Abstract
Background: Early resuscitation with blood components or products is emerging as best practice in selected patients with trauma and medical patients; as a result, out-of-hospital transfusion (OHT) programs are being developed based on limited and often conflicting evidence. This study aimed to provide guidance to Canadian critical care transport organizations on the development of OHT protocols.
Methods: The study period was July 2021 to June 2022. We used a modified RAND Delphi process to achieve consensus on statements created by the study team guiding various aspects of OHT in the context of critical care transport. Purposive sampling ensured representative distribution of participants in regard to geography and relevant clinical specialties. We conducted 2 written survey Delphi rounds, followed by a virtual panel discussion (round 3). Consensus was defined as a median score of at least 6 on a Likert scale ranging from 1 (“Definitely should not include”) to 7 (“Definitely should include”). Statements that did not achieve consensus in the first 2 rounds were discussed and voted on during the panel discussion.
Results: Seventeen subject experts participated in the study, all of whom completed the 3 Delphi rounds. After the study process was completed, a total of 39 statements were agreed on, covering the following domains: general oversight and clinical governance, storage and transport of blood components and products, initiation of OHT, types of blood components and products, delivery and monitoring of OHT, indications for and use of hemostatic adjuncts, and resuscitation targets of OHT.
Interpretation: This expert consensus document provides guidance on OHT best practices. The consensus statements should support efficient and safe OHT in national and international critical care transport programs.
The transfusion of blood components such as red blood cells (RBCs) and plasma is increasingly common in prehospital and transport medicine.1–3 In addition, the potential benefits of out-of-hospital administration of whole blood or blood products such as fibrinogen and prothrombin complex concentrate in selected patients are being investigated. In this report, we use the umbrella term “out-of-hospital transfusion” (OHT) to refer to the transfusion of whole blood, blood components such as RBCs and plasma, or blood products such as fibrinogen and prothrombin complex concentrate. Although the increasing practice of OHT suggests general consensus on a likely clinical benefit, evidence regarding the effect of OHT on morbidity and mortality is limited and conflicting.2,4–6 The generalizability of the limited evidence is further complicated in that the feasibility and potential benefit of OHT are dependent on multiple regional factors such as geography, patient factors and health care configuration. For example, 2 secondary analyses of the data sets from the Prehospital Air Medical Plasma (PAMPer) and the Control of Major Bleeding After Trauma (COMBAT) clinical trials suggested that OHT was beneficial if transport times were greater than 20 minutes and that a benefit present in blunt trauma does not translate to a benefit in penetrating trauma.7,8 In addition, out-of-hospital management of acute hemorrhage extends beyond OHT and includes factors such as administration of tranexamic acid, avoidance of hypothermia and physical means of hemorrhage control where possible.9,10 Efficient and effective implementation of OHT requires a combination of medical and logistic considerations that span multiple specialties. This is particularly relevant in countries like Canada, with long transport times to tertiary care centres, and remote communities that have limited or no access to physicians or blood components and products at their local health care facilities.11
We invited an expert panel to provide expert opinions on out-of-hospital hemorrhage management and, in particular, OHT to develop national consensus recommendations to guide OHT practice and to begin to optimize the effectiveness and safety of OHT.
Methods
We used a modified RAND Delphi process to create an expert consensus document on the development of OHT protocols by Canadian critical care transport organizations (CCTOs). The study period was July 2021 to June 2022.
Study design
We used a modified RAND Delphi process to establish recommendations for the development of local or regional OHT protocols. The Delphi technique is deemed a relevant source of evidence in health care research and is particularly important if randomized controlled trials are unavailable to set health care policies.12 It is a systematic, interactive method that relies on a panel of experts to converge on consensus statements after a series of iterative written surveys.13 Based on the study team’s experience with a recent Delphi study on in-hospital massive hemorrhage protocols,14 we modified the original technique by adding a panel discussion to the written survey rounds. This was to allow an exchange of information and opinions between participants of different backgrounds and levels of expertise. We also chose the RAND/UCLA Appropriateness Method,13 in which the participants were encouraged to edit the list of recommendations during the written survey rounds, as well as add further recommendations or comments in free-text fields.
We did not specify the number of written survey rounds a priori. Based on previous similar research,14 we estimated that 2–4 written survey rounds would be required to achieve saturation for feedback and stagnation for consensus. The study team reviewed all feedback and progress toward consensus after the second written survey round to decide whether further written rounds would be of benefit. Given the considerable geographic distance between participants, we used an online survey tool (JotForm, https://www.jotform.com/) for the written survey rounds of the Delphi study and an online meeting platform for the panel discussion.
Data source
At the start of the process, the study team created a list of 41 statements relating to OHT, covering the following domains:
General oversight and clinical governance
Storage and transport of blood components and products
Initiation of OHT
Types of blood components and products
Delivery and monitoring of OHT
Indications for and use of transfusion adjuncts
Resuscitation targets to guide transfusion.
The lead author (J.V.-F.) drafted the initial statements and domains based on clinical experience using OHT in CCTOs in the United Kingdom and Canada, after which B.N., J.L. and S.M. each provided written comments and revisions, resulting in a second draft. The third and final draft was agreed on during a meeting of the entire study team (B.N., J.V.-F., J.L., S.M.).
Participants
The study team created a list of subject experts for study participation from personal contacts, with the following inclusion criteria: senior clinician in a CCTO, or in-hospital trauma care with an interest in transfusion or in a transfusion service involved in OHT, and current clinical practice in Canada. In addition, potential participants on this list were given the option to nominate further experts for potential participation in the study. During this selection process of potential study participants, we use purposive sampling based on professional background, clinical specialty, and location of practice. Given the relatively small pool of eligible experts in Canada, we sought a sample size of 15–20 participants to achieve good representation. 15 Potential participants were contacted via email, with 2 further follow-up emails in 2-week intervals. The recruitment email contained a short summary of the study objective and design (Appendix 1, available at www.cmajopen.ca/content/11/3/E546/suppl/DC1), and participants completed a written consent form for participation. There was no financial remuneration. The study team did not participate in the written survey rounds; B.N. and J.V.-F. moderated the panel discussion but did not express opinions on statements discussed.
Delphi process and statistical analysis
In each written survey round, participants were asked to score each of the recommendation statements on a Likert scale ranging from 1 (“Definitely should not include”) to 7 (“Definitely should include”). Participants were also asked to propose wording changes to existing statements, add comments or add additional statements they considered important. Participants were blinded to the other participants’ identities and responses during the written survey rounds. Once all participants had submitted their ratings and comments, the research team calculated median Likert scores for each statement and reviewed all comments. The research team was blinded to the identity of the participants during this phase of the Delphi process. The following outcomes were possible after each written survey round:
Median score 6–7: consensus achieved. Statement included as written, or with minor adjustments based on participants’ comments if these changes did not alter the meaning of the statement. These statements were excluded from further rounds.
Median score 6–7 with critical commentary: if 1 or more participants suggested relevant changes to a statement that changed some or all of the original meaning, these changes were incorporated, and the revised statement was included in the next round.
Median score 3–5: the research team reviewed the participants’ comments and updated the relevant statements accordingly. All statements were included in the next round.
Median score 1–2: unless there were participants’ comments clearly in favour of these statements, they were considered as rejected by the panel and removed from the process.
Merging of 1 or more existing statements: if participants’ comments suggested a substantial improvement of statements by merging them into 1 item, the resulting merged statement was then included in the next round.
New statements: new statements suggested by participants were added in their respective domain and included in the next round.
This process was repeated until the study team determined that there was stagnation of consensus and saturation of information from the free-text feedback. Only statements requiring further review to achieve consensus (median Likert scale score 6–7 with critical commentary or median score 3–5) were reviewed in an online meeting of participants (round 3), which allowed discussion and clarification of statements. For technical reasons, participants’ identities and responses were not blinded during the panel discussion. The meeting was recorded and transcribed by an automatic transcription service, and the recording and transcription were made available to all participants. If participants preferred to remain anonymous, they were given the option to not actively participate in the panel discussion but, rather, to review the recording and transcription and provide written feedback to the study team. All participants (including those who were unable to attend the virtual meeting) were then asked via email to review the recording or transcription and indicate whether the statements crafted during the meeting should be included in the document. Consensus in the online meeting was defined as agreement by all participants to include a given statement.
Presentation of results
The final agreed-on statements were included in a table with domains and order of statements updated based on participants’ feedback. In addition, the study team drafted a rationale for each statement, based on current literature and participants’ comments during the Delphi rounds. All participants were given the opportunity to review the recommendation statements and the corresponding rationales in their final form before completion of the study.
Authorship
After reviewing the participants’ contributions to the research project, the study team decided to offer coauthorship to all participants during the final Delphi round. All participants consented to authorship and reviewed the final manuscript.
Ethics approval
Research ethics board review and approval was provided by the Research Ethics Office, Unity Health Toronto (REB 21-155).
Results
We invited 29 subject experts, of whom 17 (7 females [41%] and 10 males [59%]) agreed to participate in the study. All participants held senior positions within their respective organizations. Table 1 provides an overview of the participants’ backgrounds.
Of the 12 subject experts who did not participate, 1 declined and 11 did not respond. As part of the purposive sampling strategy, the study team attempted to recruit additional participants from provinces that were underrepresented, but these efforts were ultimately unsuccessful. Overall, the study team identified 21 subject experts, and a further 8 were nominated by potential participants.
After reviewing the results of the written survey rounds 1 and 2, the study team concluded that no further progress on consensus could be achieved through further written rounds, and we proceeded to the online panel discussion. The final modified RAND Delphi structure used in this study therefore consisted of 3 rounds: 2 written surveys of recommendation statements, followed by a panel discussion. All participants completed rounds 1 and 2 of the modified Delphi process, and 13 participants (76%) attended the virtual panel meeting (round 3). All participants who were unable or wished not to attend the virtual meeting reviewed the recording or transcription, or both, and provided further commentary if required. All 17 participants reviewed the final list of statements. Table 2 shows the progression toward consensus for all the statements.
Of the 41 initial statements, 21 were accepted with no or only minor changes after round 1, and 5 were merged with others. The remaining 15 statements were modified according to participants’ comments and included in round 2, together with 7 additional new statements suggested by participants. In round 2, a further 9 statements were accepted, and 2 statements were merged with another. No statements received median scores of 1 or 2 in either of the first 2 rounds. The remaining 11 statements were discussed in the virtual panel meeting. During the panel meeting (including feedback from participants who were unable to attend), consensus was achieved on all but 2 statements. Table 3 contains a comprehensive list of the 39 final consensus statements and their rationales, Box 1 lists the 9 quality metrics, and Table 4 shows the 2 statements for which no consensus was achieved.
Suggested quality metrics for quarterly review by the critical care transport organization’s medical advisory committee
Strongly recommended
Number of wasted blood components and products (absolute number and proportion of total blood components and products)
Transfusion-related errors (i.e., ABO/Rh incompatibility, compromised blood products)
Independent double checks of blood components and products
Proportion of patients receiving OHT who met protocol indications
Proportion of blood components and products successfully traced to final disposition (i.e., transfused, returned to transfusion services, wasted)
Recommended
Proportion of patients with OHT where receiving facilities were notified of need for further in-hospital transfusion, before arrival (pre-alert).
Proportion of patients who received tranexamic acid within 1 h of first contact with CCTO (if within 3 h of injury or acute postpartum hemorrhage)
Proportion of patients who had temperature of > 35°C by time of arrival at receiving hospital
Proportion of patients of child-bearing potential who received O Rh(D)-negative RBCs
Note: CCTO = critical care transport organization, OHT = out-of-hospital transfusion, RBC = red blood cell.
Interpretation
Through a modified RAND Delphi process, we developed 39 expert consensus statements and 9 quality metrics on the transfusion of blood components and products in the prehospital and retrieval setting. This guidance document specifically addresses OHT and the CCTOs responsible for implementing and assuring the quality of OHT. Although some of the guidance in this document is specific to the Canadian setting, to the best of our knowledge, this is one of the very few documents providing guidance on OHT internationally.66 We hope it will prove useful to CCTOs in Canada and other countries around the world. The consensus statements cover various aspects of OHT, from logistics to clinical aspects and quality-assurance measures. As such, we consider the multidisciplinary makeup of the expert panel participating in the study to be an important strength of this research.
The 2 domains for which gaining consensus was more challenging were domains 3 (initiation of OHT) and 4 (types of blood components or products). This slower, and, in the case of 2 statements, failed progress toward consensus in these domains likely reflects the lack of clear evidence and considerable variation in practice in these areas.67 From our experience during this modified RAND Delphi process, we stress the benefit of an exchange of information among the subject experts, particularly between patient-facing clinicians and transfusion specialists, as well as the importance of striking a balance between specific and flexible guidance statements.
The importance of dialogue among subject experts is reflected in several statements that found consensus only after the panel discussion. In particular, statements in domain 4 did not achieve consensus (or, in 1 case, rejection) until round 3. Transfusion medicine experts were able to outline the current estimates of the risk of Rh(D) sensitization, which was considerably lower than many patient-facing clinicians had assumed.44 On the other hand, logistical considerations, the higher proportion of patients of child-bearing potential receiving OHT in some participants’ CCTOs, and the higher risk of errors in the critical care transport setting compared to in-hospital practice resulted in agreement to primarily recommend O Rh(D)-negative RBCs for CCTOs. Other important discussion points during the panel meeting were the limited availability of plasma49 — which contrasted with a desire by many patient-facing clinicians to stock blood components and products that could provide clotting factors and volume7 — and the consideration of alternatives to plasma, such as prothrombin complex concentrate and fibrinogen.14
Statement 3.1, regarding the indication to commence OHT, can be seen as an example of the panel’s attempt to balance specific guidance with flexibility. Although there are multiple scores and algorithms to predict the requirement for massive transfusion for patients with trauma in the emergency department, none of the current methods to decide on which patients benefit from early transfusion in trauma achieve particularly high specificity or sensitivity.38 In addition, most of these scores have not been validated in the prehospital setting or in nontraumatic causes of major hemorrhage. The authors of a recent systematic review on the topic concluded that the process to trigger major hemorrhage protocols should be “individualized to hospital resources and skill set to aid clinical judgment.”38 This conclusion holds particular truth in the context of OHT in the setting of the unique geographic challenges faced by CCTOs in Canada. The patient population requiring OHT might be as diverse as a patient with trauma transported via a 30-minute flight from the scene of an accident to the nearest trauma centre, a patient with a perioperative major hemorrhage in a smaller hospital requiring a 90-minute interfacility transfer to the nearest tertiary care centre, or a patient with postpartum hemorrhage in a remote nursing station with no access to blood products or laboratory testing, and transport time exceeding 2 hours.11,30 We believe that this expert consensus document can help to overcome such challenges through a nationwide approach to OHT protocols that provides specific guidance while taking into account the variability in geography, patient factors, in-hospital and prehospital blood product availability, and other available resources.
Regarding the statements of domain 4 (types of blood components and products), currently there is a lack of solid evidence to support strong recommendations. However, research in this field is developing at a steady pace.68–70 Although we attempted to incorporate a level of flexibility to accommodate this limitation, this guidance document will need to be reviewed and updated in the future, in keeping with statement 1.5. For example, the Prehospital Lyophilized Plasma (PREHOP-PLYO) trial,68 comparing OHT of lyophilized plasma to normal saline in patients with trauma, was published shortly after we completed our Delphi rounds. However, given the small sample (150 patients) and the current unavailability of lyophilized plasma in Canada, the results of that trial have no immediate impact on this guidance document. Appendix 2 (available at www.cmajopen.ca/content/11/3/E546/suppl/DC1) provides an overview of ongoing trials that will provide relevant results over the coming years.
Importantly, we consider this document a starting point rather than an end product in the process of ensuring consistent and equitable access to blood components and products for all patients, irrespective of geographic location. Although outside the scope of this project, we have created a national collaboration and OHT working group with all Canadian CCTOs to assure that processes are aligned as much as possible across the Canadian provinces, that emerging evidence and new technology are reviewed in a timely and efficient manner, and that quality-improvement measures are shared across organizations. This collaboration will also ideally include a pan-Canadian OHT registry with consistent data entry from all participating CCTOs for quality assurance and future research projects. This registry will allow us to measure adherence to these recommendations by Canadian CCTOs over the coming years.
Limitations
Our modified RAND Delphi study achieved representation from major relevant clinical specialties and a wide geographic distribution. However, we were not able to recruit clinicians from every Canadian province, and there was a lack of representation from obstetricians and patient representatives. As with any self-selecting group of experts, there is the risk of recruiting only participants with similar opinions. Based on participants’ comments during the survey rounds and panel discussion, the study team was reassured that a wide range of opinion was captured during the study process. Although direct participation of patient representatives would have been challenging given the very specific focus of the study, we could have involved patient and public representatives in the planning stages of the research. Finally, no pediatric specialists participated in this research, and we did not provide any specific guidance on the pediatric population. Although many of the principles in the document can be applied to pediatric patients, we recommend involving local pediatric specialists when creating OHT guidelines for this population.
Conclusion
This nationwide consensus document covers a wide range of important domains in the development of OHT protocols. It should support CCTOs in establishing and standardizing OHT, to ensure efficient and equitable use of this valuable resource.
Acknowledgements
The authors thank Melissa McGowan, Emma O’Neil and Richard Wu for their administrative support during the course of this research.
Footnotes
Competing interests: Brodie Nolan reports research funding from Canadian Blood Services and Physicians’ Services Incorporated Foundation. Andrew Shih reports payments for consulting on educational materials in relation to bleeding management from Octapharma Canada and payments for advisory board participation in relation to clotting factor concentrates from CSL Behring. He has received speaker honoraria from Octapharma Canada and CSL Behring. Octapharma Canada reimbursed travel expenses for attending a meeting for the FARES-II trial comparing clotting factor concentrates to plasma in cardiac surgery. He is vice-chair of the National Advisory Committee on Blood and Blood Products. No other competing interests were declared.
This article has been peer reviewed.
Contributors: Johannes von Vopelius-Feldt, Joel Lockwood, Sameer Mal and Brodie Nolan contributed to the study conception and design, and analyzed the data. Johannes von Vopelius-Feldt and Brodie Nolan drafted the manuscript, with contribution from Adam Greene. Andrew Beckett, Jeannie Callum, Adam Greene, Jeremy Grushka, Aditi Khandelwal, Yulia Lin, Susan Nahirniak, Katerina Pavenski, Michael Peddle, Oksana Prokopchuk-Gauk, Julian Regehr, Jo Schmid, Andrew Shih, Justin Smith, Jan Trojanowski, Erik Vu and Markus Ziesmann contributed to data collection and interpretation. All of the authors revised the manuscript critically for important intellectual content, approved the final version to be published and agreed to be accountable for all aspects of the work.
Funding: Brodie Nolan received funding for this study from the St. Michael’s Hospital Medical Services Association Innovation Fund (no. SMH-22-013).
Data sharing: Anonymized data can be shared on reasonable request by contacting the corresponding author.
Supplemental information: For reviewer comments and the original submission of this manuscript, please see www.cmajopen.ca/content/11/3/E546/suppl/DC1.
This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) licence, which permits use, distribution and reproduction in any medium, provided that the original publication is properly cited, the use is noncommercial (i.e., research or educational use), and no modifications or adaptations are made. See: https://creativecommons.org/licenses/by-nc-nd/4.0/
References
- © 2023 CMA Impact Inc. or its licensors