From a scientific standpoint, the world is more prepared than ever to respond to infectious disease outbreaks; paradoxically, globalization and air travel, antimicrobial resistance, the threat of bioterrorism, and newly emerging pathogens driven by ecological, socioeconomic, and environmental factors, have increased the risk of global epidemics.Footnote ^1, Footnote ^2, Footnote ³ Following the 2002–2003 severe acute respiratory syndrome (SARS), global efforts to build global emergency response capabilities to contain infectious disease outbreaks were put in place.Footnote ^4, Footnote ^5, Footnote ⁶ But the recent H1N1, Ebola, and Zika global epidemics have shown unnecessary delays and insufficient coordination in response efforts.Footnote ^7, Footnote ^8, Footnote ^9, Footnote ¹⁰ In a thoughtful and compelling essay,Footnote ¹¹ Thana C. de Campos argues that greater clarity in the definition of pandemics would probably result in more timely effective emergency responses, and pandemic preparedness. In her view, a central problem is that the definition of pandemics is based solely on disease transmission across several countries, and not on spread and severity together, which conflates two very different situations: emergency and nonemergency disease outbreaks. A greater emphasis on severity, such that pandemics are defined as severe and rapidly spreading infectious disease outbreaks, would make them “true global health emergencies,” allowing for priority resource allocation and effective collective actions in emergency response efforts. Sympathetic to the position taken by de Campos, here I highlight some of the challenges in the definition of severity during an infectious disease outbreak.

De Campos asserts that the discussion about accepted definitions of pandemics is not only semantic. In fact, declaring a disease outbreak a ‘pandemic’ has vast global and local implications beyond planning and response efforts, including political, social, and economic consequences.Footnote ^{12 ,} Footnote ^{13 ,} Footnote ^{14 ,} Footnote ^{15 ,} Footnote ^{16 ,} Footnote ¹⁷ A true pandemic, she argues, would need two components: transmissibility and severity. Thus, noncommunicable diseases independent of their severity, such as obesity,Footnote ¹⁸ or tobacco addiction,Footnote ¹⁹ and nonsevere communicable diseases, should not be considered global health emergencies in the context of pandemic preparedness and response. A pandemic would thus be characterized by sustained transmission across countries,Footnote ²⁰ and a high impact on mortality and morbidity.Footnote ²¹ In this context, disease spread and severity would establish whether an infectious disease outbreak is a global health emergency or nonemergency, and this would “avoid definitional ambiguities that generate moral confusions.”Footnote ²² Unfortunately, determining the severity of an outbreak is not always straightforward.

The severity of an infectious disease outbreak can be characterized as a function of the virulence or pathogenicity of an agent to cause disease, and on its transmissibility, often operationalized as the expected number of new cases generated by an infected individual. Transmissibility is affected by the modes of transmission of an agent, susceptibility or exposure of potential hosts, and/or the duration of patients’ infectivity. But in some ways, the ‘severity’ of a disease also depends on the characteristics of local communities, including early detection and response capabilities, healthcare facilities, health workers, preventive technologies, communications, and community awareness. For example, a quick diagnosis, adequate isolation, and good infection control practices can be the difference between life and death for an Ebola infected patient, as was shown by Ebola patients treated in the USA during the 2014–2016 West Africa crisis. The large majority of those patients survived.

Perhaps the exclusion of severity in the definition of pandemics reflects the difficulties of estimating the real impact of an outbreak early on, when the local and global communities need to provide a timely and effective response, and does not reflect a lack of political will, of understanding the importance of a clearer definition, or even political and financial interests. Modellers working in emergency response typically need to predict the expected size, duration, and effects of an outbreak at an early stage. Some of the challenges they face include poor data quality and limited data availability (for example, inadequate epidemiological and clinical surveillance, reporting delays, poor epidemic research governance), short turnaround times, difficulty in communicating results (for example, interpretation of estimates, model assumptions, uncertainty), insufficient understanding of the pathogen and its transmission dynamics (for example, natural history of disease, modes of transmission, human mobility and exposure, contact networks), limited or unavailable specific and sensitive diagnostic tests, uncertainty in the effectiveness, impact, and costs of countermeasures, and several layers of complexity in international collaboration.Footnote ^{23 ,} Footnote ^{24 ,} Footnote ^{25 ,} Footnote ^{26 ,} Footnote ²⁷ Reliable data are essential for determining outbreak severity, but are often not available.Footnote ²⁸

Most resources for detecting and responding to emerging disease outbreaks are globally allocated in high-income countries, where major infectious disease outbreaks are less likely to occur.Footnote ²⁹ For example, in the past decade the World Health Organization has reported about 300 infectious disease outbreaks, from limited local outbreaks to global events involving several thousand deaths across various countries, such as the 2014–2016 Ebola outbreak in West Africa.Footnote ^{30 ,} Footnote ³¹ Most of these infectious disease outbreaks have occurred in low- and middle-income countries, often in areas with limited disease surveillance and response capabilities.Footnote ³²

There is thus substantial uncertainty about the scope and duration of most epidemics, and of their potential effects, particularly during the early phases of the outbreak when response efforts need to be planned and coordinated, and funding secured. Understandably, there might be a tendency for public health officials and/or disease experts to err on the side of caution, and overestimate the expected impact of epidemics. But at least for modellers, there is evidence for both overestimating and underestimating the size of an epidemic, largely because emergency response requires estimates at the emerging phase of an outbreak, when data may be sparse or of poor quality.Footnote ^{33 ,} Footnote ^{34 ,} Footnote ³⁵

Administrative, logistical, scientific, and political complexities are part of any local or global emergency response; yet the world lacks a global institution to coordinate response efforts, with corresponding funding, and authority to make decisions.Footnote ³⁶ De Campos highlights the role of International Health Regulations (IHR), a multilateral treaty adopted by all members of the World Health Organization “to prevent, protect against, control, and provide a public health response to the international spread of disease.”Footnote ³⁷ The IHR went into effect in 2007, and required countries to generate capabilities, including human resources, disease surveillance, emergency management, and laboratory infrastructure, to be able to rapidly detect and effectively respond to infectious disease outbreaks.Footnote ³⁸ But the H1N1 influenza in 2009 and, five years later, the 2014–2016 Ebola crisis showed that few countries had complied with IHR.Footnote ^{39 ,} Footnote ⁴⁰ These epidemics thus highlighted the urgent need for a response system for public health emergencies.Footnote ^{41 ,} Footnote ⁴² The lack of a coordinating institution adds uncertainty about the global capacity to set up a coherent, effective, organized, and sustained response, and limits the usability of severity-based definitions. Achieving adequate local and global preparedness to set up a proportionate precautionary response requires a reasonable characterization and understanding of an epidemic. Probably the most sustainable way of doing this is by strengthening local public health systems, including healthcare infrastructure, health personnel, disease surveillance, and laboratory capabilities. Improving local infectious disease detection and response preparedness capabilities not only helps prevent and stop an outbreak before it becomes a pandemic, it also promotes population health and well-being.Footnote ^{43 ,} Footnote ^{44 ,} Footnote ⁴⁵ Epidemic responses could then be adapted to local conditions of disease transmission and severity.Footnote ⁴⁶

Recent infectious disease outbreaks have highlighted the increasing risk of pandemics in a tightly interconnected world.Footnote ^{47 ,} Footnote ^{48 ,} Footnote ^{49 ,} Footnote ^{50 ,} Footnote ⁵¹ Despite adoption of the IHR treaty, many countries do not have the capabilities needed to rapidly detect and effectively respond to infectious disease outbreaks. Further, there are substantial gaps in the scientific knowledge required to characterize the threat of emergent pathogens—such as the Zika, SARS, Ebola, or influenza viruses—with confidence, and often available data are inadequate for accurately estimating the potential spread and severity of an outbreak at an early stage. I concur with de Campos’ thesis that the standard definition of pandemics may be inadequate for both theoretical and practical purposes, and that a combination of both transmission across countries and severity (i.e., high morbidity and mortality) of an outbreak would allow distinguishing between emergency and nonemergency situations. I have tried to highlight some of challenges toward the goal of achieving this conceptual clarity, not to discourage the goal, but to help set a plausible pathway. As illustrated by Peter DoshiFootnote ⁵² for the influenza outbreaks that have been labelled ‘pandemic,’ it can be challenging to standardize severity. It can be even harder to predict it.