Discussion
In this article, Sapporo city was chosen as a region suitable for investigating the transmission of a novel influenza through simulations because of the small daily influxes and outflows of people to and from the region in comparison with other major cities in Japan [28]. In many cities, including Sapporo city, many students and workers commute in crowded trains and buses every day, which will accelerate the spread of infection [13]. Therefore, high-risk casual contact groups based on commuters among wards were introduced into the model.
On the condition that the effective reproduction number R t = 1.6, the baseline scenario showed that, on average, 58% of the entire population in the 100-trial simulations would be infected, including the 19% who would be asymptomatic. In past pandemics, the basic reproductive number (R 0) was estimated to be 2.0–3.0 for “Spanish flu” A (H1N1) [38] and 1.68 for “Asian flu” A (H2N2) [35]. For the swine flu A (H1N1) in 2009, R 0 has been estimated to be 1.4–1.6 [21] and 2.2–3.1 [3] based on data originating from the epidemic in Mexico and to be 2.2 based on the epidemic in Japan [39]. Therefore, further studies will be needed for a situation with a highly effective reproductive number that may cause more damage.
The implementation of TAP as a single intervention was highly effective in suppressing the outbreak during the early period of the first 90 days. Naturally, this effectiveness was enhanced by shortening the necessary tracing period and raising the coverage of tracing. For a necessary tracing period of 2 or 4 days, there were no differences in the amount of antiviral drugs needed for prophylaxis with a 30–70% coverage; when the tracing period was increased to 6 days, however, these amounts rose, on average, to 147.3 and 155.9% for 50 and 70% coverage, respectively, in comparison with 30% coverage. These large differences may result from having a wider spread of the influenza infections before the onset of treatment with antiviral drugs for the situation of a 6-day tracing period, while shorter tracing periods (2–4 days) can contain most persons with an infection risk at an early stage of the outbreak.
The implementation of STAP as a single intervention would also be highly effective in suppressing the outbreak during the early period of the first 90 days, but STAP was inferior to TAP in terms of the ripple effect of the administration of antiviral drugs, most likely because the STAP intervention, unlike TAP, only distributes antiviral drugs to students, including those at a low risk of infection, notwithstanding influenza infection spreading among adults.
School closure and restraint brought about a delay in the peak and a decrease in incidence at the peak. However, it is actually difficult to implement school closure for such a long period, as was assumed in this model, where closure would be continued until no further patients were detected in the ward. Therefore, the effectiveness of school closure may be overestimated. In the outbreak of swine flu A (H1N1) in Japan in 2009, it was reported that the rapid spread of infection disrupted medical institutions [36]. Implementing a policy of school closure and restraint may be one approach to preventing an acute disruption of services at medical institutions. However, school closure also brings about the long-term absence of parents from work, which leads to economic loss [40]. The importance of restraint was recognized by only 46.2% citizens in Japan and the loss of precautions against infection may spread rapidly providing that a novel influenza infection has a low mortality [37]; therefore improved recognition of the necessity of restraint is needed.
The specific combination of the three interventions TAP, school closure, and restraint was shown to be the most effective of all the combined interventions (scenarios 1–5) in terms of suppressing infection. It resulted in a average decrease in total patients to 0.02% of that of the baseline scenario because of concomitantly preventing the spread of infection at an early stage by prophylaxis and decreasing infection risk by school closure and restraint. On the other hand, the combination of prophylaxis with antiviral drugs (TAP/STAP) with either school closure or restraint would reduce the amount of patients to some extent but would not be able to shorten the epidemic period because the influenza infection would be spread among people, unlike the combination of prophylaxis with both school closure and restraint.
The TAP intervention should be stopped after the influenza infection has spread widely due to the priority shifting from prophylaxis to the treatment of patients [5]. Moreover, it may be difficult to identify persons in close contact. It has been reported that the spread of the infections made it difficult for doctors to diagnose swine flu A (H1N1) cases and that it was also hard to trace persons in close contact during this outbreak [36]. In contrast, STAP is easy to implement because antiviral drugs are distributed only at schools, although STAP does have a weak ripple effect due to drug wastage. Therefore, it is necessary to improve the methods of distribution in STAP, such as its range and frequency.
The results of this study will be helpful in planning strategies for the suppression of an outbreak of a novel influenza not only Sapporo city but also in other cities.