Quelling the Great Fire
National Architecture to Support Local COVID-19 Responses
When a fire breaks out whosoever discovers the event raises the alarm, this is common practice and it works. It is the responsibility of the leadership of the affected organisation or building to evacuate the area and for the emergency services to mount a credible response before it gets out of control. For rapid onset events such as fires, decision making cannot be pushed up the chain of command, because time is of the essence.
During a fire there is no time available to assess whether the alarm was justified or what type of response might be needed until the evacuation is completed and the emergency services are on the scene. Through this well known set of standard operating procedures, evolved over two centuries or more, we are able to minimise the risk of fire in our communities, despite the presence of a multitude of hazards from gas stoves to nuclear power stations.
The mechanism to control a pandemic effectively, despite the ongoing presence of risk, should be similar. COVID-19 and other coronavirus infections will remain present within our population and future outbreaks will require very rapid action if they are to be effectively identified and contained. This requires providing a ‘fire alarm’ (trigger for action) as close as possible to the outbreak - within reach of the person who identifies it first - and ensuring responsibility for managing the containment measures is driven down to a rapidly deployable on-site local response command.
Unlike a fire, the identification of a coronavirus outbreak is not immediately obvious. Individuals might experience symptoms, or witness members of their family, colleagues or social circle suffer from symptoms. Those feeling unwell can elect to be tested at home or in medical facilities or by community testing teams.
Yet we know from the experience of Leicester City Council that information on people who suspect infection, or even those who have tested positive for the virus, is taking far too long to reach those with responsibility to mount a response to contain a potential outbreak. Indeed, unlike other emergency responses, there is a lack of clarity about which agencies and individual responders should be deployed to mount an effective response. We have no standard operating procedures to respond to a COVID-19 alarm.
Yet many of the building blocks of a containment system have already been put in place. What is missing is a well-drilled organizational model to harness these to a community level ‘response system’ to IDENTIFY new infections effectively, raise the ALARM, rapidly RESPOND and effectively CONTAIN any potential outbreak. I say potential outbreak, because any incidence of infection has the potential to become a full-blown outbreak, as any wastebasket fire has the potential to become an inferno if left unchecked.
Private sector self-reporting data collection like The Key’s https://corona-help.uk/ can provide real-time data on the incidence of COVID-19 symptoms.
What are the building blocks? Bearing in mind that like a response to a fire, time is of the essence, we need to:
IDENTIFY new infections. The best way to identify infection is through testing. Yet testing can take time and requires the affected person to seek or be sought out by testing teams, either online or via a medical professional or testing station. The prior stage is self-reporting, where an individual reports their own symptoms, or those of someone close to them. Community testing can help to capture asymptomatic infections too, as a matter of course during medical or dental appointments for example. This can be extended in high risk environments, say to users of a gym or sports club, or by random testing in shopping malls or areas of towns and cities with many bars and restaurants.
Raise the ALARM. In many ways this is the most complex and least developed element of the containment system. Who can raise an alarm? What is the criteria for a response? Using the fire analogy, anyone identifying an uncontained hazard should be able to raise an alarm, although there should also be significant legal penalties for deliberately raising a false alarm. Can this modality be applied to coronavirus infection? Clearly one person reporting cold-like symptoms is insufficient to justify a community-level response, however a cluster of such reports may merit rapid investigation and focussed testing. Positive tests require follow-up both in terms of contact tracing and enhanced public health warnings and advice in affected localities. Data processing algorithms can be linked to self-reporting systems (via self-reporting apps such as The Key’s Corona-Help.UK or online reporting instruments such as NHS 111 or similar automated phone lines) to trigger alarms if certain criteria are met, i.e an abnormal cluster of symptom reports. These alarms then need to be linked to a well-trained response mechanism with agreed standard operating procedures.
Rapidly RESPOND. Once an alarm is raised a localised emergency response should be triggered immediately as there is no time to lose. This means a well-trained response system needs to be in place with properly designed and tested operating procedures. An incident commander needs to be available to provide onsite leadership with the authority to command a group of disparate response resources and the legal instruments to enforce the lockdown of a local neighbourhood temporarily (until conditions are met that indicate the hazard has been contained). Again a good analogy would be the incident command system operated by the UK Fire and Rescue Service.
CONTAIN any outbreak. The response will normally require the localised lockdown of a small locality around the cluster of reported or proven infections, rapid testing and contact tracing and the isolation and monitoring of affected people. This requires the police, health and care services, PHE test and trace and local authorities to act in unison under a single but localised command structure for the duration of the event. Many events (like many fires) will either be rapidly contained or prove to be of low risk. Nevertheless the system will pick up those events that could lead to dangerous outbreaks too. The establishment of legal instruments to enable these incident teams and commanders to contain the virus and to discourage deliberate false alarms are critical if the system is to operate effectively, as is the development of a national architecture for the delivery of these localised responses (the fire service is again a good analogy).
____________________________________
The Key Data’s 3TP concept is a layered response system for COVID-19 containment
The evidence of fire fighting shows that it is clearly a mistake to develop a national command structure for COVID-19 responses: what we need is a national support structure for locally commanded responses. COVID-19 outbreak containment needs to be rapid and local. They can’t wait for decisions to filter up the chain of command to Whitehall. Nevertheless a national architecture is required to ensure these localised responses have the tools to do the job: the organisation, authority, data, skilled and trained personnel and proven operating procedures to manage the pandemic effectively at the local authority level.
The mistake we have made so far is to centralise control of response resources rather than focusing on building the capacity and infrastructure for those responses at the centre of government. Devolved incident commands need to take charge of local responses and manage response teams in every local authority area. Bodies such as the newly created Joint Biosecurity Centre and Test and Trace in DHSC need to be strategic, focussing on collecting and delivering the data in real-time, developing the capacity, systems, organisational models and standard operating procedures and providing funding for highly localized emergency responses.
The problem is that the aptitudes needed to conceive and lead these types of civil defence organisations lie in areas which have so-far been peripheral to the response - the emergency services and armed forces. While the specialised skills needed for a COVID-19 response lie within institutions not normally engaged or familiar with emergency response modalities: civil service institutions such as the DHSC, Public Health England, NHSx as well as local government authorities. Models developed for the execution of highly consultative and mostly non-time critical public health measures or the implementation of central or local government policies and systems are simply not fit for purpose in a rapid emergency response environment.
Yet in a democratic system the last thing we want to do is to centralise ‘emergency powers’. Especially where the risk is in reality highly localised and the high threat levels temporary, even if an ongoing lower level hazard exists. One of the problems we face is that the most recent set of threats encountered have been related to terrorism from within our communities, and it is from the counter-terrorist field - from the Home Office and Security Service and related tech companies - that the core of the Joint Biosecurity Centre team have been drawn.
Counter-terrorist responses have been highly reliant upon covert surveillance and the resulting need for secrecy, and as such have been shadowy and centralised. COVID-19 is the opposite end of the spectrum from terrorism. No-one has a desire to conceal or encourage infection, in fact the more people who are aware of the threat and have access to the surveillance data and can trigger mechanisms for response the better. While privacy and data protection concerns remain crucial to create public trust in the system, the culture of secrecy around government surveillance that has emerged from two decades of counter-terrorism response needs to be resisted by JBC in our efforts to gather early warning and forecasting data for responders. This need is in order to ensure a return to normal life in the presence of COVID-19.
The failed NHSx app was probably an outcome of this counter-terrorism legacy. The system was over-engineered as both a very intrusive surveillance system that could collect data for forecasting and as a proximity tracking contact tracer to provide an alert system for both individuals and the national response. In the end, although it was conceived as a silver bullet, it proved to be a single point of failure.
Because the NHSx app specification involved the collection of highly sensitive personal data (as has become the norm in counter-terrorism surveillance) the app’s data processing system needed high levels of data protection, and thus it is unclear if much of this data could have been easily made available in real-time to local responders. Sophisticated data processing and cleaning and layers of vetting and permissions would have been essential, and although this could be automated, data would still have needed to be highly protected by NHSx to prevent privacy breaches. As a result data would have probably not been available in real-time to contain outbreaks and the app’s primary role of contact tracing was fatally damaged by the intrusive nature of its data collection function, excluding it from the Apple and Google API, and thus damaging its functionality on iPhones.
In fact a layered response system was required, with plenty of built-in redundancy combining multiple and open sources of data and both manual and automated contact tracing systems. Most of these privacy-centric data collection systems had been built by the private sector and academia and were freely available, and yet this data source was largely ignored by the government. Since the failure of the app, the Department of Health and Social Care (DHSC) has instead tried to shoehorn the app's functions into a manual test and trace programme, which manually collects data to feed into the centralised and automated data-processing and forecasting system built for the app. As a result all the same flaws with the app are apparent, and data is not reaching local responders in real-time so they can organise effective local responses, as was clearly apparent in Leicester.
________________________
If we want an effective system to minimise the risk of COVID-19 outbreaks in our communities we need to look for local-response models to our existing and proven civil defence structures, in particular the fire and ambulance services, melded with the excellent technical resources of JBC and Test and Trace. These need to be enhanced by open-source real-time data collection and an effective alert system, informed by the wider tech industry and all of the excellent non-governmental systems that have been developed to help provide early warning and support contact tracing. There is no silver bullet here. We need an effective system that draws upon all of the available and trusted data collection and contact tracing resources.
James Fennel MBE is the Founder and CEO of The Key Data Limited https://thekey-partners.com/ which has been supplying real-time data via jHub and Project OASIS to the NHS since March 17 2020. He has over 30 years global experience in crisis management with international humanitarian organisations, private security contractors and as a Regional Conflict Adviser for the British Government.