Review of the literature on collaboration.

Scholars of public policy, organizational partnerships, team science, and multisectoral collaboration have produced a series of theoretical frameworks to describe cross-boundary collaborations and identify which practices make them successful [13–15]. The focus on collaboration and partnership is not unique to any one discipline, yet there is very little cross-fertilization of research across disciplines. This research builds upon the existing literature on cross-boundary collaborations and applies it to One Health collaborations. The conceptual framework for this study focuses on three critical phases of a successful cross-boundary collaboration: adequate starting conditions, an effective process of collaboration, and attention to the outcomes of collaboration [16–20].

Starting conditions.

There is a general consensus in the literature on cross-boundary collaborations that starting conditions-the conditions in place before any collaborative process begins—impact the process, structures and outcomes of collaborative engagement. These include prior history (e.g. successes, failures, existing partnerships), the environment (e.g. resource imbalances, stakeholder incentives), and relational dynamics (e.g. balances of power, who convenes or facilitates the collaboration, and how and by whom problems are defined) [16,17]. The presence or absence of such conditions influences successes and challenges encountered during the collaborative process.

Process.

Beyond starting conditions, many scholars point to the process of collaboration itself and the structures in place to support effective collaboration [13,14,20,21]. Although the terms used for collaboration vary, scholars focusing on the process of collaboration point to the importance of leadership, shared goals, trust and mutual understanding, institutional structures and resources, communication, and data management.

Measuring outcomes.

A review of the literature on collaboration suggests a lack of validated metrics for measuring collaborative effectiveness and performance. Several scholars of cross-boundary collaborations, citing works published between 2005 and 2019, highlight the importance of measuring the outcomes of collaboration and lament the challenges of describing and evaluating collaborations in a uniform way [12,14,16,17,20,22–24]. This underscores the importance of understanding which factors support collaborative efforts, and how teams can evaluate their performance and outcomes in association with these factors.

The literature on cross-boundary collaborations and its attention to the starting conditions, processes, and outcomes of collaborative approaches have informed this study on the factors that enable effective One Health collaborations. The following questions guided this study: What factors (systems, structures, processes, skills, competencies, decisions, and actions) enabled two or more disciplines or sectors to collaborate effectively in a health event?

Selection of relevant articles.

The search included peer-reviewed articles available to-date in the U.S. National Library of Medicine’s PubMed database that were searched using specified MeSH (Medical Subject Headings) terms. Although the multidisciplinary research team has extensive experience in One Health, they were not trained in sensitive search strategies [26]. The research team thus elected to work with a University of Minnesota research librarian to develop MeSH terms for this study. Table 1 provides a list of the key terms used to identify articles discussing multisectoral health events and collaborations. To avoid tautology, it was a deliberate decision to not use “One Health” as a search term. Instead, drawing upon the researchers’ extensive experience in One Health, various terms were used to describe One Health and similar multidisciplinary and cross-sectoral health collaborations. The underlying assumption was that any articles explicitly addressing One Health would be captured using these key terms. This initial MeSH search identified 2,630 non-duplicated articles. This scoping review was an inductive study of the literature and was conducted in order to support more hypothesis driven research for One Health. By design, the authors elected to limit this literature review to the PubMed database at the outset of the study. PubMed is peer-reviewed and peer-led database. Articles are selected and included based on scholarly and quality criteria by literature review committees and are tagged by keyword and by article structure, contributing to more accurate retrieval than other databases (e.g. Google Scholar); accurate retrieval supports the search results are reproducible and reportable, which is critically important for a scoping review of the literature in which it is important for other researchers, no matter their location, to repeat the study. The decision to use one database reflects the exploratory nature of this study and the Author’s intent to propose further hypothesis-driven research that may include additional databases. This methodological choice is in line with Arksey and O'Malley (2005) who attest that decisions must be made at the outset of a study to clarify reasons for the coverage of the review in terms of time span and language [26] [p.23-24].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Article search terms (yielded 2,630 unduplicated results).

https://doi.org/10.1371/journal.pone.0224660.t001

Initially, citations and abstracts of these articles were screened in two phases. The articles were reviewed for inclusion based on the criteria outlined in Table 2. In the first screening, 179 abstracts met initial inclusion criteria and full articles were procured and reviewed. In the second phase of screening, two further criteria were included to better achieve scoping review objectives. The research team divided into transdisciplinary pairs which included a reviewer from the health sciences and one from the social sciences. Each of the articles that met the initial inclusion criteria were divided among the team members and then independently reviewed according to the modified screening criteria. Articles were included if both reviewers agreed that they met all initial requirements. In instances where the transdisciplinary reviewers did not agree, the articles were brought to a full research team meeting and reviewed jointly until consensus among all researchers was achieved. This same method of collaborative review was used for the second round of screening and resulted in 50 articles for the final analysis. The PRISMA diagram below (Fig 1) illustrates the article search, screening, and review process.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA diagram.

https://doi.org/10.1371/journal.pone.0224660.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Article screening criteria.

https://doi.org/10.1371/journal.pone.0224660.t002

Analysis.

The interdisciplinary team conducted an analysis of the 50 articles that explicitly addressed multisectoral collaboration in response to an actual health event. Each reviewer coded approximately 5–10 selected articles using the qualitative data analysis software, MaxQDA [28]. Descriptive codes were identified in advance to ensure that baseline data reflected the One Health aspects of the articles reviewed. All other codes emerged from the data using a grounded theory approach [29,30]. Preliminary and axial coding procedures are outlined in the following section and ensured that inductive and deductive thinking could be related.

Preliminary coding.

A set of predetermined, descriptive codes were used to denote the location and nature of the health event in the articles, including specific infectious agents, relevant disease vectors or hosts, and the various entities involved in the collaboration. Each paper was coded for the predetermined codes outlined in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Preliminary descriptive codes.

https://doi.org/10.1371/journal.pone.0224660.t003

Predetermined codes were also used to identify the entities involved in each health event response. The team used the code “roles” to identify individuals or groups who participated in the coordinated response in a formal role based on individual expertise and formal training. While many of these roles represent professions in the health sciences, this category also included representation from the social sciences, media and community relations, government, and engineering. Other articles focused on types of training, identified by the research team as “disciplines,” (e.g., clinical epidemiology [31] or food safety [32]), or specific professions (e.g., toxicologist [33] or information technology specialist [34] versus specific professions). A third type of classification in the literature was more general categorization of sectors involved, such as the traditional designations of Public, Non-profit, and Private/For-profit.

Axial coding.

Axial coding was used to construct linkages between “data sets” or, in this scoping review, articles regarding intersectoral collaboration. Axial coding is a qualitative research technique to relate data together in order to reveal codes, categories, and subcategories, as well as patterns in the data [35]. This grounded theory is an iterative process that combines inductive and deductive thinking.

Each article was first coded to identify any area of text where authors analyzed collaboration around a specified health event. In this process, it became quickly apparent that the review team would need to differentiate between actual and hypothetical forms of collaboration reported. All articles included in the analysis at this stage were retrospective analyses of actual health events, yet many were actually prospective in their analysis and discussion. As an example, several of these articles included suggestions based on what should happen in an ideal scenario, rather than what occurred in practice, thus leaving out key details of the actual event. Therefore, a first round of organizing codes differentiated between collaborations that actually happened versus ideal scenarios and hypothetical lessons, allowing the research team to focus the analysis on what actually happened (Table 4). The text was further coded to reflect whether the authors were reporting a success factor of collaboration, or a challenge of collaboration. Both the successes and challenges reported in the literature were related during the grounded theory thematic analysis and informed the final thematic results reported.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Organizing codes.

https://doi.org/10.1371/journal.pone.0224660.t004

After the first round of axial coding was conducted to organize the data, the authors employed a deductive framework developed from the review of literature on multisectoral collaboration [13]. Aligned with this framework, the research team distinguished between starting conditions for collaboration, the process of collaboration itself, and the outcomes of collaboration (Table 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Deductive codes informed by the literature.

https://doi.org/10.1371/journal.pone.0224660.t005

Finally, the review team re-examined the passages coded as “actually happened” and “successes”. These codes were then related to the deductive codes of starting conditions and process-based factors. An Excel table was used to organize axial codes into a table of final results.

Limitations.

The primary limitations of this scoping review are three-fold. First, the literature analysis relies on peer reviewed publications alone, which may have underrepresented collaborative efforts that are more commonly encountered in grey literature. Future work may be expanded to include these types of sources.

Second, there was no consistent framework or language for reporting the successes or challenges of collaboration, and thus, important content may have been missed during the search and review [36]. The scoping review team tried to overcome this with two strategies, which included building an expanded list of search terms and conducting an iterative review process using two independent transdisciplinary reviewers. Both methods offset this limitation and might have minimized the likelihood of missing specific content. Third, the researchers could not identify specific metrics for evaluating performance and collaboration in the literature. This meant that an evaluation baseline was not present. However, the research team believes that the final subset of articles represents a diverse crosssection of transdisciplinary efforts around emerging health events.

Individual factors.

Relevant success factors at the onset of a One Health event include an individual’s education and training, as well as prior experience and existing relationships. Many authors identified existing or previous education and training as enabling factors for collaborative success [37,40–42,44,47,49,51]. Formal technical education and training of individual workers prior to a health event was critical to prepare the necessary human resources for response efforts. Authors noted that foundational technical training during an event was often not possible [41,42], but that preemptive and collaborative planning did support the development of key relationships, and in some cases, the development of shared protocols used in the response. The absence of formalized training opportunities before an event, both individual technical training and collaborative, were frequently reported as a gap and a challenge to effective One Health response [40–42, 49]. Shared competencies were suggested as a strategy for standardizing protocols and performance across multiple individuals and organizations [49]. Multiple sources also reported the importance of prior experience in collaborative response efforts and how this established existing relationships to support the work, both formal (i.e. required communication through standard operating procedures) and informal (i.e. loosely structured and based on personal relationships and ongoing professional engagement) [34,45,52,53,56,57]. When instituted before a health event occurs, these starting conditions were reported to support a more effective collaboration processes.

Individual factors that supported the process of collaboration were most frequently reported as workers having access to necessary education and training that was available ad hoc or as “just in time” training to support operations during the health event. Examples reported include the use of shared training across organizations to additionally support institutional alignment and partnership with community-level organizations to provide training [39,42,49]. Many of these trainings were reported to be rigorous and responsive with continuous follow-up to support compliance [31,45,58]. Williams et al [45] discussed how ongoing multisectoral disease specific training supported workers to operate within new and existing systems while simultaneously sharpening their technical competence. These training and capacity-building opportunities were reported to provide a platform for better collaboration for outbreak response [49]. However, ad hoc trainings do not replace or diminish the need for foundational technical training, as formalized education and training were reported as a critical starting condition to facilitate quick mobilization in the case of a health event. Our literature review uncovered the role for both strong university-based education and training, and the role that ad hoc or “just-in-time” training can play to meet immediate operational needs during process-based response.

Organizational factors.

Factors reported to enable organizational-level collaboration were broadly applicable to both the starting conditions and the processes of collaboration. Organizational structures, culture and resources were cited as important elements for creating an enabling environment for effective One Health collaboration. The organizations serve to connect the individual worker with a network of One Health actors.

The organizational structures that support collaboration were often discussed as a success factor. These structures include, but are not limited to, the policies and protocols or systems established within organizations to support technical implementation and collaborative efforts. Policies and protocols reported to be supportive included technical guidelines and standard operating procedures, as well as management, response and communication strategies and protocols [34,42,50,59,60,62]. In addition, organizations reported the need for functional systems for information and resource management and sharing and reporting both surveillance and laboratory results [41,43,45,48,55,59,63,64,66,67]. These systems were reported to benefit from being adaptive, flexible/reflexive and improved through iterative feedback and monitoring and evaluation [38,46,48,60].

Organizational culture was reported in multiple key areas [31,35,38,41,48,49,60,61,68–70,81]. The role of organizational leadership was discussed at length in many of the reviewed publications. Authors recognized and identified the importance of having strong and engaged leadership [31,34,52,68] and the need for leadership to support the iterative and developmental review of collaborative processes [60]. In addition, organizations benefited from having a culture that supported accountability, ownership, cultural engagement and diversity [53,68]. Trust and credibility were consistently reported as a key element of organizational success [38,41,49,70], as was the need for both an understanding of each other's processes and systems based thinking [34,48,53,70]. Authors reflected on the importance of cultural awareness, transparency of communication processes [31,34,53,61] and the engagement of diverse stakeholders who were able to reflect community needs [53].

Human resource-related factors appeared in all three levels of analysis. Research suggests that workers need to be trained at an individual level, have defined roles and responsibilities at an organizational level, and need to be able to mobilize their efforts at a network level. At an organizational level, Human resources are made up of individual contributors and also function as collective entities that reflect employees’ prior experiences, existing relationships, and the collective institutional knowledge of its members [31,34,45,49] serve to benefit the organizations in which they work. Clear roles and responsibilities were consistently reported [34,42,45], as well as awareness of systems in place to support ongoing engagement, operations and information sharing [31,45]. Additionally, several authors highlighted the importance of a reflexive workforce, i.e. human resources that were readily available and could be mobilized quickly and efficiently to respond to health event in a rapidly changing context [31,42,45,57].

Network factors.

Starting condition factors reported to enable collaboration at the network level included network structures, existing relationships, available resources in the face of a health event, and the political environment in place to support these efforts. Pre-existing network structures were reported to provide a foundation for effective collaborative efforts to occur across participating organizations. Established Multisectoral Coordinating Mechanisms (MCMs), also referred to as One Health platforms or joint task forces, were often reported as key to assisting with collaboration across a network [34,41,42,45,48,52,60,72]. Organizational and network structures provided operational standards that crossed relational and organizational boundaries at all levels of the system—individual, organizational and network—which supported the development of formal relationships at each level. Analysis suggested that these systems and relationships need to be in place before the health event. MCMs provide a formalized operating foundation in which organizations and individuals could contribute, and formalized roles and responsibilities supported effective human resource mobilization in both organizations and networks [34,42,45,72]. These structures were often supported by formal policies or agreements such as bilateral agreements or Memoranda of Understanding (MOUs) [41,45,72]. In addition, operating procedures such as the Incident Command System (ICS) also supported effective mobilization of multiple organizations within the MCM [60]. Finally, the importance of formal structures were repeatedly emphasized as a response to “lessons learned” during challenging responses. On the contrary, lack of existing structures was reported to prevent efficient multisectoral engagement in the preparedness and response to health events [37,42,71]. Several sources indicated that reporting structures and policies at local, regional, national and international levels support continuity of response and effective implementation in response to health events [45,48,49,58,60,62,65,72,73]. These reporting structures and policies allowed for information flow between stakeholders, and the coordination of response efforts across a diversity of individuals and organizations participating in preparedness or response efforts [31,49,58,65,74].

Established structures created a foundation for network relationships that support effective outbreak response to a health event [31,40,42,45,47,49,54]. Development of formal and informal relationships prior to a health event allowed individuals, organizations and networks to more effectively respond once an emergency arose. The existence of structural agreements in any form such as MCMs, MOUs, shared Standard Operating Procedures (SOPs) or bilateral agreements were reported to support the further development of existing relationships to promote ongoing engagement prior to and throughout a health event [55,74]. Preemptive planning for potential disease threats was reported to strengthen connections and relationships and support multisectoral disease training, sometimes leading to shared protocols [45]. Additionally, the creation of common goals [34,47] and clearly defined, previously established, roles and responsibilities for individual actors and network partners were reported as necessary in network operations [34,40,42,45,49,54]. Cultural awareness and the inclusion of diverse stakeholders from government, nonprofit, and private sectors from the national to community level, was consistently reported as a success factor for collaborative efforts if included from the start [33,34,37,42,51,52,56,59,61,75].

Availability of resources, including human resources, that can be easily and efficiently mobilized in a health event was considered an important factor for response [31,34,37,41,44,45,49,52,54,74,82]. Authors also noted the importance of a supportive political environment to aid in the development and institutionalization of effective collaborative structures [41,48,65,72]. A supportive political environment was reported to enable the flow of available financial, human and material resources and empowered decision making [72]. Readily available resources supported rapid mobilization of collaborative efforts when a health emergency occurred. This is particularly impactful given that the absence of these resources and actions was noted across the literature as challenges to effective health response.

Network leadership and management processes were critical to effective multisectoral response efforts. Leadership engagement during a health event allowed for the mobilization and needed support for management processes. By utilizing existing structures and decision-making power, leaders and lead agencies can support managers and the process of management across organizations and networks. Emergency response protocols, such as the ICS, were frequently reported as mechanisms to this end, by concretely providing a leadership and management structure to support ongoing multi-organizational response. It was particularly useful for identifying a lead agency and establishing structures for regular meetings and communications. In the process of collaboration, relevant network factors included network leadership, management, and the effective and efficient mobilization of resources for response. For example, strong and engaged network leadership was noted as an important success factor for collaboration. When established prior to a health event, factors reported to support network collaborations included identifying a lead agency [41,52], promoting information sharing and joint decision-making across a network [49,60,65], and convening regular multisectoral meetings [53,58,60]. In addition, strong leadership was integral for strategic risk communication across the network [45].

Effective network management during an outbreak was reported in the areas of management practices, monitoring and evaluation (M&E), communication, awareness and ongoing stakeholder engagement. Management practices included case and task management through the MCM [41], regularly scheduled meetings [53,58,60] and development of shared response guidelines and management protocols across the network [58,81]. These management practices, when paired with existing structures, can support rapid start up response in the face of health events [51]. Monitoring and evaluation allowed for the iterative review of the collaborative processes during response efforts, as well as the outcomes of the collaborative process [31,34,38,45,48,55,60]. Monitoring and evaluation was reported as integral in being able to show the outcome of interventions as beneficial or not [31,38,45].

The importance of communications cannot be overemphasized and was repeatedly reported as an integral factor for building relationships, trust and supportive organizational culture, and for contributing to effective response processes. Both the characteristics and the methods of communications were highlighted as important. Characteristics of successful communication included the need to be frequent and honest [44,45]; timely and consistent [45]; reflexive and flexible [59] and prioritized (risk-based) [45], and streamlined [54,70]. Additionally, characteristics included the need for communications that build trust [49] and have a clear purpose [31,44,70]. These elements were widely reported to support effective communication within and across organizations [31,41,44,48,49,51,53,54,58,60,70,77,78,80].

Communication was deemed most effective when it was regular, frequent, and designed to foster awareness and support the engagement of a range of stakeholders, from local through national, regional and international levels. The MCMs, or the use of ICS, were often cited as important organizing structures for ongoing communication during a health event, supporting meetings, data collection and information sharing [43,48,58,60], underscoring the importance of starting conditions to support communications. Multiple methods of communication were reported including electronic communications, list-servs, contact lists and regular meetings; in many cases these were supported through existing MCMs [48,58] Monthly meetings [53,58,60] and establishing clear lines of communication [31,44,51,77,78] were reported as critical. These methods were supported by the use of a variety of methods and platforms such as press briefings, websites, television, newspaper, teleconferencing, listserv, available contact lists, local/ regional/ cross-border meetings and periodic reporting [44,45,49,53,58,62,77]. Additionally, leadership and management processes played a key role in supporting or challenging communication; high-level support, resource allocation, and use of good practices across an organization are foundational for good internal and external communication.

Closely linked with communication was the reported importance of building shared awareness and diverse stakeholder engagement. Awareness included information sharing, education campaigns, jointly coordinated communications and public release of reports with all members of the network and with the public [31,38,44,45,52,54,55,61,70]. Engagement of diverse stakeholders before, during, and after the response was reported as essential; these stakeholders included community and local actors, national governments, intergovernmental organizations, and operating partners [45,49,51–53,75]. To facilitate communication across stakeholder groups, Adams et al. [60] and Butler et al. [70] underscored the importance of transparent joint communications specifically between responders and community leaders for efficient and effective response. Butler et al. [70] further reported the success of joint interviews held with stakeholders to support shared understanding of response needs. Diverse partners, including foreign militaries, were reported to support foundational infrastructure that allowed other international partners to stay involved when supporting a response effort when they would not have been able to serve effectively on their own [39,71,79]. Common goals, common interests, and perspective sharing amongst stakeholders were reported to support an effective response to a health event [38,49,53].

Resource mobilization and allocation during an event, relies heavily on the starting conditions, as well as the communication, leadership and management during the process of collaboration. A number of authors pointed to the importance of being able to mobilize both the material and human resources. Once again, the involvement of diverse stakeholders, the use of MCMs and management systems such as ICS were attributed with the ability to draw upon existing resources. Processes characterized as successful included establishing a supply chain with standardized access, delivery, allocation and flow [34,41,68,80]. Human resource mobilization benefited from online recruitment [45] as well as reflexive Human Resource protocols to ensure that positions were filled and workers are incentivized and rewarded for participation [31,57].

Outcomes reported.

Although the researchers created a code to capture reported outcomes of collaborative efforts, only a small number of authors reported outcomes of their collaborative processes. Outcomes were consistently missing or under-reported in the literature reviewed, and this is likely a result of One Health outcomes being difficult to characterize. The few reported included the outcomes of cost reduction and improved safety [34], decreased mortality [41], reduction in MRSA (Methicillin-resistant Staphylococcus aureus) cases [31], increased stakeholder buy-in [45], and a report that multisectoral professional development opportunities resulted from the response [47]. However, implementation of M&E activities was one of the major gaps in the reports of One Health collaboration. The majority of articles reviewed never discussed the evaluation of either the process of collaboration or the resulting outputs or outcomes. This creates a pivotal challenge in understanding how to improve One Health operations. The authors noted that outcomes of collaborative efforts were consistently missing or underreported in the literature reviewed.

Language in collaboration.

Language used to describe One Health work continues to be a challenge when working across disciplines. Each discipline contributing, and specifically those authors reporting on these interactions, bring their own nomenclature and vernacular [36]. As also discussed in the limitations of this work, we encountered challenges in how authors reported on which entities were involved in the response to a health event. Organizations, sectors and disciplines were characterized in different ways, making is difficult to find a standard classifying system for the coding.

Considerations for the evaluation of One Health.

Despite an emphasis on the importance of iterative improvements to collaboration, the implementation of monitoring and evaluation activities was one of the major gaps in the reports of One Health collaboration. The majority of articles reviewed never discussed the evaluation of either the process of collaboration or the resulting outputs or outcomes. This creates a pivotal challenge in understanding how to improve One Health operations. It became clear in the literature review that there was no standard framework for how to evaluate One Health processes [12, 36]. Although networks and collaborators such as the Network on the Evaluation of One Health are making important advances, practical evaluation tools are still needed [83]. Some authors from public affairs, such as Emerson and Nabatchi (2015) et al.[19] have proposed a framework for evaluating outputs, outcomes, and what they refer to as “adaptations” of collaborative processes [18,19]. Their work is one of the first to propose an integrated framework that captures collaborative results at all levels of the system, from the target population to the participating organizations, and the network as a whole. The results of this scoping review are intended to support the next steps in supporting One Health evaluation.