This report uses mathematical models that describe person-to-person transmission of an infectious disease to evaluate the alternative interventions available for the control of an emerged pandemic of influenza. It uses data from past pandemic and currently circulating influenza strains to guide the choice of values for model parameters. However, recognizing that parameter values may be quite different for a newly emerged pandemic influenza strain, the emphasis is on a relative comparison of the effectiveness of interventions and on identifying the circumstances under which measures will be most effective.

Border control

Even with 100% sensitive screening of symptomatic arriving travelers there remains a substantial probability (very roughly, 0.3 when the travel duration is 12 hours) that an infected traveler passes through border screening undetected.

The probability that an outbreak, initiated by a single infected traveler who enters Australia undetected, leads to a major local epidemic can be reduced substantially by

  1. actions that promote early presentation of the infected arrival, and
  2. partial home quarantine of travelers arriving from at-risk regions,

making this probability quite small. Nevertheless, such measures delay the time until a local epidemic begins only marginally (several days or, at best, a few weeks), assuming that the epidemic in the source region gathers momentum. Reducing the number of travelers from the source region delays the local epidemic only marginally unless the number of arrivals from the source region is near zero.

Border control is only useful for preventing disease entry, and provides very little further benefit once an epidemic has gathered widespread momentum within Australia. A criterion for initiating the Australian response is suggested in Section 3.3.

Limiting transmission

Calculations indicate that by themselves, the interventions
  1. isolating cases upon diagnosis,
  2. closing non-essential workplaces and/or schools, and
  3. restricting travel within Australia,

are only modestly effective at limiting the transmission of influenza-like infections. This means that their effect is modest when the basic reproduction number (R0) of the infection exceeds 2. However, as R0 becomes closer to 1 the effectiveness of these interventions, to limit transmission, increases. In particular, the addition of isolating diagnosed cases and closing non-essential workplaces can have a major impact when other interventions bring the effective reproduction number R close to 1. Closing schools does reduce the attack rate in children, and would reduce the overall attack rate effectively if school children were found to have a much higher risk of infection than adults. In combination these interventions can be moderately effective.

By themselves, the interventions

  1. personal infection control and distancing (for example: avoiding close contacts, wearing a P2 mask and frequently washing hands),
  2. quarantining affected households, and
  3. use of antiviral drugs for targeted prophylaxis,

have greater potential for limiting transmission. In combination they can be quite effective. This assessment of the use of antiviral drugs for targeted prophylaxis assumes that their effectiveness against the newly emerged strain is as for circulating strains of influenza. A variety of calculations indicates that the effectiveness of each of these interventions decreases rapidly as the delay in introducing it increases.

The intervention of prohibiting mass gatherings is difficult to evaluate. Its effectiveness is sensitive to the probability of attending a mass gathering event during the infectious period and the mean number infected at such gatherings. Unfortunately data are not available to estimate these quantities.

It is encouraging that all of these interventions, used in combination and with good compliance, seem capable of eliminating a newly emerged pandemic influenza infection with a basic reproduction number as high as 10, although the level of compliance required seems difficult to achieve in practice.

Top of page

Antiviral drugs

Calculations indicate that using antiviral drugs (AVs) only for treatment has a modest effect on transmission. It delays the epidemic peak only when nearly everyone is treated and R0 is less than 1.5.

The indiscriminate use of AVs for prophylaxis limits transmission minimally, and is simply wasteful. In contrast, providing AVs to individuals who are likely to become exposed, or to have had a recent exposure, can substantially reduce or delay most of the transmissions in a local epidemic if the reproduction number is less than 2. Specifically, calculations suggest that timely prophylaxis for 50% of contacts for every case can delay the peak of a local epidemic by about one year if the basic reproduction number is less than 1.7. This calculation assumes that no other interventions are in place. If the reproduction number is much higher, delays can still be achieved, but a greater fraction of contacts would need to be traced – for example, for a reproduction number of 3.3, 90% of contacts would need to be traced to achieve a six month delay. This strategy delays the central part of the local epidemic (including its peak), but there is little change in the eventual size of the epidemic. Apart from the practical difficulty of distributing the antiviral drugs to the right people, this strategy relies on an effective vaccine becoming available by the time the antiviral stockpile runs out, because transmission will accelerate once the stockpile has been used up.

It is important to note that the majority of the stockpile must be able to be distributed during a short amount of time: in many situations modeled, the most severe part of the epidemic lasts only one month, and therefore the stockpile can only be effectively used if Australia has the capacity to dispense tens of thousands of courses per day.

Note that our calculations have not taken account of the possibility that the virus develops resistance to the antiviral drugs. Moreover, our assumptions concerning the effect of AVs are based on trials with currently circulating strains of influenza A. It is essential that an assessment of the effectiveness of AVs against the new influenza strain be made in the event of a pandemic.

Health care workers

Our calculations indicate that a substantial proportion of infections will be due to health care workers (HCWs) if they are not protected by personal protective equipment (PPE) and AVs when tending infected cases. The number of individuals infected by HCWs is reduced substantially by the use of PPE and prophylactic use of AVs for HCWs.

With targeted prophylaxis of influenza-dedicated health care workers and no use of AVs in the community, our models suggest that the stockpile will last over the course of the first wave of an epidemic. However, this will not delay the peak of the epidemic.

If antivirals are also used to prevent community transmission by giving prophylaxis to contacts of new infectious cases, the stockpile will run out before the peak of the epidemic, unless elimination is successful. However, the peak of the epidemic can be delayed through this strategy for several months, which may facilitate wide-spread distribution of a vaccine before the peak of the epidemic.

The number of cases requiring hospitalisation at any one time rises above 10,000 for a significant period of time for an epidemic with a high attack rate. The peak number of cases requiring hospitalisation can be reduced through treatment of infectious cases, and delayed through prophylaxis of case contacts.

Document download

This publication is available as a downloadable document.

Using Mathematical Models to Assess Responses to an Outbreak of an Emerged Viral Respiratory Disease(PDF 873 KB)