Why we need an intelligence-led approach to pandemics: supporting science and public health during COVID-19 and beyond

Science is not intelligence

The dynamics and scale of the COVID-19 pandemic have stressed every health ecosystem: from social determinants to public health, basic and advanced clinical care as well as high-level policy decisions around resource allocation. There has been a huge call upon a wide variety of scientific domains from infectious disease modelling, risk stratification, medical countermeasures and health economics to provide the necessary expertise and evidence to inform so-called ‘science-led policymaking’. Yet the reality is that the pandemic has overtaken the ability of traditional scientific practice in many areas to deliver the necessary insights, resulting in substantial information asymmetry.

The philosophy of science from Karl Popper’s empirical falsification, to David Sackett’s evidence-based medicine reflects a methodological approach that requires time. The taken-for-granted Mertonian norms dictate the ethos of modern science: communism (common ownership of discoveries); universalism (claims to truth evaluated according to universal criteria); disinterestedness (reward for selflessness); and organised scepticism (ideas subject to rigorous community scrutiny). ¹ Changes are, by and large, gradually tested and introduced. Experience and expertise then ‘stress test’ these in the real world, with clinical and public health ‘natural selection’ determining whether successful practices and interventions become embedded or are discarded.

But this is when time is not the enemy, and where science meets permissive political, social and clinical conditions. What the COVID-19 pandemic has exposed is that contemporary science alone quickly reaches its limits when faced with policy requirements from a global pandemic that has transformed into a political and national security threat. COVID-19 demonstrates that scientific information is not intelligence. The nature of intelligence versus information is an essential distinction here. ² The intelligence cycle encompasses a number of domains, beginning with ‘direction’, in which intelligence requirements are identified and knowledge gaps recognised. The following steps are ‘collection’ of necessary information, ‘processing and analysis’ whereby the information is examined by specialists, and ‘dissemination’ to the necessary audiences. ³ This method balances uncertainty amid time-pressure, providing rapid situational awareness, exploitation (providing explanatory hypotheses that best fit the facts, and that have the least evidence against them), ongoing estimates and, finally, strategic notice (long-term horizon).

The ‘science-led’ approach to COVID-19 did not prepare countries adequately for the demands of a global pandemic and furthermore has exposed process deficits that have led to the failure to properly collect, analyse and disseminate the vast quantities and breadth of pandemic-related information. ⁴ The consequences for public health, economies and clinical practice around the world have been dramatic, from poorly implemented national lockdowns, to the failure to properly implement testing, through to dramatic changes such as closing routine surgical services and diagnostics. ⁵ Normative methods for scientific and clinical research have not delivered timely high-quality intelligence. Instead, publication-based dissemination has led to data either being put into the public domain too quickly without sufficient peer review or, paradoxically, too slowly. Despite the myriad of COVID-19 collaborative groups, academic competition has led to information fragmentation, research misconduct and the dominance of particular domains, e.g. modelling at the expense of other equally important methods. Harmonising the involvement of essential scientific insights with the scope of intelligence methods operating amid uncertainty acknowledges that rapidly spreading pandemics such as COVID-19 place exceptional pressures on information systems to produce timely and accurate assessments in the face of incomplete data.

Public health and the limits of the UK approach

The failure of the UK’s approach to the COVID-19 pandemic has manifested in uncertainty and fragmentation. The UK response has been characterised by a series of missed opportunities to deliver timely specialist public health practices, in tandem with shortcomings incorporating these into the broader architecture of domestic emergency preparedness. The initial critical failure has been the absence of concrete assessment of the actual threat. A repeated failure to establish a meaningful testing system to identify the underlying epidemiologic structure of the spreading outbreak lost vital time to understand the demographics of risk and to implement an effective and timely control system. At the time of writing (mid-June 2020), the government’s test and trace programme has only been in place for a few weeks. Over-reliance on modelling compared to real-world data on the prevalence of SARS-CoV-2 has characterised the UK’s approach to COVID-19, which runs directly counter to the early and sustained advice of the World Health Organization as well as the experience of ‘first wave’ countries. Lessons from countries such as South Korea, Iceland and Germany^6–8 have placed in the foreground the central importance of testing; knowing that the UK was anticipating COVID-19 following outbreaks in China, South East Asia and Italy, the lack of joint threat assessment intelligence has been a notable failure, which resulted in delayed and bluntly deployed containment measures alongside clinical resource redeployment without focus on a sharply delineated threat.

Issues of authority (i.e. a clear direction) in the UK reflect a series of command and control failures. The Scientific Advisory Group (SAGE) has acted as the primary advisory group to the government, with subsidiary groups such as SPI-B and SPI-M supporting further theme-specific portfolios. The delicate relationship between ‘the science’ and government action has been well publicised; however, it is clear that current institutional processes for pandemic management lack sufficient power to direct action necessary to fulfil the informational requirements of the designated expert hub. Despite SAGE advice on 16 March that emphasised the ‘critical importance of scaling up antibody serology and diagnostic testing’, ⁹ no such concrete plan was effectively implemented by government reflecting a substantial disconnect between advice and operations.

The overt disconnection between the NHS and Public Health England has illuminated the fact that existing pandemic response mechanisms are unsuited to the needs of an intelligence-led process for high-threat pandemics. At an early stage, the absence of an all-sources pandemic intelligence cell has been a major flaw exposing the rigidity of UK political culture when it comes to responding to time-critical population threats. The lack of real-time data on hospitalisations, intensive care unit bed capacity, useable ventilator numbers and COVID-19-confirmed versus ‘query COVID’ bed occupancy reflected an operational framework that was unable to deliver actionable intelligence and have sufficient authority to link this directly to tasking operations.

There has been a notable lack of Human Intelligence (HUMINT) during the pandemic response; the dearth of real-world data has made it challenging to contextualise emerging information and calibrate it with real-time signal. As part of the analysis stage, the role of ‘red-teaming’, a system for testing an organisation’s detection and response capabilities is an important exercise in determining system integrity and identifying unforeseen vulnerabilities. Such systems again have been absent in the roll-out of novel clinical and public health protocols despite a huge mobilisation of government and health system resources, surveillance technologies and exceptional public measures. Finally, the dissemination of information from government has tended to focus on the media rather than the NHS, or directly to the public. Other countries such as South Korea have shown the way by identifying dissemination strategies that manage multiple vectors of communication to specialist and lay groups, all of whom have been required to modify their practices in line with intelligence estimates. ⁶ Although in the UK the Chief Medical Officer, Chief Scientific Officer and Deputy Medical Officer have been present at press briefings, the limitations placed on them by political actors have driven a lack of clarity over the distinction between scientific advice and over-riding political imperatives. Such communications have been ad hoc and politically determined rather than existing as a structured mechanism of the pandemic intelligence machinery.

Towards an intelligence-led approach

UK pandemic preparedness and response to COVID-19 has been a failure. We have experienced one of the worst levels of COVID-19-related morbidity and mortality in the world and, in addition, the manner in which non-pharmaceutical interventions (i.e. lockdown and contact tracing) were implemented, and the economic downturn will result in significant long-term death and disability from diseases such as cancer and mental health. These failures are magnified in the context of a history of NHS fragmentation and politicisation. Repeatedly, the issue of the integration of hospital IT systems has been raised as an important goal for clinical efficiency; however, central government has failed to attend to the required technical and investment needs that would permit real-time oversight in times of crisis. Disconnection via the outsourcing of specialist services has compounded this and been made visible again in the awarding of the test and trace contract to Serco, a private non-specialist general services provider. As focus in recent years has been placed on efficiencies and cost-savings, lessons from other countries and our own exercises such as Operation Cygnus have failed to motivate political actors to invest in a robust intelligence and operations approach to pandemics. These deficits in public health have been mirrored in the security sector where we have seen a steady decline in biosecurity and biointelligence capacity and capability since the Butler Review into Iraq's weapons of mass destruction in 2004. ¹⁰ This has led to serious lacunae in professional intelligence experience and expertise in the field of biosecurity.

Robust science and public health are essential for future pandemic preparedness and integrating these domains within an intelligence-led approach is a pressing imperative. The new Joint Biosecurity Centre may hold promise; however, considerable attention is necessary for this initiative to match the scope and influence of, for example, the Joint Terrorism Analysis Centre (JTAC) which delivers services across government departments and agencies, determines threat levels and coordinates investigations. Such an approach would be a crucial step-change to harness the strengths and expertise of UK intelligence while embedding multi-lateral cooperation across Five Eyes partners. ¹¹ COVID-19 has dramatically exposed the fatal limits of our current methods; future public health threats will necessitate agile systems for generating robust actionable intelligence that can operate in conditions of significant uncertainty and complex externalities.