Pandemic simulations and modelling
Pandemic risk inevitably remains largely underspecified and unspecifiable. A pandemic is a likely and yet a highly uncertain high-consequence event. The current chapter looks at how in practice these high margins on uncertainty' are handled by' policymakers and practitioners - including medics, scientists, veterinarians, climate modellers, urban planners, behavioural scientists, and logistics experts. It discusses the strategies deployed to translate projections into actionable plans, following a move towards an ‘all-hazard approach’ to preparedness. The latter aims to achieve flexible strategies but results in vague and underspecified plans of action. The tension between keeping complexity in the picture, to account for multiple outcomes of infection and factors, multiple actors and impacts of a pandemic, and achieving a simplification of the picture, for the sake of agility and transferability of plans from one crisis to the other, means that important dimensions are overlooked.
In this chapter, we discuss the increasing reliance on modelling and the deployment of various forms of simulation exercises. We analyse the strengths and limitations of these tools, identifying the untested assumptions inbuilt in current practices and the negative impacts they can produce. The aim of the chapter is to encourage a more considered and critical approach to these tools, with a view to achieving better ways of balancing the tension between aiming for simplicity' of plans and keeping in sight the complexity of pandemic situations and response.
Vague plans and empty networks: the impasse of all-hazard preparedness
The 2009 swine flu pandemic showed us that countries need to be able to escalate and de-escalate their response more flexibly than in the past, given that the effects of that pandemic were uneven and generally milder than anticipated through modelling forecasts (Mansnerus 2015). They also came to affect different regions differently and at different times. In a pandemic, countries must no longer be locked into a type of full-scale, all-out response at all times (WHO 2013, 2017b). Flexibility across countries and within a single country has been shown to be of paramount importance.
This lesson has driven countries towards adopting a new ‘allhazard approach’ to preparedness and planning. In Chapter 2, we illustrated the point with reference to specific policies and plans (see section 2.6) and showed that this way of practicing preparedness is informed by guidelines diffused nationally and internationally (see also Lakoff 2008; Holmberga and Lundgren 2018; Public Health England 2014a, 2014b; Hoffman 2013; Canada 2019). A one-size-fits-all approach is no longer understood to be helpful (WHO 2013, 2017b). Instead, the WHO concluded that responses need to remain flexible and alert to the need to scale up and scale down measures regionally (Ibid). This advice has been cascaded as best practice and has been informing the updating of pandemic preparedness plans post 2009 (ECDC 2017; Holmberga and Lundgren 2018; Public Health England 2014a, 2014c, 2014d).
The all-hazard approach is predicated on the basis that plans can then become transferable and deployable in a wide range of circumstances. The same plan might perhaps be invoked against a wide range of pathogens and types of crises (from naturally occurring outbreaks to deliberate releases of biological agents). In principle, it can be deployed notwithstanding the fact that the mode of transmission of different pathogens will vary quite substantially - and so will the measures to be devised to arrest or delay transmission. Transmission of a respiratory disease caused by a pandemic influenza strain will not be stopped via isolation and quarantine (at best it can be somewhat marginally delayed), since it is diffused via airborne means including in aerosol format and in very small droplets and may occur also when asymptomatic, during the incubation
Pandemic simulations and modelling 131 period. A respiratory disease like SAKS, in contrast, only becomes infectious after onset of symptoms, and therefore could be more easily contained. A haemorrhagic fever like Ebola requires contact with bodily fluids and infected surfaces to spread, so infection could be arrested by applying extremely stringent but traditional methods of infection control. In contrast to flu, however, which has a 2 to 5-day incubation period on average, Ebola’s incubation period is up to 21 days, and therefore it poses different challenges for detecting the infected and for tracing their contacts.
Tempting as it is to rely on an all-hazard approach and devise plans that are believed to be flexible enough to be mobilised in a wide range of circumstances, the result is that the ‘plans’ developed are incredibly vague and underspecified (Hoffman 2013; Lakoff 2008; Canada 2016; Canada 2019; Wolf 2016). Canada (2016, 2019), who focused on preparedness against biological threats in Finland and other Nordic countries, and Wolf (2016), who investigated pandemic preparedness in Germany, both lament the vagueness and emptiness of the supposed preparedness plans in these countries, which effectively do not provide clear answers or responses to a set of circumstances when they materialise.
Further layers of uncertainly linked with the vagueness of plans arise also in relation to the triggers for moving from one stage of response to the next above or below. If these triggers are defined quite generally and opaquely, confusion might follow in implementing measures. Flexibility for escalation and de-escalation may engender the risk of uneven response even when this is not required. For example, it is often unclear what functions may need to be taken up centrally (i.e. nationally, and in collaboration with any devolved authorities), regionally and even locally at the city scale (Hoffman 2013; Teo, Yeoh and Ong 2005; London Resilience Partnership 2014). Uncertainty also exists as to how evenly and rapidly the plans can be rolled out by different localities (Hoffman 2013). Regional and city-wide responses can vary from being extremely well-structured and formalised, in large capital cities like London for example (London Resilience Partnership 2014), to being practically non-existent. Most other cities in the UK, for example, do not have a city-specific pandemic plan in pre-pandemic times. Instead, the plan they will follow is to implement locally the national guidelines.1
Vague plans are envisaged to be fleshed out when a specific outbreak occurs. They are then to be made into more tailored plans for various sectors and groups to ensure business continuity, tackle high disruption or absenteeism and ensure essential service provision is unaffected - for example this involves drafting plans for educational settings, prisons or workplaces, transport or waste disposal services. Guidance and planning advice from international bodies suggest forming multi-agency plans, which involve a range of responders — from primary healthcare personnel to first responders like police and fire service.
Plans and practical guidance for and with various sectors are much better specified for pandemic flu but absent or underspecified for other pathogens. In the UK, for example, ahead and during the 2009 swine flu pandemic, a veritable flurry of plans was released to direct action in sectors including cleaning services and refuse collection, funeral and burial services, the hospitality industry, fire and rescue services, healthcare settings, businesses and other occupational settings (UK Department of Health 2008a, 2008b, 2008c, 2008d, 2009a and 2009b). Yet, the same cannot be claimed in relation to subsequent infectious disease threats with pandemic potential, like Middle East Respiratory Syndrome (MERS). For the coronavirus induced MERS first identified in 2012, hospitals and healthcare first responders tended to have guidelines and plans in place (including through simulations exercises — see Central Manchester University Hospital and NHS 2015), but other agencies or sectors did not develop plans for dealing with this respiratory novel infectious pathogen.
To some extent, therefore, previous pandemic flu plans can act as a blueprint in different pandemic contexts. They too can be considered adaptable to new circumstances and, as such, they complement existing vague pandemic plans. But this is not always a successful translation. For example, during the early stages of the COVID-19 infection, laws were drafted in the UK about various measures including quarantine and safe burials and funerals. Because they were based on the blueprint of pandemic flu plans, the proposed legislation envisaged body cremation as the only safe option. This option proved controversial, since it disregarded the cultural practices and preferences of various ethnic and religious communities in the UK, and since it was unfounded on epidemiological and logistic
Pandemic simulations and modelling 133 grounds in the case of the COVID-19 infection. As a consequence, it was contested in Parliament and amended out of the legislation that has now been introduced (House of Commons 2020).2 Despite the fact that an effective solution was in the end found and voted in (Coronavirus Act 2020), the case illustrates some of the pitfalls of relying on plans for different pathogens and in adapting vague and underspecified plans under pressure and time constraints.