Conceptualisation of the two methods.

The developed participatory approach methods were inspired by established participatory frameworks, namely the Delphi method, Focus group, and the Nominal Group Technique [24,25].

Both methods shared a common six-step framework (Fig 1) and were managed by four facilitators specialized in veterinary medicine and animal welfare, two of whom were trained in participatory approaches and focus groups. In Method A, four steps (training, scoring task, data analysis, and review task) were condensed into a single 3-hour in-person session. During the scoring task, experts had the opportunity to freely propose respiratory health indicators and assess them. In contrast, Method B relied on a pre-defined list of indicators and was split into a 1-hour online scoping session (training), an asynchronous individual work (scoring task and data analysis), and a final 2-hour online session (review task).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Comparative workflow of the two participatory approach methods (A and B) used for the assessment and selection of respiratory pig heath indicators by veterinarians.

https://doi.org/10.1371/journal.pone.0358467.g001

Narrative review.

A comprehensive narrative review was first conducted in march 2024 across the PubMed and CAB Abstracts databases to identify scientifically valid pig health indicators. The search equation was: (« Indicators » OR « criteria » OR « assessment ») AND « health » AND (« swine » OR « pigs » OR « pig breeding » OR « pig farms » OR « pork »). Articles were selected if they evaluated pig health using one or several validated indicators. In total, 129 articles were reviewed. In addition, references coming from other sources were consulted to complete the list, such as Zimmerman’s reference book “Diseases of swine” [23]. The extracted indicators were firstly categorized by biological system (i.e., respiratory, digestive, locomotor, urinary, nervous, reproductive or cutaneous). Within each system, indicators were further classified into six categories (clinical signs, biological analysis, slaughterhouse data, necropsy data, performance data and environmental data), encompassing both animal-based measures (ABMs; i.e., indicators assessing animals’ reactions to their environment) and resource-based measures (RBMs; i.e., indicators evaluating environmental factors and management practices). Table 1 presents the pre-defined list of pig health indicators obtained for the respiratory health system.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. List of respiratory health indicators in pigs extracted from the narrative review.

https://doi.org/10.1371/journal.pone.0358467.t001

Development of the scoring matrices.

Scoring matrices were designed using Microsoft Excel (v1808). Their objective was to collect experts’ assessments of respiratory health indicators. A total of six scoring matrices were developed, each corresponding to a specific indicator category (clinical signs, biological analysis, slaughterhouse data, necropsy data, performance data and environmental data). The overall structure of the matrices was identical for both methods, with each row representing an indicator and each column an information to be collected. In Method A, the rows were intentionally left blank, allowing experts to propose and assess the indicators they considered relevant. In Method B, the rows were pre-filled with the respiratory health indicators identified through the narrative review. For each indicator, experts followed the same steps:

• They first provided characteristics on the indicators, i.e., whether they used it to investigate respiratory diseases or not. If not, they provided a justification and proceeded to the next indicator.
• If the indicator was used, expert specified the measurement method and the physiological stages to which it applied (suckling piglets, growing pigs, finishers, or sows).
• Experts, then, assessed the indicator’s relevance by scoring four attributes (feasibility, reliability, specificity, and sensitivity). Each attribute was scored through a statement adapted from definitions reported in the literature [12] (Table 2). Experts expressed their level of agreement with each statement by using a 4-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree). To capture the multidimensional nature of respiratory health, specificity and sensitivity were scored in relation to common respiratory diseases.
• Experts were allowed to add any comments regarding their scores. In Method B, experts could also add and assess indicators that were missing from the pre-defined list.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Definitions for the four attributes defining the relevance of pig respiratory health indicators and given to the veterinarians.

https://doi.org/10.1371/journal.pone.0358467.t002

Tables 3 and 4 represent a scoring matrix provided to experts during the scoring task and adapted and formatted for publication purposes.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Scoring matrix (indicator’s characteristics section) used by the experts during the scoring task.

https://doi.org/10.1371/journal.pone.0358467.t003

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Scoring matrix (indicator scoring section) used by the experts during the scoring task.

https://doi.org/10.1371/journal.pone.0358467.t004

Group dynamic.

Method A involved four veterinarians (two men, two women), whereas Method B involved five (three men, two women). Both panels included experts with different levels of professional experience (ranging from 4 to 39 years in Method A and from 5 to 30 years in Method B).

Regarding group dynamic, several points can be highlighted. Although experts in Method A were instructed to fill the scoring matrices individually, they interacted with each other and the facilitators numerous times. Nevertheless, all participants submitted their filled matrices within the allocated one hour period. During the review task, each indicator was discussed by the panel, although discussions were brief. Experts provided rationales and clarifications for their scores during these discussions. The pre-defined list of indicators, initially planned as a proxy to identify potentially missing indicators, was not used because the review task exceeded the allocated time. By the end of the task, the group reached agreement on a final set of respiratory health indicators.

During the online review task in Method B, experts discussed the results and provided rationales for their assessments. Differences in experts’ opinions regarding the scores of some indicators were identified and discussed by the panel. Agreement on whether indicators should be selected or excluded was reached during the discussion. By the end of the task, participants reached agreement on a final set of respiratory health indicators.

Expert understanding.

Experts suggested minor revisions to the documentation. In Method A, during the meeting before the scoring task, experts asked questions about the project and the documentation and requested additional information. At the end of the meeting, all experts confirmed that they understood the instructions.

In Method B, experts gave a positive feedback on the scoping session. They had questions mostly about the project or the documentation, and asked for more details on the aim of each step. One expert initially expressed concerns about the workload, but facilitators explained that the four-week scoring period allowed participants to complete the task at their own pace. Then, during the scoring task, two experts (out of five) contacted the facilitators for clarifications on the scoring matrices. Following the review task, experts also provided positive feedbacks regarding the results matrices and the colour coding system, which they considered useful for rapidly identify indicators of interest and those requiring discussion during the review task.

Time management.

In Method A, during the scoping session, answering experts’ questions (e.g., regarding the project’s objectives and instructions) and confirming their willingness to participate required nearly 30 minutes instead of the 15 minutes initially planned. Despite this delay, the one hour scoring task was maintained as scheduled. However, two of the four experts reported that one hour was insufficient to fill and double-check the scoring matrices. They suggested allowing several days to carry out the task in subsequent sessions. One expert considered the possibility of adding comments to the scores particularly valuable, although time-consuming. Finally, as delays accumulated, less than an hour remained for the review task. Because each expert had proposed a different set of indicators during the scoring task, facilitators first had to collect feedback from the other experts before discussing each indicator. Consequently, experts could only briefly review the results, limiting opportunities for in-depth discussion and consensus building. Overall, experts from method A devoted approximately 3 hours to the participatory approach.

For Method B, the one-hour scoping session was sufficient to answer experts’ questions and demonstrate how to fill the scoring matrices. Moreover, following the scoring task, all scoring matrices were returned on time. All five experts reported that four-week completion period allowed them to organize their time and workload according to their own professional constraints. This period also allowed facilitators to carry out a detailed data analysis for the review task. Experts estimated the time to fill the matrices to be approximatly four hours. Regarding the online review task, all of the participants completed it within the two hours allocated. Overall, experts devoted approximately 7 hours to the participatory approach.

Flexibility.

Experts’ feedback during Method A showed that carrying out the entire participatory approach during a unique 3-hours meeting was not ideal. Moreover, regarding the scoring task, one expert reported difficulty identifying which indicators to include in the matrices. Ultimately, the whole panel indicated that pre-defined list of indicators should be preferred for this work, as it could save time during the scoring process and because it better reflects their routine approach to assessing pig health on farms.

In Method B, experts gave positive feedbacks regarding the division of the participatory approach into several sessions, especially the four-weeks scoring task, which was perceived as a strength. They also expressed no concerns about using a pre-defined list of indicators.

In addition, finding between experts and facilitators a common time for starting the participatory approach and perform the scoping session was easier and faster in Method B than in Method A.

Group dynamic.

Results in both methods suggest that the use of small panels composed of colleagues may have influenced group dynamics and the functioning of the participatory process. The interactions between experts facilitated the exchange of perspectives and the explicit formulation of practical knowledge and reasoning that may not have emerged from individual assessments alone. Familiarity can contribute to a supportive discussion environment, facilitating open discussion, disagreements, and the sharing of ideas [26–29]. Similarly, small group sizes can provide participants with greater opportunities to express and explain their views and allow facilitators to explore the reasoning underlying individual assessments in greater depth [30–32]. These characteristics likely facilitated consensus between both panels. Furthermore, involving veterinarians as both animal health experts and future end-users strengthened the validity and overall acceptance of the selected indicators and process [33,34].

Nevertheless, the possibility of a “group think” effect cannot be excluded. Within groups, participants may consciously or unconsciously self-censor or adapt their opinions to those expressed by others in order to preserve group cohesion [35,36]. Such dynamics may facilitate agreement without necessarily reflecting fully independent judgement. Although experts expressed different views and discussed disagreements during the review tasks, the influence of existing professional relationships and group dynamics cannot be excluded [37,38]. For this reason, the term agreement was preferred rather than consensus throughout this study, as the process did not include formal procedures to establish consensus. More structured participatory approaches, such as the Delphi technique, seek to limit some forms of social influence through individual responses, iterative rounds, and controlled or anonymised feedback [36,39]. These features can reduce the influence of dominant participants and conformity pressures but may also limit the direct interactions through which experts explain, challenge, and refine each other’s reasoning. The methods developed in this study therefore represent a trade-off between preserving independent expert judgement and allowing direct exchanges of contextual and practical knowledge.

Expert understanding.

Maintaining expert motivation and ensuring a clear understanding of the participatory process were important considerations in the implementation of both methods. Previous studies have similarly highlighted the importance of sustained participant motivation and a clear understanding of the tasks were for effective participation [32,40,41]. The dedicated training provided before the scoring task probably contributed to this objective by clarifying expectations and reducing uncertainties. However, in both methods, the numerous questions and requests for clarification raised by participants indicated that some aspects of the documentation and initial training could be improved. In particular, providing additional context on the overall project and including more examples illustrating how to score indicators could facilitate participants’ understanding of the task.

Likewise, regular interactions with the participant during the approach may also be important. In both methods, facilitators remained available to answer questions and, when applicable, followed expert’s progress during the scoring task. Such interactions may help address uncertainties and limit participant disengagement as previously recommended in the literature [24,42,43]. In future participatory approaches, these interactions will be even more frequent.

Time management.

Time management emerged as a key consideration in the implementation of the participatory methods. Veterinarians have limited availability, making the balance between the time required from experts and the quality of their contribution an important challenge when designing participatory approaches.

Results from Method B suggest that distributing the workload over a longer period may offer advantages despite requiring a greater cumulative time commitment from experts. Indeed, a longer scoping session provided more opportunities to clarify the approach and address expert’s questions. The asynchronous scoring task, although it required a greater cumulative time commitment from experts, provided them time to reflect on their assessments, revise their scores and provide detailed rationales. In parallel, the interval between the scoring and review tasks allowed facilitators to conduct descriptive analyses and identify indicators requiring further discussion. Although shorter sessions could reduce the time commitment required from experts, this may come to the detriment to the quality of the information collected [32].

In contrast, Method A illustrated some of the limitations associated with completing all stages of the participatory process within a single session. Time constraints was detrimental to the training, the scoring task or the discussions and overall, the completeness of the final set of relevant respiratory health indicator. Moreover, experts had to carry out a difficult task in a reduced amount of time, which could have limited opportunities for them to reflect on their assessments. As highlighted in the literature, excessive time pressure can encourage experts to seek rapid agreement without fully expressing and exploring contradictory views [44].

Overall, these findings suggest that providing sufficient time for training and scoring indicators may support richer assessments and contributions.

Flexibility.

The level of flexibility in both methods played an important role in shaping the outcomes. Although the Method B required a greater cumulative time commitment, experts particularly valued the four-week scoring period, which allowed them to organize the task according to their professional availability. Thus, the burden associated with participation may depend not only on the total time required but also on how flexibly this time can be allocated. This is consistent with recommendations from the literature to limit the burden placed on experts, as excessive work may reduce motivation, and the quality of their contributions [32]. Carrying out the scoring task individually may also have provided an additional advantage by limiting some of the social influences associated with group work, such as the dominance of particular participants or conformity pressures [34,45]. However, flexibility and autonomy need to be balanced with sufficient guidance throughout the process. Previous studies suggest that participant autonomy can promote ownership of the process, whereas appropriate facilitator guidance can help maintain engagement, clarify expectations, and keep participants focused on the objectives of the task [31,46,47]. Method B combined these two dimensions by allowing experts flexibility during the individual scoring task while maintaining contact with the facilitators. This balance between autonomy and guidance may therefore be an important consideration when designing participatory approaches involving professionals with limited availability.

Finally, the use of online meetings in Method B provided additional logistical flexibility. Experts could participate without travelling and could more easily integrate the sessions into their professional schedules, potentially reducing barriers to participation [48]. Although online meetings can sometimes hinder in-depth discussions, it is believed that the effect was limited in this study, as experts knew each other, thus facilitating exchanges [49]. Face-to-face meetings (Method A) remain valuable to foster richer interactions and immediate feedback, but can provoke counterproductive behaviours and be more constraining in terms of time and logistics for veterinarians [36]. The choice between online and face-to-face formats should therefore consider both the opportunities for interaction and the practical constraints associated with expert participation.