Research Article: Health system capacity in Sydney, Australia in the event of a biological attack with smallpox

Date Published: June 14, 2019

Publisher: Public Library of Science

Author(s): Chandini Raina MacIntyre, Valentina Costantino, Mohana Priya Kunasekaran, Eric HY Lau.


Planning for a re-emergent epidemic of smallpox requires surge capacity of space, resources and personnel within health systems. There are many uncertainties in such a scenario, including likelihood and size of an attack, speed of response and health system capacity. We used a model for smallpox transmission to determine requirements for hospital beds, contact tracing and health workers (HCWs) in Sydney, Australia, during a modelled epidemic of smallpox. Sensitivity analysis was done on attack size, speed of response and proportion of case isolation and contact tracing. We estimated 100638 clinical HCWs and 14595 public hospital beds in Sydney. Rapid response, case isolation and contact tracing are influential on epidemic size, with case isolation more influential than contact tracing. With 95% of cases isolated, outbreak control can be achieved within 100 days even with only 50% of contacts traced. However, if case isolation and contact tracing both fall to 50%, epidemic control is lost. With a smaller initial attack and a response commencing 20 days after the attack, health system impacts are modest. The requirement for hospital beds will vary from up to 4% to 100% of all available beds in best and worst case scenarios. If the response is delayed, or if the attack infects 10000 people, all available beds will be exceeded within 40 days, with corresponding surge requirements for clinical health care workers (HCWs). We estimated there are 330 public health workers in Sydney with up to 940,350 contacts to be traced. At least 3 million respirators will be needed for the first 100 days. To ensure adequate health system capacity, rapid response, high rates of case isolation, excellent contact tracing and vaccination, and protection of HCWs should be a priority. Surge capacity must be planned. Failures in any of these could cause health system failure, with inadequate beds, quarantine spaces, personnel, PPE and inability to manage other acute health conditions.

Partial Text

Smallpox is a category A bioterrorism agent, despite being declared eradicated in 1980[1]. The virus is retained in high security biosafety level 4 laboratories in the United States and Russia [2]. The variola genome is fully sequenced and could be synthesized in a laboratory [3]. This was previously thought to be unlikely, but Canadian researchers synthesized a closely related orthopox virus in 2017 and published the methods in 2018, thus highlighting the feasibility of de novo synthesis of smallpox [4]. Smallpox may re-emerge from deliberate or accidental release [5], and is a high-consequence event for which preparedness planning is needed [6]. Due to ageing, advances in medical therapies, transplantation and people living with immunosuppressive conditions such as HIV, the immunological status of the population has also changed dramatically since eradication of smallpox, with almost one in five people living with immunosuppression in Sydney, Australia [7]. A high proportion of people are unvaccinated, and vaccine-induced immunity in cohorts vaccinated before 1980 is waning [8–10]. In a modified SEIR deterministic model of smallpox transmission, we have shown that unprecedented rates of immunosuppression will result in increased morbidity and mortality of smallpox [7]. Planning for an epidemic of Planning for an epidemic of smallpox includes health system preparedness and resilience. In addition to vaccination, identifying and isolating cases to prevent further spread is influential in epidemic control [2].

To determine the capacity of the health system in Sydney, a city of 5.3 million people in Australia, during an epidemic of smallpox. Specifically, we aimed to determine hospital bedcapacity for isolation, public health workforce capacity for contact tracing and health care worker (HCW) personal protective equipment (PPE) requirements under different attack scenarios. We also aimed to test a worst case scenario among the range of possible attack scenarios and identify modifiable factors which would prevent a worst case scenario.

We constructed a modified SEIR model for smallpox transmission based on a model published in our previous study [7]. Model parameters and their estimation have been previously described [7]. We assumed that the virus has not been genetically modified and that there is minimal residual immunity in the population from previous vaccination, as described in our previous study [7]. We assumed an initial attack size of 100, 1000 or 10,000 infected. Case isolation was assumed to reduce transmission to zero [2]. Given antivirals would be commenced after diagnosis and isolation, we assumed this effect would only apply in the healthcare setting and would not add to interruption of transmission above the effect of isolation alone, with the main transmission risk being in the community for undiagnosed or early cases prior to hospitalisation. We assumed that antivirals would therefore have no effect on community transmission, acknowledging that they would likely reduce morbidity and mortality for treated cases. We estimated number of hospital beds needed to control the epidemic, PPE requirements for clinical HCWs and public health workers required for contact tracing, under different scenarios.

We estimated 14595 public hospital beds and 5618 private hospital beds in Sydney. We estimated there are 100638 clinical HCWs in Sydney, the majority (65%) aged 30–49 years old, 51% nurses and 18% doctors [27]. We estimated a public health workforce of 1500 nationally, with approximately 330 public health workers in Sydney.

In the case of a smallpox release in Sydney, a high-income, well-resourced city of over 5.3 million people, health system impacts may be substantial under some scenarios as shown in our model. We showed if smallpox arises overseas and is imported as a single case into Australia by travel, control will be far easier than under an attack scenario. We showed that influential factors on epidemic impact are the size of the initial attack, time to commencing the response, case isolation rates and contact tracing for ring vaccination. Whilst both are influential, case isolation is more influential than contact tracing. These public health interventions depend on physical and human resources, including clinical and public health workforce. Whilst the size of an attack may not be within our control, other the influential factors are modifiable and potentially within our control. If the initial attack size is 100–1000 and the response is rapid, an outbreak of smallpox can be controlled with case isolation, contact tracing and vaccination. However, if the response is delayed to 30 days or longer (which equates to about 2 weeks after the first symptoms occur), or if the attack infects 10000 people, epidemic control will be much more challenging, and the health systems impacts will be substantial. In the worst-case scenario, available hospital beds will be exceeded in less than 40 days. The requirement for hospital beds for isolation of cases will vary from up to 4% to 100% of all available beds depending on the size of initial release and speed of response. Even in the mid-range scenario of 1000 initial cases, up to 40% of all available hospital beds will be required for smallpox control. This does not account for the facilities required for quarantine of contacts, which must additionally be planned for, and in the worst-case scenario would require over 400,000 high risk contacts to be quarantined. Quarantine and isolation capacity are critical to epidemic control. Planning for surge bed capacity using available guidelines should be undertaken [30], and back up plans such as the use of community halls, school buildings, hotels or other large buildings should be made to ensure that that other viable isolation sites are pre-designated as smallpox treatment centres and available. During the 2009 pandemic of influenza, which was reportedly not as severe as expected, studies reported a tripling of patient presentations to hospital [31]. Plans for managing hospital bed capacity in the event of a large initial attack should also be made, including designation of specific treatment facilities, cancellation of elective surgery and decanting of patients with non-urgent other conditions into private hospitals or other facilities. The capacity for hospital beds for non-smallpox patients who require urgent hospitalisation must also be considered, and in some scenarios, the care of patients with urgent non-infectious conditions such myocardial infarction or stroke, may be compromised by lack of hospital capacity and staffing shortages.