Introduction
The recent spread of avian influenza A (H5N1) in wild and domestic poultry has raised global concern over a pandemic outbreak in humans [1]. Since 2003, about 450 cases of avian influenza A (H5N1) infection in humans have been reported in 15 countries, mainly in Southeast Asia [2]. In the study reported here, we have attempted to predict the effectiveness of intervention strategies against an epidemic of a novel influenza that will be caused by new strains of influenza achieving the same transmission ability in humans as avian influenza A (H5N1) by simulating various scenarios targeting Sapporo city, the capital of Hokkaido in Japan, with a population of about 1.9 million. Influenza A (H1N1), which is a new flu virus of swine origin, was first detected in Mexico in March 2009 [3] and has spread rapidly across the globe. A total of 209,438 cases of influenza A (H1N1) infection in over 170 countries, including 2,185 deaths, had been officially reported by late August 2009 [4]. The Ministry of Health, Labor, and Welfare of Japan [5] designed a “Pandemic Influenza Preparedness Action Plan” which, among other measures, includes the administration of antiviral drugs and school closure in the case of an influenza pandemic.
Mathematical models have been developed for influenza transmission. A series of studies using stochastic models were carried out between 1964 and 1976 [6–9] and, recently, there have been several analytical studies on various containment strategies against an influenza pandemic for Southeast Asia [10, 11] and for the USA [12]. Using an individual-based model (IBM), Ohkusa and Sugawara [13] recently investigated the spread of influenza in a metropolitan area in Japan as a result of infection occurring on crowded trains.
The objective of the study reported here was to evaluate the possibilities for suppressing a pandemic through interventions. To this end, we constructed an IBM for the transmission of a novel influenza virus in Sapporo city that takes personal information, such as age, household, habitation, social activity group, casual contact group, and behavioral patterns, into account, thereby resulting in a more realistic model. A series of interventions were explored: (1) targeted antiviral prophylaxis (TAP) through prescribing antiviral drugs for symptomatic patients and persons in close contact with them; (2) school-age targeted antiviral prophylaxis (STAP) by prescribing antiviral drugs for school-aged children; (3) school closure; (4) restraint.
The simulation results showed (1) that both TAP and STAP interventions could be effective in suppressing an outbreak within an early period of 90 days but that STAP would be inferior to TAP in terms of the ripple effect of the administration of antiviral drugs and (2) school closure and restraint could bring about a delay in the peak of infection. The combination of three interventions—TAP, school closure, and restraint—would decrease the number of total patients to 0.02% of that with a no-intervention situation and would be highly effective in suppressing infection. In comparison, although the combination of TAP with only school closure or restraint would have some degree of effectiveness in reducing the number of patients, this strategy would be unable to shorten the epidemic period. The results of this study will be helpful in planning intervention strategies against a future influenza pandemic.