Kidd in the bud.
Therefore, despite the simplifications used in the current model, the researchers conclude, "Our results lead to a set of generalizable recommendations for disease control that are broadly applicable to the current COVID-19 crisis as well as to future pandemics."
*Important notice: medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.
Incidence and risk factors of an invasive fungal lung infection among COVID-19 patients