Mapping of stakeholders in avian influenza surveillance in Canada

Abstract

Introduction

Highly pathogenic avian influenza viruses are highly transmissible and lethal in wild and domestic birds and can infect other mammals. Effective avian influenza surveillance and response requires coordinated, cross-sectoral efforts involving many organizations and individuals. A detailed understanding of who is involved and their role in surveillance and response is necessary for optimizing efforts. However, a comprehensive map of stakeholders and their roles in AI surveillance and response is currently lacking in Canada.

Purpose

The aim of this study was to identify stakeholders and their roles in avian influenza surveillance to support effective surveillance and response in Canada. This map supplements existing information, including the Canadian Animal Health Surveillance System Poultry Surveillance Stakeholder Map, by comprehensively mapping specific sectors and organizations involved in avian influenza surveillance.

Findings

The final stakeholder list included 234 stakeholders involved in avian influenza surveillance (7 international, 60 national, 167 provincial/territorial). Stakeholders could have one role, multiple roles, or be involved in all steps of the surveillance cycle. The most common AI surveillance role was action and dissemination of information (n=141; 60.3%). There were 66 stakeholders (28.2%) involved in all steps of the surveillance cycle.

Significance

This process identified and characterized stakeholders involved in surveillance and response to avian influenza outbreaks in Canada, improving awareness amongst stakeholders of who is involved and what their roles are. This map is intended to facilitate proactive communication and collaboration with the long-term goal of mitigating the impact of highly pathogenic avian influenza outbreaks in Canada.

Introduction

Highly pathogenic avian influenza (HPAI) is a viral zoonosis of global animal and public health concern [1,2,3]. HPAI viruses are highly transmissible and lethal in wild and domestic birds [2, 3]. As such, HPAI outbreaks can be difficult to control, resulting in extensive losses [2, 3]. Since January 2022, over 100 million birds in the United States and 11 million domestic birds in Canada have been culled due to an outbreak of H5N1 avian influenza virus [4,5,6]. Concerningly, this virus has also infected multiple mammalian species in the United States and Canada, likely driven by close contact with infected birds [7,8,9]. Ongoing surveillance is essential for effective outbreak response and control in non-human species and preparedness for transmission to humans [10,11,12].

Public health surveillance is the “ongoing, systematic collection, analysis, and interpretation of data, integrated with the timely dissemination of data to those responsible for preventing and controlling disease and injury” [13]. Animal health surveillance is a “tool to monitor disease trends, to facilitate the control of infection or infestation, to provide data for use in risk analysis, for animal and public health purposes, to substantiate the rationale for sanitary measures and for providing assurances to trading partners” [14]. The general goals of surveillance include describing the current burden of disease, monitoring disease trends, and identifying outbreaks [15]. Surveillance methods used to achieve these goals are varied and include both passive and active approaches [13, 15].

There are two separate avian influenza (AI) surveillance systems for domestic poultry and wild birds in Canada [16]. The Canadian Notifiable Avian Influenza Surveillance System (CanNAISS) aims to prevent, detect, and eliminate H5 and H7 subtypes in domestic poultry, as these subtypes are most likely to be highly pathogenic to poultry, causing high rates of morbidity and mortality [5, 16]. Cases of notifiable AI in Canadian domestic poultry are monitored using passive surveillance in domestic poultry experiencing signs of notifiable AI, targeted surveillance upon AI detection, pre-slaughter surveillance in commercial poultry, hatchery flock surveillance, and voluntary surveillance in the poultry genetics exporter sector [16].

AI surveillance in wild birds and other wildlife is collaboratively conducted by the Canadian Wildlife Health Cooperative (CWHC), Environment and Climate Change Canada (ECCC), the Public Health Agency of Canada (PHAC), and the Canadian Food Inspection Agency (CFIA) and through testing conducted by provinces and territories [17,18,19]. Provinces and territories coordinate detection, submission and testing of dead birds, and trapping and swabbing of live birds [20]. Live bird samples and dead birds are submitted to the relevant provincial/territorial laboratory for screening [20]. Both live and dead wild birds are initially screened for the presence of AI virus. If positive, they are then tested for H5 and H7 subtypes with subsequent confirmation by the CFIA if they are positive [21].

These surveillance systems, while using coordinated efforts, are stand-alone systems that each share data to different extents, making it challenging to monitor the total burden of disease in Canada. Poultry AI data is reported publicly as the number of infected premises per province and estimated number of birds affected [5]. The same data is reported on an interactive dashboard through the Canadian Animal Health Surveillance System (CAHSS) website [22]. Wild bird and wildlife data is reported on an interactive dashboard with details concerning each case including date of collection, date of result, result status, species name, province found, strain type, and lineage [23]. Live and dead wild bird survey results are shared in summary biweekly reports, indicating the number of wild birds tested by province and test result [21]. The CFIA and provincial/territorial partners, CAHSS, Community for Emerging Zoonotic Diseases (CEZD), industry, and animal health professionals are in frequent communication to ensure stakeholders receive reliable and current information to assist with responses to AI cases and outbreaks [17].

AI is a significant animal health concern; its zoonotic and pandemic potential also makes it an important human health concern [24]. Surveillance of poultry, wild birds, and other wild and domestic animals requires extensive stakeholder involvement and resources to appropriately capture the burden of disease in Canada. Outbreak response requires stakeholders at all levels of government and across multiple sectors to work together in an efficient manner to facilitate the rapid response necessary for effective disease control. To ensure the most effective response, a clear understanding of who is involved and their role in AI surveillance is essential. The objective of this study was to map AI stakeholders in Canada by their organizational level (provincial/territorial, national and international) and role. Ultimately, this information will help to optimize the collaboration and communication that takes place between stakeholders and allow for more informed decision making to better prevent and reduce the impacts of AI outbreaks in Canada.

Methods

A snowball approach was used to generate a list of AI stakeholder organizations in Canada [25]. An initial list of organizations was developed in consultation with a subject matter expert and study co-author (SS) at the University of Guelph. Additional organizations were identified by searching the phrase “avian influenza” on the Government of Canada website [26] and reviewing all returned webpages for the names of other stakeholder organizations. Next, the websites of identified stakeholder organizations were similarly searched to identify additional organizations. This process continued until no new organizations were identified. Contact information for each organization was retrieved using publicly available online information. The organization’s general email was used unless an individual’s specific email was provided. If multiple individuals were listed for a single organization, the most appropriate individual at the author’s (EJ) discretion was contacted. Individual representatives from each identified stakeholder organization were then contacted to check for completeness of the list and provide the names of any missing organizations. After compiling the final list, organizations were assigned to a level (international, national, provincial/territorial) and visualized using concentric circles. Organizations that spanned more than one province/territory were assigned to the federal level to avoid duplications. Next, the role of each stakeholder organization in AI surveillance was determined by searching their websites and assigning listed roles to the following steps of the surveillance cycle: set objectives, data collection, data consolidation, data analysis, data integration, action and dissemination, feedback, and evaluation [27]. Stakeholder roles could be assigned to more than one step of the surveillance cycle.

Ethical approval

Research ethics approval was not required as per the University of Guelph Research Ethics Board.

Results

Stakeholder identification

Of the 156 stakeholder organizations initially identified through web searches and subject matter expertise, 55 had publicly available contact information. The definition of a stakeholder organization was not pre-defined and included both umbrella organizations (e.g. Government of Canada), subsidiary organizations (e.g. Canadian Food Inspection Agency, which is part of the Government of Canada), and independent organizations. Many organization websites had a “contact us” form that does not allow for disseminating the stakeholder list and these organizations were not contacted. Those with public contact information were emailed by EJ requesting review of the list for completeness. Over a three-week period, responses from 32 organizations were received (58.2%) and 71 additional stakeholders were added to the list. Individuals who responded would include others from the organization who could provide input, thus resulting in more individuals than organizations who responded. Seven stakeholders were identified through further searches of newly identified organization websites, yielding a total of 234 stakeholders (Fig. 1).

Fig. 1

Flow chart indicating the number of stakeholders identified at each stage

Level of organization

Stakeholders exist at all levels of organization, including international (n = 7), national (n = 60), and provincial/territorial: Ontario (n = 26), British Columbia (n = 19), Saskatchewan (n = 19), Alberta (n = 18), Québec (n = 16), New Brunswick (n = 13), Manitoba (n = 11), Nova Scotia (n = 11), Newfoundland and Labrador (n = 9), Prince Edward Island (n = 9), Yukon (n = 7), Northwest Territories (n = 5), and Nunavut (n = 3) (Fig. 2).

Fig. 2

Concentric circles indicating the level of organization for each stakeholder and their province/territory for provincial/territorial stakeholders. The inner orange circle includes international stakeholders, the middle blue circle includes national stakeholders, and the outer green areas include provincial/territorial stakeholders

Role in AI surveillance

The most common AI surveillance role was action and dissemination of information (n = 141; 60.3%). Other roles in the surveillance cycle were evenly distributed across stakeholder organizations with 58–71 organizations assigned to each role (Fig. 3, Table 1). There were 66 stakeholders (28.4%) involved in all steps of the surveillance cycle, including the Government of Canada, CFIA, CanNAISS, National Centre for Foreign Animal Disease, ECCC, Canadian Wildlife Service, and CEZD.

Fig. 3

Bar chart indicating the number of stakeholders who perform each step of the surveillance cycle or the number who perform all steps of the surveillance cycle

Table 1 List of stakeholders, their organizational mandate, and role in the surveillance cycle

Discussion

This study sought to identify and map AI stakeholders in Canada by their level of organization and role in AI surveillance using an established framework of the surveillance cycle [27]. Using a stakeholder-engaged approach, we identified 234 stakeholders across all levels of organization, located in all provinces and territories, and functioning in all roles of the surveillance cycle. Given the importance of AI to animal and human health, it is essential that all stakeholders and their roles are specified to allow for rapid collaboration and response to AI cases in Canada to limit its spread and detrimental cross-sectoral impacts. With over 200 stakeholders identified, the AI surveillance network is large and complex with many participating organizations that need to act cohesively for effective outbreak management. Additional stakeholders not specified here include individual farmers, veterinarians, and other poultry industry workers, who are in regular, direct contact with poultry and represent the front line of surveillance, response, and management.

Given two distinct AI surveillance systems in Canada, each system will have its own unique challenges, in addition to general challenges of disease surveillance in Canada [28]. Barriers to effective public health surveillance in Canada include fragmented data systems, recruiting and retaining skilled employees, and lack of standardized approaches for data collection and sharing [28]. These challenges are relevant to AI surveillance and persist due to chronic underfunding and under resourcing of surveillance efforts [20]. For example, following the 2004 AI outbreak in British Columbia, approximately $3.5 M of financial support was provided annually for wild bird surveillance, dwindling to ~ $600 K by the early 2010s [20]. Ongoing lack of resources has limited the scope of wild bird surveillance conducted in Canada and subsequent ability to detect early warning signals [20]. The effectiveness of the wild bird surveys as an early warning system has not been evaluated; however, given the low sample sizes and opportunistic sampling, there is likely suboptimal detection of AI in wild birds [20]. In addition to lack of resources, there have been limited efforts to include upstream risk factors in surveillance systems, which may help improve early warning capabilities [20].

Wild bird and poultry surveillance data are not shared in the same format, at the same frequency, or through the same channels, presenting a challenge in estimating or monitoring the total burden of disease in all birds in Canada at a specific point in time [5, 23]. Other unique challenges for effective wild bird surveillance include migratory patterns and the remoteness of species, which can inhibit representative data collection [29, 30]. To optimize surveillance and response efforts in Canada, it is essential to minimize barriers and maximize facilitators of surveillance. A near real-time integrated surveillance system providing both wild bird and poultry data presented using harmonized data formats would promote better understanding of the total burden of AI in birds in Canada. Timely, accurate, comprehensive information is essential to rapid outbreak response and management.

Surveillance of zoonotic diseases of animal and human health importance, like HPAI, requires strong collaboration. Most AI stakeholders in Canada do not contribute to every role in the surveillance cycle. Thus, close collaboration is necessary to ensure that the full surveillance cycle is effectively and efficiently completed [31]. Collaboration in large surveillance networks can be facilitated by international coordinating bodies and targeted country participation, allowing for resources to be appropriately allocated across time and space to maximize the usefulness of the information collected [32]. Coordinating bodies can create and implement standards for surveillance programs, allowing for easier data sharing and greater data comparability, supporting both surveillance and research efforts [32]. In addition, collaboration can maximize laboratory and personnel capacity during outbreaks through sharing of resources and avoiding unnecessary duplication of efforts [32, 33].

Rapid dissemination of the best available information from trustworthy sources is another essential component of AI surveillance in Canada. Of all surveillance roles, dissemination and action were under the purview of the largest number of stakeholders. It is imperative that action can effectively follow dissemination of surveillance information, and that dissemination of information is not viewed as the sole and final action but rather part of an iterative cycle. Surveillance data is often disseminated in reports or publications containing technical jargon that makes it less interpretable by diverse audiences to inform action [34]. For AI outbreaks, there are two primary groups who must receive information: decision makers and those who are at risk of exposure or disease [34]. Information must be disseminated differently to these groups; technical reports and publications are not sufficient [34]. To overcome this challenge, evidence-based tools using different user-friendly formats (email, dashboards, and/or reports) have been and should continue to be developed and implemented to aid in rapid dissemination of data [34, 35]. The language and format of information dissemination must be adjusted to the appropriate target audience to ensure rapid and effective outbreak response [34]. In addition to these adjustments, it is important to be aware that different stakeholders have different abilities or jurisdictions in which they can respond or have authority to respond. The information that they receive must then be applicable to their jurisdiction in which they can respond and to the level at which they can act. Allowing for stakeholders to tailor the information they receive, particularly in dashboards and reports, can support a timely response from that stakeholder organization.

One tool that can be used to improve AI surveillance and response is a decision support system (DSS) [36]. A DSS is a computerized information system that supports the human decision-making process to inform action [37]. These systems integrate diverse data sources to better inform risk prediction models that can support a more rapid response [38]. These systems have been used in numerous sectors including, environmental management, healthcare, finance, and public health [39, 40]. A DSS has been developed for AI using Indonesian data and several others are under development for use in national and provincial contexts [38, 41]. While these systems can support decision-making processes, there are other activities that must be completed to enable implementation. Knowledge of who is involved and their role(s) in the surveillance cycle would allow researchers to collaborate with specific stakeholders and tailor DSS outputs to their needs. The development of this stakeholder map represents an initial step to support implementation of an AI DSS in Canada.

This study has several strengths, including a specific focus on AI surveillance that adds to existing efforts to map related areas, such as the CAHSS Poultry Surveillance Stakeholder Map, which is focused on poultry surveillance [42]. Given that AI can affect other mammals and humans, our map adds to and includes other organizations outside of the poultry sector. An important limitation of this study is that only English-speaking stakeholder organizations with publicly available contact information were contacted, and thus the addition of some organizations may have been missed. This map is expected to change over time as organization mandates change or new organizations are created. As such, this map represents a point-in-time view of stakeholder involvement that will importantly need to be updated and revised over time. However, many of the listed organizations are long-standing and thus changes are expected to be incremental. Another limitation of this study is that the surveillance roles were determined using publicly available information, which may not fully represent organization roles in AI surveillance. Related to the use of publicly available information, the relationships between stakeholders could not be reliably determined and were not elucidated. Future work could directly collect this information from organizations to determine their roles according to the organizations and the relationships between them in AI surveillance. This kind of detailed stakeholder mapping can further support system improvement by identifying areas of inefficiency, such as duplicative efforts, that may be resolved through improved stakeholder awareness and connection.

Conclusion

In this study, we sought to identify stakeholder organizations involved in AI surveillance in Canada. Using a web-based search and snowball sampling method, 234 stakeholders were identified across all levels of organization and with different surveillance roles and responsibilities. This study will serve as a reference for stakeholder identification and engagement to optimize AI surveillance and related tools, including decision support systems, to ultimately improve detection and control of AI and reduce its negative health and economic impacts.