Some interventions implemented in isolation were found to have a modest effect on reducing transmission, by which we mean (roughly) that they cause a substantial reduction in the number of eventual cases and a lower epidemic curve when R is 1.5 but have little effect when R=2.5 or larger. They can still play a substantial role in the overall control if other interventions, such as personal infection control and distancing, are able to bring the effective reproduction number close to 1.
Isolating individuals as soon as possible after diagnosis was found to have a modest effect on reducing transmission, particularly under the assumption of a peaked infectiousness function. This result is sensitive to the assumption that the potential for infecting others prior to becoming symptomatic is substantial. Evidence suggests that this is likely for currently circulating strains of the influenza virus, but it may not be the case for a newly emerged influenza virus. Ferguson et al. (2005) and Longini et al. (2005) assume that infectiousness begins after becoming symptomatic. The factor by which isolating cases reduces R is sensitive to the shape of the infectiousness function (flat or peaked), but it is not sensitive to the magnitude of R0 . The reason why isolating cases has a modest effect stems from the fact that, in practice, an influenza case is likely to present after a considerable part of their potential to infect others has passed.
Personal infection control and distancing measures such as wearing a P2 mask, frequently washing hands and avoiding unnecessary close contacts, have considerable potential to reduce transmission. Not only can they reduce the eventual attack rate, but they also lower the epidemic curve substantially and delay the peak of the epidemic. As a consequence, the epidemic is less disruptive under this control measure. However, this intervention relies on the co-operation of individuals. Community members are likely to comply with requests to practice personal infection control and distancing if severe illness or death is seen to be likely for the emerged pandemic influenza, but compliance is likely to be low if individuals think that the health risk from pandemic influenza is similar to that from currently circulating influenza. Personal infection control and distancing can be very effective because the transmission rate is reduced by both susceptible and infected individuals seeking to avoid infectious contacts.
Closing schools was found to have a modest effect on reducing transmission. This conclusion goes against the expectation of some experts, because children are thought to be infectious longer and their behaviour and hygiene suggests they play a major role in the transmission. However, this perception of children’s higher susceptibility and infectivity does not seem to translate to transmission at school being a major driver of community transmission. The modest effect of closing schools is found for choices of parameters values that are consistent with available data on influenza attack rates of children relative to adults from past pandemics, even when we allow for plausible rates of unobserved asymptomatic infections. We should note, however, that closing schools does assist in reducing the attack rate in school children, and so this measure may play a role in helping to protect school children – particularly if they were unwilling to adopt personal infection control measures.
The effect of closing non-essential workplaces depends on the number of individuals that modify their behaviour. Per individual, the effect is less than that of closing schools, because adults spread less infection than children, but there are many more adults than school children, so the overall effect may be greater. If all children stay home from school and a large proportion of adults stay home from work, there is the potential to eliminate disease spread for low values of the reproduction number.
The conclusion that restricting travel within Australia provides limited scope for delaying the spread of infection from one capital to another when R>1 is, like the results on border control, quite robust. Once the incidence of infection starts its exponential growth in one capital, travel from there has to virtually stop completely to prevent geographic spread of the infection. However, there does seem scope for remote areas to be able to delay importation of the infection into their area by carefully monitored travel restrictions into their area.
To reduce transmission substantially, quarantining households has to be timely. Then household members who may be infected are prevented from circulating in the community prior to onset of symptoms. In other words, quarantining prevents some infected individuals from circulating in the community earlier than is possible by isolating them once they become a diagnosed case. The cost is that some individuals will be quarantined unnecessarily. Compliance may be a problem, but can potentially be boosted by offering quarantined household members AVs to protect them from infections/severe illness. Timely quarantining of affected households that are large is particularly effective.
Timely use of AVs for prophylaxis of individuals likely to be, or to have been, exposed to an infected person reduces transmission in a similar way to personal infection control and distancing. That is, there is a reduced chance of being infected and, when infected, there is a reduced potential to infect others. Therefore, just like personal distancing, timely use of AVs for prophylaxis has the potential to reduce transmission substantially and hence delay the peak of the epidemic. However, in contrast to personal distancing, targeted use of AVs for prophylaxis has some practical difficulties. These include:
- Many courses of AVs will be dispensed unnecessarily.
- It takes time to identify, reach and dispense AVs to potentially exposed community members.
Timely dispensing of AVs for prophylaxis is therefore difficult to achieve in practice. Dispensing AVs for prophylaxis to an individual who is already infected reduces transmission only by lowering that person’s infectivity, from the time when AVs are administered.
Our assessment of the reduction in transmission resulting from using AVs for targeted prophylaxis is reliant on estimates of effectiveness made from studies on circulating influenza. The effectiveness against a newly emerged influenza virus may be different.Top of page
Combining the interventions
The above comments often relate to interventions used singly, which enables us to compare their effects relative to each other. In practice interventions will be used in combination. A useful way to judge how effective the interventions are in combination is to estimate their effect on the reproduction number and, in particular, seeing if the combination can bring the effective reproduction number below 1. The question of whether R<1 can be achieved depends critically on the value of R0 . One must therefore face the possibility that the newly emerged pandemic influenza is very infectious, quite different from circulating influenza and that antecedent immunity is minimal. In other words, the baseline R0 may be higher than we have assumed. For a crude calculation, let us suppose that ‘isolating diagnosed cases’ and ‘personal infection control and distancing’ are able to reduce R by a factor of 0.8 and 0.5, respectively. Further, suppose that ‘household quarantining’ and ‘use of AVs for prophylaxis’ are each able to reduce R by a further factor of 0.5. The factor of 0.5 for these two interventions is more speculative since these interventions are not independent of ‘isolating diagnosed cases’ and they depend on how quickly households are quarantined and targeted AVs are dispensed. Achieving these fractional reductions for the four interventions simultaneously in the field would be challenging. Combining these interventions gives an effective reproduction number of 0.8×0.5×0.5×0.5×R0 = 0.1R0 . Setting this to 1 suggests that we can obtain an effective R below 1 even if R0 is as large as about 10. If this could be achieved the infection can be eliminated with probability 1 if that combination of interventions is sustained until fade-out occurs.