Highlights
-
• No standardized cognitive aid exists for first medical responders at MCIs.
-
• A 2-round Delphi process with 20 experts yielded consensus on 31 items.
-
• Ergonomic optimization used the Cognitive Aids in Medicine Assessment Tool.
-
• The tool targets early scene organization and incident command coordination.
-
• Modular design supports adaptation to diverse international EMS systems.
Introduction
Mass casualty incidents (MCIs) represent a major challenge for prehospital emergency medical systems. They are defined by a mismatch between available resources and the number and/or severity of casualties, requiring an approach in which organization takes precedence over individual care.Reference Shackelford, Remley and Keenan 1 These situations are exceptional and unpredictable, and few health care professionals are regularly exposed to them or receive specific training for their management.Reference Rush, Lauria and DeSoucy 2
During the early phase, the first medical responder plays a primarily organizational role. They contribute to escalation of the response system, establish contact with the Incident Commander (IC), and help structure the initial medical response.Reference Lesaffre, Tourtier and Violin 3 This phase is characterized by uncertainty, time pressure, and information overload, which may lead to cognitive overload or “freezing,” impairing decision-making.Reference Brook 4 , Reference Baker 5
Cognitive aids (CAs) are tools designed to reduce the risk of error while improving the speed and reliability of task execution in crisis situations.Reference Blanié, De Saint Maurice and Kurrek 6 Their use has been associated with reduced mortality in critical care settingsReference Lipitz-Snyderman, Steinwachs and Needham 7 and with improved performance in simulation.Reference Truchot, Balança and Wey 8 However, their implementation in emergency medicine remains inconsistent.Reference Bijok, Jaulin and Picard 9 –Reference Blanié, Roulleau and Benhamou 11
Existing tools are mostly procedure-oriented or hospital-based and do not address the early organizational role of the first medical responder during MCIs.Reference Blanié, Kurrek and Gorse 12
MCIs involve high stress and time pressure, making CAs particularly relevant. Nevertheless, the reflex actions to be incorporated into such a tool lack consensus, and no standardized recommendations currently guide their development. These constraints limit the availability of reliable, reproducible data derived from systematic studies,Reference Culley and Effken 13 necessitating the use of an alternative method for selecting relevant actions. The Delphi method enables structured expert consensus for complex and poorly standardized situations.
To our knowledge, no published CA specifically addresses the organizational role of the first medical responder during the initial phase of an MCI.
The primary objective was to develop a CA for first medical responders in MCIs based on expert consensus achieved through a Delphi process. The secondary objective was to optimize the ergonomic design of this CA using the Cognitive Aids in Medicine Assessment Tool (CMAT).
Materials and Methods
Study Design
This methodological study aimed to develop a CA specifically dedicated to the first medical responder involved in MCIs. The study was conducted in three sequential phases:
-
1. Identification of relevant items through a comprehensive review of the scientific literature
-
2. Content agreement through expert consensus using a Delphi methodology
-
3. Ergonomic optimization of the CA using CMAT
The study received an ethical opinion from the ACASAN Ethics Evaluation Committee (IRB00014579, French Armed Forces Health Academy—École du Val-de-Grâce, Paris, France). Data collection and analysis were conducted between June and August 2025.
Selection of Experts and Delphi ProcessReference Jones and Hunter 14
The objective of the Delphi method was to achieve expert consensus on the selection of items to be included in the CA through multiple rounds of individual questionnaires. The Delphi process followed established methodological principlesReference Bourrée, Michel and Salmi 15:
Anonymity: Participants were unaware of the identity of other panel members, preventing dominance, authority, or affiliation bias.
Iterative rounds with controlled feedback: Experts participated in successive rounds and received aggregated results from the previous round prior to reassessment.
Quantitative analysis of group responses: Experts were invited to justify their ratings when necessary.
Experts were selected according to the following criteria:
Board-certified emergency physicians or anesthesiologist–intensivists with recognized prehospital practice, representing professional societies (French Society of Disaster Medicine or the French Society of Emergency Medicine, through its MCI committee), or affiliated with specialized operational units (Paris Fire Brigade, Marseille Naval Fire Battalion, or the National Fire Officers Training School).
Willingness and availability to participate in the study
In addition to their clinical expertise, panel participants had direct operational experience in MCIs in both civilian and military settings. Several experts had previously participated in MCI command, coordination, triage, or field management roles, and many were involved in national-level disaster medicine teaching and simulation-based training programs.
Eligible experts were contacted by email. The recruitment message outlined the rationale and objectives of the study, as well as the preliminary list of CA items. In accordance with methodological recommendations, a panel of 10 to 15 experts was considered sufficient to validate a Delphi processReference Bourrée, Michel and Salmi 15; therefore, 20 experts were invited to account for potential attrition.
Item Selection
An exhaustive literature review was conducted to identify relevant items for inclusion in the CA. The search combined the following Medical Subject Headings (MeSH) and keywords using Boolean operators (AND/OR): “triage,” “prehospital,” “mass casualty,” “major incident,” “exceptional health situations,” and “cognitive aids.”
Articles published in English or French were eligible without time restriction. Studies addressing early prehospital organization, triage coordination, disaster medical response, or CAs in emergency settings were included. Purely hospital-based protocols unrelated to prehospital organization and duplicate publications were excluded.
PubMed® and Google Scholar® were searched, and reference lists of selected articles were screened manually. The second edition of the Vademecum of Disaster Medicine and MCIs was also consulted.Reference Julien 16 Titles and abstracts were screened for relevance, and key operational actions were extracted to generate the preliminary item list.
The preliminary list of items was organized into 6 chronological sections reflecting the operational timeline of the first medical responder:
-
1. Before arrival on scene
-
2. Initial scene survey upon arrival
-
3. Environmental and situational assessment (radio report)
-
4. Establishment of casualty collection points and a forward medical post
-
5. Management of casualties at collection points
-
6. Evacuation
In parallel with the Delphi process, thirteen additional questions were submitted to the experts (Supplementary material Appendix 1). These questions aimed to collect expert opinions on operational context, triage principles, temporal organization, and overall relevance of the proposed actions. Nine questions were rated using a Likert scale, and 4 were open-ended. These questions were not included in the Delphi consensus process and were used solely to guide item refinement between rounds.
Delphi Phase
An online questionnaire was distributed via the DRAG’N SURVEY platform (Créteil, France). Experts were asked to rate the relevance of each proposed item using a 9-point Likert scale, ranging from 1 (totally inappropriate) to 9 (totally appropriate). A free-text field allowed participants to provide comments or suggest additional items.
Prior to dissemination, the questionnaire was peer-reviewed by physicians from the Marseille Naval Fire Battalion to ensure clarity and comprehensibility.
The Delphi questionnaires were administered in French, which is the operational language of the participating experts. The final CA was subsequently translated into English for publication purposes.
Item Selection Process
Consensus criteria were defined a priori in accordance with HAS methodological guidelines. 17
First Delphi Round
Item evaluation was based on both the median Likert score and the distribution of ratings.
-
• Strong agreement was defined as a median score ≥7 with all ratings within the 7-9 range.
-
• Strong disagreement was defined as a median score ≤3 with all ratings within the 1-3 range.
-
• Items with a median ≥7 but with at least one rating in the 1-3 range were classified as showing relative agreement. These items were not accepted in the first round and were reassessed in the second round after revision.
-
• Items with a median between 4 and 6.5 were considered indeterminate.
Items classified as indeterminate were excluded after the first round. Items not reaching strong agreement and not classified as indeterminate were revised based on qualitative expert feedback and considered for reassessment in the second round.
-
• All other rating distributions were considered as showing no consensus.
In accordance with HAS Formal Consensus Method guidelines, item revision between rounds is typically guided by an inter-round meeting of the expert panel. In the present study, this deliberative function was fulfilled by qualitative comments provided by experts in the free-text fields of the first-round questionnaire, which served as the primary basis for item modification, merging, rewording, and creation prior to the second round. Items associated with consistently unfavorable qualitative feedback were excluded after the first round due to limited perceived operational relevance.
Second Delphi Round
In the second round, revised items were reassessed using the same predefined criteria. Items achieving strong agreement or relative agreement were retained in the final CA, while items remaining indeterminate were excluded.
In summary, after the first round, items with strong agreement were accepted, items classified as indeterminate were excluded, and all remaining items were revised based on qualitative feedback and considered for reassessment in the second round, unless consistently negative comments justified their exclusion. After the second round, items achieving strong agreement or relative agreement were retained, while items remaining indeterminate were excluded.
Development of the Cognitive Aid
The selected items were organized into a format suitable for use in emergency situations, following minimalist design principles that emphasize essential information only.Reference Borshoff 18
The CMAT methodologyReference Evans, McCahon and Barley 19 was used to optimize the ergonomic design of the CA. CMAT was used as a structured framework to guide the ergonomic optimization of the CA rather than as a formal validation tool. Originally inspired by aviation emergency checklists, 20 CMAT evaluates CAs across 3 domains: physical characteristics, content, and layout/format. The CMAT includes 43 items, originally selected using a Delphi process, and enables systematic assessment of clarity, relevance, usability, and accessibility. The CA was tested in simulation scenarios; however, its impact on clinical performance and non-technical skills has not yet been evaluated.
Currently, no standardized protocol exists for assessing CAs in simulation settings.
French Organization of MCIs and Definition of the First Medical Responder
The French prehospital emergency medical system is physician-led. At an MCI, overall scene command is held by the IC, who coordinates both the firefighting/rescue and medical sectors. The medical sector is led by the Medical Incident Director, a physician operating under the IC’s authority.
The “first medical responder” refers to the physician arriving first on scene. During the critical early phase, no formal medical command structure has yet been established: this physician acts alone, assumes de facto the future functions of the Medical Incident Director, and must simultaneously conduct the initial scene size-up, establish contact with the IC, and begin structuring the medical response under conditions of high uncertainty, time pressure, and information overload. Victim identification and tracking are ensured through the SINUS system (Standardized Digital Information System for Victims), which assigns each casualty a unique prehospital identification number maintained until hospital admission.
This CA was developed within this framework primarily because no equivalent tool previously existed in the French context. Nevertheless, the functional responsibilities it targets—verification of life-saving interventions, triage coordination, casualty collection organization, structured communication with the IC, and prioritization of unstable casualty evacuation—are not exclusively physician-dependent. Within the Incident Command System (ICS) framework widely used internationally, these responsibilities are functionally comparable to those of a senior paramedic acting as initial medical coordinator. These functional similarities suggest that the CA may be adaptable to paramedic-led EMS systems, although this requires local validation and contextual adaptation.
Results
Participants
The Delphi process consisted of two successive rounds. 17 All 20 invited experts participated in the first round (100%), and 18 completed the second round (90%). The 2 non-responders in round 2 were excluded from that analysis. The characteristics of the expert panel are summarized in Table 1. Most participants were emergency physicians (95%), with more than 10 years of clinical experience in 75% of cases. The majority had training as Medical Incident Commander (85%) and experience in disaster medicine teaching (95%), and 40% were involved in research activities.
Characteristics of the expert panel. Demographic and professional characteristics of the experts participating in the Delphi process, including medical specialty, institutional affiliation, clinical experience, qualification as Medical Incident Commander, teaching experience in disaster medicine, and research involvement related to mass casualty incidents
Table 1. Long description
The table is organized into two columns: Characteristics and Number of participants (percent).
Profession:
* Emergency physicians: 19 (95 percent)
* Anesthesiologist–intensivists: 1 (5 percent)
Affiliation/status:
* Paris Fire Brigade: 6 (30 percent)
* Marseille Naval Fire Battalion: 8 (40 percent)
* French Society of Emergency Medicine: 3 (15 percent)
* National Fire Officers Training School: 1 (5 percent)
* French Society of Disaster Medicine: 2 (10 percent)
Clinical experience:
* Greater than 10 years: 15 (75 percent)
* 5 to 10 years: 5 (25 percent)
Qualification as Medical Incident Commander:
* Yes: 17 (85 percent)
* No: 3 (15 percent)
Experience in teaching disaster medicine:
* Yes: 19 (95 percent)
* No: 1 (5 percent)
Research activity related to disaster medicine or M C Is:
* Yes: 8 (40 percent)
* No: 12 (60 percent)
Abbreviations: MCI, mass casualty incident.
Affiliation reflects either current or previous operational assignment within the French prehospital emergency medical system, or recognized involvement in relevant national scientific societies, and does not necessarily indicate the primary employer.
Delphi Process Outcomes
Differences in item counts between rounds reflect removal, merging, and creation processes (Figure 1). Detailed item-level data are provided in Supplementary Appendix 2.
Flow diagram of the Delphi process.
Flow diagram illustrating the selection, modification, and validation of items across the two-round Delphi process. The figure details expert participation, item evaluation, merging, rewording, removal, and the final retention of items used for the development of the cognitive aid.
Figure 1. Long description
The flowchart follows a vertical progression with lateral inputs.
* At the top, a wide box labeled Literature Review / Questionnaire Development points down to the first stage.
* To the left of the main flow is a box for the Expert Panel, noting that in the first round, 20 out of 20 experts participated, and in the second round, 18 out of 20 experts participated.
* The central flow begins with the First Delphi Round box, which includes 7 demographic questions, 46 items, and 13 additional questions for evaluators. Results show 20 completed questionnaires, 43 items without consensus, 2 indeterminate items, 1 item with relative agreement, and 0 items accepted.
* An arrow leads from the first round to a lateral box on the right detailing item modifications: 19 items merged into 7 items, 5 items removed, 5 new items created, and 17 items reworded.
* The main flow continues down to the Second Delphi Round box, containing 1 demographic question and 34 items. Results show 20 items with strong agreement, 12 items with relative agreement, and 2 indeterminate items.
* An arrow leads from the second round to another lateral box on the right: 3 items reworded and 3 items removed (1 indeterminate item, 1 non-consensus, 1 redundant item).
* The final box at the bottom, titled Final Outcomes, states that 31 items were accepted.
First Delphi Round
A total of 46 items were evaluated in the first Delphi round. No item reached strong agreement. One item achieved relative agreement, 2 were classified as indeterminate, and 43 were classified as showing no consensus.
The absence of strong agreement reflected both the deliberately stringent consensus criteria and initial limitations in item clarity, as highlighted by qualitative feedback. For example, item 8 (concerning contact with the IC) obtained a median score of 9 but did not reach strong agreement due to missing values; item 14 (verification of life-saving interventions) obtained a median of 9 but included one extreme rating; and item 9 (assessment of circumstances and scale) obtained a median of 9 with missing values.
A detailed item-by-item breakdown of round 1 ratings, comments, and outcomes is provided in Supplementary Appendix 2.
Modifications Following the First Round
Five new items were created based on expert comments and open-ended responses: a question to the IC regarding possible zoning; verification of tactical and medical communication channels; a definition of a common operational maneuver; an estimation of initial casualty assessment (Unstable/Immediate/Delayed/Deceased/Minor); whether initial triage had been performed.
Five items were removed after the first round: 2 due to indeterminate ratings (item 1: PPE/turnout gear check; item 28: request for specialist facilities) and 3 due to lack of consensus combined with unfavorable qualitative comments (item 5: consider the 3Rs; item 15: display victim categorization; item 24: I do: list of actions performed).
Nineteen items were merged into 7 to reduce redundancy and cognitive load, including consolidation of scene assessment (items 9 and 10), access and resource staging (items 18 and 19), and casualty collection and forward medical post organization (items 29-36).
Seventeen items were reworded to improve clarity and comprehension, notably through the addition of the prefix “Question to the Incident Commander” to structure information gathering.
The exclusion of a triage algorithm reminder was a deliberate decision based on expert consensus to preserve universality, as multiple algorithms exist without demonstrated superiority; a standardized categorization framework was retained instead.
A detailed item-by-item description is provided in Supplementary Appendix 2.
Second Delphi Round
Thirty-four items were evaluated in the second round. Of these, 20 achieved strong agreement, 12 achieved relative agreement, and 2 remained indeterminate.
Two items were excluded: item 7 (designate a timekeeper), classified as indeterminate (median 5.5), and item 39 (have the 3Rs been performed?), classified as absence of consensus despite a high median score, reflecting unfavorable expert opinions regarding its limited operational added value.
Items achieving strong agreement covered key domains including contact with the IC, risk assessment, casualty estimation and categorization, structured scene size-up, organization of casualty collection and forward medical posts, and prioritization of unstable casualty evacuation. Items achieving relative agreement included communication resources, event timing, site access and resource staging, care traceability, casualty identification, and evacuation of immediate casualties.
A detailed item-by-item presentation is provided in Supplementary Appendix 2.
Modifications Following the Second Round
Three items were removed after the second round: item 7 (indeterminate), item 39 (absence of consensus), and newly created item 4, which, although achieving relative agreement, was removed during the post-Delphi optimization phase owing to redundancy with the item addressing estimation of the number of casualties.
Final Results and Development of the Cognitive Aid
A total of 31 items were retained in the final CA, including 20 with strong agreement and 11 with relative agreement.
The ergonomic design of the CA composed of these items was optimized using the CMAT methodology.Reference Evans, McCahon and Barley 19 The final version of the cognitive aid is presented in Figure 2.
Final cognitive aid for first medical responders.
Final version of the cognitive aid developed for first medical responders during mass casualty incidents. The tool is organized according to the operational timeline of the initial response and was ergonomically optimized using the Cognitive Aids in Medicine Assessment Tool (CMAT).
Figure 2. Long description
The cognitive aid is divided into two vertical pages.
Page 1/2 titled First medical responder on scene includes:
- Header: Time reference and Verification of Communication Resources.
- Section 1: Command Contact / Reconnaissance. Lists tactical channels and a checklist for evolving risk, estimated victims, C B R N hazards, predominant injuries, life-saving interventions, and operational maneuvers.
- Section 2: Predominant injuries / Activation of Care Pathway. Features a triage color-coded bar: Unstable (red), Immediate (red), Delayed (yellow), Deceased (black), and Minor (green). Includes a field for counting pediatric patients, law enforcement, and firefighters.
- Section 3: Initial scene size-up (target less than 5 min maximum 10 min). Contains prompts for I am, I see, and I anticipate. Below are request checkboxes for human reinforcements (paediatric M I C U, adults M I C U) and material reinforcements (C B R N, Field Medical post, O sub 2, Blood products).
Page 2/2 includes:
- Section 4: Casualty Collection Point C C P / Forward Medical Post. Checkboxes for location safety, organization of entry/exit routes, single exit point with Digital casualty identification system, and verification of medical triage.
- Section 5: Evacuation Phase. Checkboxes for distribution of identification bracelets, centralized casualty list, and evacuation of unstable or immediate casualties.
- Section 6: A large empty box at the bottom labeled To be locally adapted.
The CMAT framework was used to guide ergonomic optimization of the CA across 3 domains: physical characteristics, content, and layout. This iterative process resulted in multiple refinements intended to improve readability, usability, navigation, cognitive load management, and operational applicability in high-stress MCI environments.
Several modifications of the CA were directly informed by CMAT principles. A sans-serif font was selected to optimize legibility under stress conditions (CMAT attribute 1.3). The tool was organized into a 2-page format with a logical page break to maintain field usability and continuity (CMAT attribute 3.1), with explicit page numbering to facilitate navigation (CMAT attributes 1.10 and 3.5). Items were grouped into chronologically coherent operational sections to reduce cognitive load (CMAT attribute 2.1). This principle was also reflected during the Delphi process itself, with removal of items considered excessively burdensome and merging of overlapping casualty-management items. A color-coded triage categorization table was integrated to provide an immediately actionable visual reference (CMAT attributes 2.13 and 3.6). Checkboxes were added to facilitate sequential task completion and place-keeping during high-stress situations (CMAT attribute 1.11). Structured communication sections (“I am/I see/I anticipate/I request”) were formatted to support standardized reporting and coordination (CMAT attribute 3.6). A customizable free-text section (“To be locally adapted”) was included to allow site-specific adaptation and integration with local organizational protocols and operational resources (CMAT attribute 3.13). CMAT attribute numbering refers to the original published CMAT framework described by Evans et al.Reference Evans, McCahon and Barley 19
Discussion
This study describes the development of a CA dedicated to the first medical responder during MCIs, based on expert consensus using a Delphi methodology. To date, no standardized recommendations define priority medical actions during this early organizational phase, highlighting the originality of this work.
Several structured communication tools support information management during MCIs. METHANE, widely used in ICS-based EMS systems,Reference Timbie, Ringel and Fox 21 is a mnemonic structuring the transmission of incident information (Major incident, Exact location, Type, Hazards, Access, Number of casualties, Emergency services). It is fundamentally a reporting tool, organizing what the responder communicates rather than what they do on scene.
In contrast, the CA developed in this study is a sequential action guide structuring the operational tasks of the first medical responder during the initial phase of an MCI. It is therefore complementary to, rather than a replacement for, structured reporting tools. The free-text fields allow real-time collection of essential operational data, directly supporting structured scene size-up reports (e.g., “I am/I see/I anticipate/I request”). While METHANE structures communication content, the present tool structures the actions required to generate that information.
The Delphi process included 20 experts in the first round and 18 in the second, with 31 of 46 items retained (67%). Missing responses may reflect several factors, including ambiguity or insufficient clarity of certain early-round items, variability in interpretation among participants, or respondent burden associated with iterative Delphi surveys. The precise reasons for missing answers could not be formally determined. However, feedback obtained during the revision process suggested that some items initially lacked sufficient operational clarity, which contributed to subsequent rewording and refinement.
The absence of strong agreement in the first round reflects both stringent consensus criteria and initial limitations in item clarity, which were addressed between rounds.
The final list was subsequently optimized to produce an operational CA suitable for real-world use. The final CA includes 31 items organized into 6 chronological sections, covering pre-arrival preparation, initial scene assessment, communication with the IC, organization of casualty management areas, casualty management, and evacuation.
Several mechanisms may explain the potential impact of such CAs during MCIs. By reducing cognitive overload and supporting organization and structured communication, CAs may facilitate early incident management. In simulation settings, CAs have demonstrated improvements in communication and problem-solving capacities.Reference Truchot, Balança and Wey 8 In prehospital trauma care, the use of a checklist has also been associated with reduced omission errors and improved adherence to critical management steps.Reference Lefèbvre, Balasoupramanien and Galant 22 More efficient resource allocation during the early phases of incident management could also result from these mechanisms. These potential effects should now be further evaluated in simulation-based environments and real-world MCI settings.
Methodological Choices and Contribution of the Delphi Method
In the absence of standardized recommendations for CAs in MCI settings, a two-round Delphi methodology was selected in accordance with national methodological guidance. Alternative consensus methods, which are less reproducible or rely on in-person interactions, appear less suited to MCI.Reference McMillan, King and Tully 23 –Reference Connis, Nickinovich and Caplan 25 A predefined consensus threshold (median Likert score ≥7) ensured methodological rigor. The absence of strong agreement in the first round should not be interpreted solely as a methodological artifact. It also reflects the inherent uncertainty and variability of early MCI response, characterized by time pressure, limited information, and rapidly evolving conditions. This variability highlights the complexity of standardizing actions in this phase and supports the need for structured cognitive support tools. Expert feedback also highlighted issues of item clarity, cognitive load, and limited relevance of certain items, guiding iterative refinement of the CA.
The exclusion of a triage algorithm reminder reflected a deliberate expert consensus decision: multiple triage algorithms are currently in use (START ABC, START X ABC, Ten Second Triage, SALT), none of which has demonstrated clear superiority over the others in the literature, and institutional preferences vary considerably across EMS systems. Including a specific algorithm would have reduced the universality of the tool and introduced a potential source of confusion for users trained in different systems. Casualty categorization (Unstable/Immediate/Delayed/Deceased/Minor) was, however, retained as a consensus-validated item, as experts considered this framework universal and essential regardless of the specific algorithm used to assign categories. The choice of triage algorithm is therefore explicitly left to local adaptation.
Item Adaptation and Ergonomic Optimization
Item Merging
Nineteen items were merged into 7 final items to reduce cognitive load and eliminate elements deemed to provide limited added value at the stage of the first medical responder. In particular, the initial distinction between actions related to casualty collection points and those related to the forward medical post was considered inappropriate for the early phase of intervention and was therefore removed.
Item Rewording
Twenty items were reworded overall, including 17 after the first round and 3 after the second round to improve clarity and comprehension. The addition of the phrase “Question to be addressed to the Incident Commander” to selected items helped structure information gathering and reinforced the operational logic of the CA. The literature provides limited data regarding the proportion of reworded items in Delphi studies, as this parameter is rarely reported.Reference Schifano and Niederberger 26
Item Removal
Eight items were removed over the course of the 2 Delphi rounds, including 5 after the first round and 3 after the second round. Four items were removed on the basis of qualitative expert comments, identifying limited operational added value or risk of cognitive overload: item 5 (consider the 3Rs), item 15 (display victim categorization), and item 24 (I do: list of actions performed) after the first round, and item 39 (have the 3Rs been performed?) after the second round.
These removals were not intended to contradict consensus results but to ensure ergonomic coherence and usability of the CA. The removed items were characterized by limited added value or redundancy, illustrating the distinction between item selection through expert consensus and the subsequent development of an operational tool.
Strengths of the Study
The primary strength of this study lies in the achievement of expert consensus across multidisciplinary specialties addressing a complex and poorly standardized issue. The study is innovative, as limited data are available in the literature regarding the early medical management of MCI.Reference Timbie, Ringel and Fox 21 , Reference Borgers, Van Boxstael and Sabbe 27 By focusing on the first medical responder and the initial organizational phase, this work addresses a critical gap in current guidance.
Although developed within a specific national framework, the CA is built upon operational principles—scene size-up, triage coordination, casualty collection, and structured communication with the IC—that are common to most prehospital mass casualty response systems.
Limitations
Limitations Related to the Complexity of MCI Settings
Operational contexts vary considerably across MCIs, making standardization of recommendations and development of a universally applicable tool challenging.Reference Shang 28 , Reference Hasson, Keeney and McKenna 29 The absence of validation of the CA in real-life conditions represents a major limitation. Although simulation-based data suggest improvements in technical and non-technical performance,Reference Truchot, Balança and Wey 8 their translation to real operational settings remains uncertain.Reference Culley and Effken 13
Additionally, this CA was developed within the framework of the French prehospital emergency medical system, which is characterized by physician-led field response and a centralized command structure. These organizational specificities may limit the direct transferability of the tool to other health care systems with different models of prehospital care.
Limitations of the Delphi Method
The Delphi method relies on the expertise and representativeness of the expert panel. Panel size, disciplinary and geographic diversity, and sustained engagement throughout the process may introduce bias and limit generalizability.Reference Shang 28 –Reference Niederberger and Spranger 31
Experts were selected based on predefined criteria, including recognized roles in prehospital disaster response, formal training as Medical Incident Commander, and involvement in disaster medicine education. All were affiliated with institutions with high exposure to MCIs, including the Paris Fire Brigade and the Marseille Naval Fire Battalion. The Medical Incident Commander qualification, held by 85% of the panel, reflects nationally recognized expertise in MCI management. Combined with extensive teaching experience in disaster medicine (95%), the panel can be considered to represent a high level of operational expertise within the French prehospital system.
In accordance with methodological recommendations, 17 20 experts were invited to anticipate attrition, resulting in participation rates of 100% in the first round and 90% in the second round. Two experts did not complete the second round, potentially introducing a representativeness bias; however, no pattern suggested a structural issue with the questionnaire.
Consensus definitions vary across Delphi studies,Reference Grant, Booth and Khodyakov 32 , Reference Diamond, Grant and Feldman 33 and thresholds are rarely standardized. In this study, a median Likert score ≥7 was used in accordance with HAS methodological guidance, providing a predefined and reproducible criterion. The limitation to two rounds was chosen to balance methodological rigor with feasibility.
Limitations Related to Item Selection and Cognitive Aid Development
While the Delphi method enables consensual item selection, it does not directly produce an operational CA, necessitating subsequent merging and removal of items to limit cognitive load. No recommendations exist regarding the optimal number of items in Delphi-based tools. Open-ended questions and qualitative comments, which are infrequently reported in Delphi studiesReference Schifano and Niederberger 26 along with expert comments, guided ergonomic adaptation without inter-round meetings.
These inputs led to structuring information exchange around the IC, integration of temporal considerations, and consolidation of actions related to casualty collection points and the forward medical post.
Several elements were removed due to limited added value or redundancy to avoid cognitive overload.Reference Clebone, Burian and Tung 34 These decisions aimed to ensure usability during the initial phase of MCIs, when cognitive burden and time pressure are maximal. This step reflects the distinction between expert consensus and subsequent ergonomic design of an operational CA.
Limitations of the CMAT Methodology
The main limitations of the CMAT methodology relate to its external validity. CMAT was initially developed and validated for difficult airway management algorithms.Reference Evans, McCahon and Barley 19 Certain aspects of the tool may not fully capture the complexity of MCI response scenarios.Reference McLanders, Marshall and Sanderson 35 Inter-rater reliability of CMAT has been reported as variable, and interpretation of its items may lack standardization. In this study, CMAT was used to modify item formulation and optimize the ergonomic design of the CA.
Future Perspectives
Preventable mortality in MCIs is largely determined by the quality and timeliness of early organizational decisions made by first responders. Previous studies have highlighted that initial scene organization, casualty movement, and triage directly influence subsequent care and outcomes.Reference Shackelford, Remley and Keenan 1 –Reference Lesaffre, Tourtier and Violin 3 Delays or errors during this phase propagate throughout the entire response chain, reinforcing the importance of structured decision-making under conditions of uncertainty and time pressure.
The CA developed in this study is intended to support the first medical responder from the moment of arrival on scene. It follows a “read and do” approach, guiding sequential execution of key organizational tasks while enabling real-time collection of essential operational data, including casualty number and categorization, predominant injuries, and resource needs. These data directly support the structured scene size-up report expected within the first minutes of intervention. In more experienced teams, the tool may also be used in a “do and confirm” approach for verification and documentation.
The tool is designed for individual use, with potential sharing within the team in more structured environments. The format allows use in laminated or digital versions, although real-world validation of the latter remains to be established. Informal simulation-based exercises have demonstrated feasibility and informed minor ergonomic refinements, but the operational impact of the tool on clinical performance and non-technical skills remains to be evaluated.
Effective implementation of CAs requires prior structured training.Reference Lefèbvre, Balasoupramanien and Galant 22 Simulation-based approaches, including tabletop and high-fidelity scenarios, represent the most appropriate modality to prepare first responders for the cognitive and organizational demands of MCIs. Integration of this CA into disaster medicine training programs therefore constitutes a necessary step toward its operational deployment and future evaluation.
Conclusion
MCIs represent a major challenge for first medical responders, who are exposed to substantial cognitive overload and unusual organizational constraints. The absence of standardized recommendations justified the use of a Delphi methodology to select the key items required for the development of a dedicated CA.
Through a comprehensive review of the literature and expert consensus, this study identified relevant items tailored to the early phase of medical intervention. Ergonomic optimization using CMAT resulted in a structured, clear, and potentially usable support tool for emergency contexts.
This CA is intended as a decision-support tool and cannot replace formal training of first medical responders, which remains essential in the management of MCIs. Rather, it provides a framework of actions designed to guide initial intervention, the operational use of which will allow progressive adaptation and future refinement.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/dmp.2026.10394.
Author contribution
Concept and study design: J.G. Data collection: J.G., S.E., S.T., and A.R. Data analysis and interpretation: J.G., S.E., N.C., and S-P.C. Manuscript drafting: J.G., S.E., and E.M. Supervision: A.R. and L.G.
All authors read and approved the final version of the manuscript and agree to be accountable for all aspects of the work.
Competing interests
None.
Ethical approval
The study received an ethical opinion from the ACASAN Ethics Evaluation Committee (IRB00014579, French Armed Forces Health Academy—École du Val-de-Grâce, Paris, France).
Disclosure of AI assistance
ChatGPT (version GPT-5, OpenAI, 2025) was used to assist in the English translation and linguistic polishing under full human supervision. All intellectual content, data interpretation, and conclusions remain entirely the responsibility of the authors.
