Building a better global health security early-warning system post-COVID: The view from Canada

The COVID-19 pandemic in the People’s Republic of China (PRC) and its rapid spread in 2020 has laid bare the weaknesses and vulnerabilities of the existing international system for monitoring major disease outbreaks. A powerful globalized disease met a deeply flawed globalized response mechanism. The failure of that response begs important questions about what went wrong and what work might be ahead to provide a stronger collective security framework for global health.

The international system in place when COVID-19 struck was of long standing, and was the product of lessons learned from an earlier coronavirus outbreak—the Severe Acute Respiratory Syndrome (SARS) epidemic of 2002–2003. SARS was the first global health emergency of the twenty-first century. In the aftermath of SARS, where timely and accurate health information reporting proved in short supply, especially with regard to the original outbreak in the PRC, major efforts were made to build a new international process for monitoring and reporting on disease outbreaks and to embed it in international law. The new approach was meant to be comprehensive in its reach and scope, incorporating natural biological outbreaks as well as man-made biosecurity threats from chemical, biological, radiation, and nuclear sources, and was even linked to health-related responses to natural disasters and climate change impacts. ¹ Canada was an important early shaper and contributor to the global health surveillance system, based in part on its own experiences of SARS.

The post-SARS system operated at two inter-connected levels—the international and the national. The World Health Organization (WHO) was placed at the pinnacle of the system and was to be supported by national systems for disease outbreak reporting and monitoring. Information flow was meant to be two-way, with national health systems reporting on disease outbreaks to the WHO and the WHO monitoring national outbreaks, interfacing with countries affected by disease outbreaks and reporting globally. The system depended on both national capabilities as well as the WHO resources. It also depended on strong international collaboration and on information transparency.

The teeth in the system were provided by the International Health Regulations (IHR), which were fundamentally revised in 2005. ² The IHR had existed since 1969, but its scope was previously limited to a handful of listed communicable diseases, and reporting on outbreaks by WHO member states was often treated as discretionary. The 2005 revision was meant to both broaden the scope of IHR coverage and remove discretionary reporting. It did so by creating mandatory responsibilities for both the WHO and member states for disease outbreak reporting, processes for information sharing, and thresholds for the declaration of global outbreaks. One recent commentator argued that the IHR represented a “grand bargain”—real transparency with regard to national disease outbreaks in exchange for protections against arbitrary imposition of travel sanctions and other national costs. ³ What the IHR did attempt was to weave together national capacities for disease surveillance with an international system capable of a genuine “collective security” response.

The various articles of the IHR spelled out the intended working of the global health surveillance system in detail. ⁴ Member states were to develop surveillance capacities to “detect, assess, notify and report events …” ⁵ It was incumbent on states to notify the WHO of any disease outbreak that might constitute a “public health emergency of international concern” within 24 hours of its identification. ⁶ This requirement was amplified in Article 7 of the IHR, which referenced the case of an “unexpected or unusual public health event” within a country’s territory and which required the provision to the WHO of “all relevant public information” about the event. ⁷ Following IHR mandated notifications, states were to continue to communicate “timely, accurate and sufficiently detailed” public health information to the WHO. ⁸ The mechanism for information exchange involved the creation of national IHR “focal points” that would be responsible for the timely exchange of data with the WHO.

The WHO, in turn, would serve as the central resource to inform all member states about disease outbreaks (Article 11) and to provide assistance. The IHR also embedded a new concept of a “Public Health Emergency of International Concern,” which was meant to be a cornerstone for a coordinated global response to a major disease outbreak. This concept was seen as so important to global response capabilities that, while it was based on a hoped-for agreement between the WHO and a state suffering a disease outbreak, there were circumstances in which the WHO director general could proceed independently to make such a declaration, with the consent of a specially constituted “Emergency Committee.” While there were specified criteria laid down for such an independent determination, the available information and a risk assessment were key. The risk assessment involved three considerations: risk to human health; risk of the international spread of disease; and, in recognition of the likely spill-over effects of any such declaration on national interests, “the risk of interference with international traffic.” ⁹

The WHO Director General, Tedros Adhanom Ghebreyesus, has already signalled that changes are needed for declarations of a “Public Health Emergency of International Concern,” which made headlines when adopted by the WHO on 30 January 2020 but appears to have made little impact on the global response to COVID-19. ¹⁰ Critics have called, instead, for a more flexible multi-tiered alert system, analogous to counter-terrorism practices in various Western countries. ¹¹

In the IHR, the commitment to the international exchange of health security information through the WHO, and to a supreme coordinating role for the WHO for managing an international disease outbreak, was buttressed by one additional and unusual instrument. While the provision of information about disease outbreaks stressed a requirement for official reporting by states, there was a recognition built into the IHR that circumstances may arise in which other forms of early warning might be useful if a state experiencing an outbreak was either unwilling or unable to provide timely information. The IHR drafters may have been thinking primarily of states with weak governance and public health systems, but IHR Articles 9 and 10 were equally applicable to authoritarian regimes that might practice tight, centralized control of information and various forms of censorship. These provisions of the IHR established a parallel stream of early warning derived from what were called “other reports.” Other reports encompassed information from sources other than official state notifications.

Where were such “other reports” to come from? That was not spelled out explicitly in the IHR. But the IHR drafters were fully aware of both state and private sector capabilities for open-source reporting. Established non-state public health early warning networks used media scanning, such as “Pro-Med.” ¹² A pioneering and unique information gathering system was also developed by Canada, in collaboration with the WHO, called the Global Public Health Intelligence Network (GPHIN). Article 9 was an implicit recognition of the value of early warning platforms, such as Pro Med and the GPHIN. The same article also spelled out the responsibility for nation-states, “as far as practicable,” to “inform the WHO within 24 hours of receipt of evidence of a public health risk identified outside their territory [emphasis added] that may cause international disease spread.” ¹³ The Canadian GPHIN system may have been a direct source of inspiration for this requirement. A global watch responsibility was meant to be another element in an international system of early warning. It also opened a possible role for national intelligence systems to contribute to global health early warning, where mandates, capabilities, and national interest allowed. At the time of ratification of the IHR, when member states were able to indicate national “reservations” and interpretations (which the PRC did by declaring its authority for public health reporting from Taiwan), only the United States appears to have paid attention to Article 9 and its implications, by indicating that the country could find Article 9 obligations “impractical” with regard to any notification “that would undermine the ability of the U.S. armed forces to operate effectively in pursuit of US national security interests.” ¹⁴

The IHR created an international system for disease outbreak reporting that was designed to learn lessons from SARS, boost requirements for national reporting, and create a functioning and robust global system for information exchange. It rested on four pillars: WHO capabilities; national reporting of outbreaks; global community monitoring; and good use of “other” sources. In the face of COVID-19, a provisional assessment suggests that none of the pillars was as effective as might have been hoped for by the IHR founders. This leads to questions about how the IHR system might need to be overhauled across all four pillars.

Analyzing the four pillars of the WHO IHR system involves a large research agenda that awaits multiple studies and reviews. Certainly, the review of the WHO by an independent panel announced on 9 July 2020, will serve as one future and significant contribution. ¹⁵ An interim report was expected in November 2020 with the full report due in May 2021. The resolution of the World Health Assembly, consisting of the member states of the United Nations (UN), adopted on 19 May 2020, initiated the lessons-learned review, and called attention in particular to:

The effectiveness of the mechanisms at the WHO’s disposal

The functioning of the International Health Regulations

The actions of the WHO and their timelines pertaining to the COVID-19 pandemic. ¹⁶

While we await the reporting from the independent panel evaluating the WHO as well as what may become a tsunami of national lessons-learned reviews, it is possible to explore in a preliminary fashion the inter-connectedness between Canada’s health surveillance system and the global architecture under the IHR. This is an important topic not only because it illuminates a national capability but also because for many years Canada has tried to position itself as a leading contributor to global health surveillance initiatives.

The key agency that links the Canadian federal public health system and the WHO is the Public Health Agency of Canada (PHAC), which reports to the health minister and is a separate entity within a federal department, Health Canada.

The origins of both the PHAC and Health Canada date back to efforts to learn lessons from SARS, particularly in regard to better coordination at the federal government level of health measures in response to disease outbreaks and a stronger capacity to interface effectively with provinces and territories, which have jurisdiction over many aspects of health policy delivery. ¹⁷ The decision to make these organizational changes also intersected with the development of Canada’s first national security policy.

The national security policy, “Securing an Open Society,” published in April 2004 described the development of an integrated security system “capable of responding to both intentional and unintentional threats. It is as relevant in securing Canadians against the next SARS-like outbreaks as it is in addressing the risks of a terrorist attack.” ¹⁸ The national security policy promised to ensure that that public health issues would “figure prominently in the Government’s integrated threat assessments.” ¹⁹

These promises may have been excellent, but they were never realized after 2004. A new threat assessment unit was launched, initially called the Integrated Threat Assessment Centre (later renamed the Integrated Terrorism Assessment Centre), housed at the Canadian Security Intelligence Service. But it never developed any health security-related expertise or reporting. The new public health organizational structure, created in 2004, remained disconnected from other elements of the Canadian security and intelligence system, and health security issues remained outside the orbit of security and intelligence reporting. ²⁰

The new PHAC was left, in essence, to go it alone in strengthening Canadian capacity for global health surveillance within the federal government system. Canada signed on to the IHR and was seen by the WHO as an exemplar of a developed nation with a strong public health system and a willingness to contribute to a concept of global collective health security, a key element in the WHO’s approach after SARS. Integral to Canada’s contribution to global health security was the work of the GPHIN, an innovative platform for collecting and assessing global media stories on health outbreaks and sharing these with the WHO, state, and non-state entities. The GPHIN started up in the late 1990s, but was infused with new life after SARS, where it had made a notable contribution and new money as a result of the Canadian national security policy in 2004. Although it was little known to the Canadian public, the GPHIN made historic contributions to early warnings of disease outbreaks in the years following SARS, including with regard to the H1N1 (swine flu) pandemic in 2009, “the outbreak of MERS (Middle East respiratory syndrome)” in Jordan in 2012, and to situational awareness with regard to the 2014–2015 Ebola outbreak in West Africa. ²¹ The media scanning platform seemed poised to make a similar contribution in the face of a new coronavirus outbreak.

A WHO joint evaluation of Canada’s public health system was conducted in 2018, led at the Canadian end by the PHAC. The WHO report scored Canada very high overall across multiple categories of assessment. It painted a picture, reflecting of course a picture that the PHAC itself wanted to paint, of a highly proficient public health capacity linked to the global architecture of the WHO IHR. The executive summary had this to say:

Institutional strengths, resources, self-awareness and technical expertise have allowed Canada, over the past two decades, to explore innovative technological paths and modi operandi, some of which, such as the Global Public Health Information [sic Intelligence] Network (GPHIN), have not only triggered changes of paradigms at global level, but have also shaped the IHR. ²²

The section of the WHO evaluation devoted to surveillance painted a picture of a jurisdictionally complex system for domestic surveillance and noted the important role played by the GPHIN for global surveillance. The GPHIN was credited for generating “most signals related to potential or acute international public health threats.” ²³ The WHO report also suggested that the GPHIN signals “are also rapidly acted upon as they trigger a cascade of actions, across jurisdictions and government sectors, ranging from assessment and monitoring to discrete or large-scale responses.” ²⁴ This was an idealized portrait. It was to be tested and found wanting with the COVID-19 outbreak.

The idealized picture of the GPHIN’s capabilities and contribution to global surveillance remained in place right up to the onslaught of a novel coronavirus in Wuhan, the PRC. As late as November 2019, a senior Canadian PHAC epidemiologist presented a rosy account of the GPHIN’s capabilities and value to a WHO meeting in Geneva. The Canadian presentation highlighted GPHIN’s uniqueness as the only media surveillance system that is state-owned, noted that it functioned through a combination of human analysis and artificial intelligence (AI) applications, and that it was the “earliest event-based surveillance system to use Big Data.” ²⁵ According to Canadian data, the GPHIN system contributed about 20 percent of the total intake available to the WHO through its own “Epidemiological Intelligence from Open Sources” (EIOS) platform. ²⁶

The GPHIN’s media scanning intake was able to ingest some 7000 news articles per day. Its filtering processes allowed it to winnow that daily intake to some 5–10 significant news items to be included in a daily GPHIN report. ²⁷ The WHO presentation also mentioned GPHIN “alert notifications” to an international clientele, even though we now know that it had stopped generating such alerts several months earlier in 2019, and would not resume them until late into the COVID-19 pandemic. ²⁸

The Canadian public was slow to learn details, and in a different story, about the GPHIN. Two weeks after Canadian authorities had belatedly begun to respond to COVID-19 with emergency measures and border closures, the President of the PHAC prepared answers for a Parliamentary Committee appearance on 31 March 2020. Tina Namiesniowski provided some details about the GPHIN’s operations, indicated that GPHIN had captured an early Agence France Presse report about a mysterious pneumonia outbreak in the PRC on 31 December 2019, and published its first daily report mentioning this news story on 1 January 2020. ²⁹ The Health Committee of Parliament was given no indication that anything was amiss with the GPHIN and its cornerstone early warning capacity.

In late April 2020, Canadian investigative journalists from the CBC and The Globe and Mail began significant news reporting on the GPHIN, which soon uncovered concerning details about its technological capacity, budget shortfalls, restraints on internal scientific reporting, management problems, and cessation of its alert reports to a global clientele. ³⁰ Perhaps most concerning of all were indications that the GPHIN’s fundamental mission for global health surveillance was being undercut by a desire on the part of Canadian officials for it to focus on more domestic, or at least North American, disease outbreak reporting. ³¹ There was some premonition of this in the presentation to the WHO in November 2019, which pointed to intelligence gathering from media sources on disease outbreaks related to vaping in the United States as a key recent success story for the GPHIN. ³² Eventually, the tide of media investigations forced Minister for Health Patty Hajdu to commission an independent, external review of the GPHIN, the terms of reference for which remain, at the time of writing, unavailable. ³³

What can be said is that a shakeup of the GPHIN, whatever its outcome and time frame, is inevitable. A potentially similar shake-up of the GPHIN’s counterpart at the WHO, the EIOS, may result from the announced independent review of the WHO.

The EIOS program at the WHO is meant to provide a system-of-systems (or “super system”) for the use of what the IHR called “other reports.” The EIOS would not be reliant on any one stream of open-source information but was meant to be able to ingest multiple streams of information globally. It was built from earlier initiatives, including the WHO collaboration with Canada on the GPHIN, the work of the multilateral Global Health Security Initiative to develop an early warning and response platform, and efforts undertaken by the European Commission’s Joint Research Centre to develop the Hazard Detection and Risk Assessment system. All of this work came together in 2017 with the creation of the EIOS under WHO auspices. The EIOS was an ambitious undertaking meant to be rolled out over a three-year period from 2017 to 2020. ³⁴ The WHO headlined the program on its website under the banner “Saving Lives Through Early Detection.” ³⁵ But COVID-19 struck while it was still in its infancy.

Details of how well the EIOS functioned await the WHO independent review. According to the WHO, by the end of March 2020 the EIOS system was collating some 228,000 news articles per day relating to COVID-19, a magnitude far greater than what the GPHIN was engaged on. As the WHO obliquely acknowledges, information flow on this scale poses its own challenges in terms of usability. “Members of the EIOS community have been working on additions to the system to help manage the unprecedented volume of articles by looking at improved ways to filter, contextualize and visualize all of the content coming in.” ³⁶

If the global health surveillance architecture set out in the WHO IHR has proved inadequate in the face of the COVID-19 onslaught, both in terms of national inputs and WHO capabilities, and assuming that equally devasting (or worse) pandemics await the world in the future, the question becomes how can the international community forge a more effective system for early warning?

The Canadian experience here is telling. Canada pioneered an open-source global health intelligence platform in the GPHIN, which was touted as an important, even game-changing, contribution to the IHR system. Yet a yawning gap between the GPHIN’s potential and real capabilities was allowed to open up. That gap was partly a product of resource deficiencies, technological limitations, and vision constraints on the part of the PHAC. The GPHIN mission appears never to have been fully embraced by the agency that housed it. Another problem was that the GPHIN output of media scans was not properly integrated into a risk assessment and decision-making process. ³⁷ During COVID-19, GPHIN reporting ballooned into voluminous, unverified, and unanalyzed daily media summaries, whose real utility was doubtful and may have been nil. ³⁸ A similar problem of putting open-source information to good use may have beset the WHO’s EIOS, with its much greater volume of data and dearth of human analytic capacity.

Moreover, the very conceptual framework of the GPHIN was undercut by bureaucratic decisions to end an “alerts” system that allowed individual GPHIN findings to be shared by Canada with international partners. Instead, the GPHIN media scraping effort was allowed to reach only a domestic Canadian audience of public health officials. ³⁹

The GPHIN was also meant to exemplify the value of open-source intelligence collection for health surveillance and prove the validity of the IHR’s effort to free itself from sole reliance on disease outbreak reporting by host countries. There can be no doubt about the value of open-source collection and reporting on global health problems, just as there can be no doubt these days about the value of open-source collection and reporting for national intelligence systems. But in the Canadian construct, the GPHIN’s effort at open-source collection and reporting was completely siloed from the efforts of the Canadian security and intelligence system and its many agencies, some with a foreign intelligence collection and assessment mandate.

The problems that hemmed the GPHIN in, and appear to have prevented it from making any significant contribution to early warning are, in turn, a guide to what needs to change in the global pandemic alert system. Both at the Canadian and international levels, the importance of open-source intelligence for global health surveillance needs to be recognized and invested in. The PHAC will have to restore pride of place to the GPHIN, recognizing the importance of both open-source collection and related risk assessment. The WHO, in turn, will have to invest more heavily in its EIOS system and find the right combination of AI applications to filter a huge volume of media stories and human analytic capabilities. Open-source intelligence reporting, risk assessment, and decision-making will all have to be better fused together, at both the national and international levels. The high conceptual and organizational walls that have traditionally separated public health surveillance from the global collection and assessment capabilities of national intelligence systems will have to come down so that both the health surveillance and foreign intelligence missions, and the decision-making that benefits from them, can achieve closer integration and knowledge exchanges. ⁴⁰

But even with reforms to the GPHIN and the WHO’s EIOS “super-system,” the architecture of global health surveillance will remain dangerously incomplete and unstable. The United States decoupling from the WHO, if it persists, will weaken the UN organization, no matter what reforms are made. The idea that open-source intelligence, especially from media scans, could be a valuable supplement to national responsibility for reporting disease outbreaks will remain, but the broader viability of the IHR’s collective health security approach remains uncertain. Some measure of retreat from a WHO-centric collective approach, however, appears inevitable, given the lack of success in confronting COVID-19 under the IHR system.

New arrangements will need to be looked to as a public health insurance policy, alongside efforts to shore up the previous WHO IHR architecture. These new arrangements may bring a greater role for national intelligence systems in the early detection and assessment of global health outbreaks. They may also strengthen existing multilateral efforts, such as the Global Health Security Initiative, to bring a health security agenda to the the five eyes intelligence partnership, a long-standing relationship between Canada, the US, the UK, Australia and New Zealand for mutual intelligence sharing and national security consultations, and to bring greater focus to health security issues for the G7 grouping. Canada’s former public safety minister, Ralph Goodale, argued in favour of utilizing the broader and more diverse array of states involved in the G20 as a forum to discuss global health security. ⁴¹

It seems likely that new forms of tension will emerge between national, multilateral, and international efforts at global health, with an uncertain geopolitical centre of gravity. An international, collective security approach is likely to be sustained to a degree in renewed multilateral frameworks that might provide some substitute for diminished expectations regarding the WHO and its IHR system. National health systems will be reinforced, at least for a time.

One lesson embedded in the COVID-19 experience is that open-source intelligence, broadened from media scanning alone, will be a core requirement of effective early warning and must be matched to a sophisticated risk assessment capacity about public health threats, made available to decision-makers at the national, multilateral, and international levels. COVID-19 over-matched both existing open-source intelligence systems and the risk assessment processes meant to make sense of such data. But it also showed their potential. Correcting flaws in open-source intelligence and risk assessment must be done in time for the next global pandemic. Canada has its own work to improve open source intelligence and risk assessments; it can also make a contribution to improvements internationally by maintaining its commitment to the WHO, reforming rather than abandoning the IHR, and helping to bring health security issues to different multilateral fora—all in pursuit of the laudable and necessary, but now weakened, goal of collective health security for the planet. ⁴²