Perspectives of Triage Team Members Participating in Statewide Triage Simulations for Scarce Resource Allocation During the COVID-19 Pandemic in Washington State

Key Points

Question  What are the experiences of individuals participating in scarce resource triage teams during the COVID-19 pandemic, and how can clinicians prepare for this role?

Findings  In this qualitative study of 41 triage team members participating in multi-institutional triage simulations in Washington state, participants described how they grappled with clinical uncertainty and ethical challenges and how the triage task could conflict with professional values and required transformation of the usual clinical mindset.

Meaning  These findings highlight challenges that triage team members may face and suggest that clinical experience, education in ethical and operational foundations of triage, and experiential training may help prepare them for this difficult role.

Importance  The COVID-19 pandemic prompted health care institutions worldwide to develop plans for allocation of scarce resources in crisis capacity settings. These plans frequently rely on rapid deployment of institutional triage teams that would be responsible for prioritizing patients to receive scarce resources; however, little is known about how these teams function or how to support team members participating in this unique task.

Objective  To identify themes illuminating triage team members’ perspectives and experiences pertaining to the triage process.

Design, Setting, and Participants  This qualitative study was conducted using inductive thematic analysis of observations of Washington state triage team simulations and semistructured interviews with participants during the COVID-19 pandemic from December 2020 to February 2021. Participants included clinician and ethicist triage team members. Data were analyzed from December 2020 through November 2021.

Main Outcomes and Measures  Emergent themes describing the triage process and experience of triage team members.

Results  Among 41 triage team members (mean [SD] age, 50.3 [11.4] years; 21 [51.2%] women) who participated in 12 simulations and 21 follow-up interviews, there were 5 Asian individuals (12.2%) and 35 White individuals (85.4%); most participants worked in urban hospital settings (32 individuals [78.0%]). Three interrelated themes emerged from qualitative analysis: (1) understanding the broader approach to resource allocation: participants strove to understand operational and ethical foundations of the triage process, which was necessary to appreciate their team’s specific role; (2) contending with uncertainty: team members could find it difficult or feel irresponsible making consequential decisions based on limited clinical and contextual patient information, and they grappled with ethically ambiguous features of individual cases and of the triage process as a whole; and (3) transforming mindset: participants struggled to disentangle narrow determinations about patients’ likelihood of survival to discharge from implicit biases and other ethically relevant factors, such as quality of life. They cited the team’s open deliberative process, as well as practice and personal experience with triage as important in helping to reshape their usual cognitive approach to align with this unique task.

Conclusions and Relevance  This study found that there were challenges in adapting clinical intuition and training to a distinctive role in the process of scarce resource allocation. These findings suggest that clinical experience, education in ethical and operational foundations of triage, and experiential training, such as triage simulations, may help prepare clinicians for this difficult role.

Surges in patient volumes during the COVID-19 pandemic have resulted in severe shortages in health care resources^1-3 and declaration of crisis capacity in several states in the United States.^4,5 Most planned protocols for allocating intensive care resources are intended to identify and prioritize patients with the greatest likelihood of survival to receive scarce resources.^6-11 While algorithms and organ failure scores have been proposed as strategies to categorize patients by estimated prognosis as part of triage plans, there is increasing concern about the effectiveness and fairness of these automated selection mechanisms.^12-16 Many allocation systems ultimately rely on clinician judgment to make final prognostic determinations.^13,17

During the COVID-19 pandemic, clinicians have reported feeling ill prepared to take on the task of triage in crisis capacity settings.¹⁸ Few clinicians receive formal education or training in approaches to health care triage.¹⁹ In fact, clinicians are typically trained to explicitly exclude considerations of resource limitation in clinical decision-making.²⁰ Nonetheless, preparation for crisis capacity has focused on the rapid deployment of triage teams, with little attention to strategies to prepare and support these team members.²¹ We aimed to elucidate the experiences and perspectives of clinicians participating in triage simulations during the COVID-19 pandemic in order to identify opportunities to better support these teams in real-world settings and to improve the allocation process.

This qualitative study was determined to be exempt from full review and requirement for informed consent because it did not constitute human participants research under the Common Rule by the Benaroya Research Institute Virginia Mason Institutional Review Board. Participants were not financially compensated. We reviewed the Consolidated Criteria for Reporting Qualitative Research (COREQ) reporting guideline in reporting the details of our methodology.

We completed a 3-phase research program to develop and test a scarce resource triage team process in Washington state during the COVID-19 pandemic. First, we conducted a Delphi study of emergency preparedness experts to develop a list of patient information items needed to inform triage team decision-making.²² Second, we held multi-institutional triage team simulations to evaluate efficiency, consistency, and effectiveness of the approach.²³ Third, we conducted a qualitative study of the triage team process and experiences of team members. In this report, we describe the results of this third component of the research program.

As previously described,^22,23 clinicians and ethicists involved in Washington state institutional emergency preparation participated in triage team simulations from December 2020 to February 2021. Triage teams consisted of at least 2 clinicians (physicians and nurses) and a team member with formal training in bioethics or experience in clinical ethics consultation.¹² Health care institutions were encouraged to refer senior clinicians and ethicists who had volunteered to participate in institutional triage teams if this were to be needed during the pandemic.

Triage team participants engaged in a 60-minute orientation session on ethical and operational aspects of triage, a 90-minute triage simulation, and a 30-minute debriefing session. During the simulation, teams reviewed a limited set of deidentified patient information items (eFigure 1 in the Supplement) and assigned each patient case to 1 of 5 prognostic categories (eFigure 2 in the Supplement). Participants also completed a survey of their demographic information and clinical background. Self-identified race and ethnicity were ascertained because these features are likely associated with participants’ personal and professional experience and perspective. Participants were asked to report their race (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or other Pacific Islander, White, prefer to self-describe, or prefer not to say) and whether or not they were Hispanic or Latino.

Simulations were silently observed by at least 3 members of the study team who took notes on team dynamics and decision-making. We used a purposive sampling approach to recruit participants for follow-up interviews, which included sampling across different simulation sessions, clinical backgrounds, and triage team roles (ie, clinician vs ethicist).²⁴ Interview participants completed a one-on-one 30-minute to 60-minute audio-recorded interview with C.R.B. (a nephrologist and researcher with expertise in qualitative methodology) or L.B.W. (a bioethicist and emergency medicine [EM] and critical care registered nurse). A semistructured interview guide was developed by C.R.B. and L.B.W. and included open-ended questions to elicit participants’ perspectives and experiences pertaining to triage team simulations (eTable in the Supplement). Throughout the period of data collection, approach to recruitment and the interview guide were iteratively refined by C.R.B. to support thematic saturation.^24,25

We conducted an inductive thematic analysis²⁶ of observation notes and written interview transcripts to identify emergent themes describing participants’ perspectives and experiences on triage simulations. Two team members with expertise in qualitative analysis (C.R.B. and K.C.V., the latter being an intensive care physician and researcher) independently reviewed and openly coded observation notes and transcripts until reaching thematic saturation (the point at which no new concepts were identified).^24,27 Both team members continued to review and code remaining transcripts to identify additional exemplar quotations. Throughout analysis, the 2 investigators used a constant comparison¹⁹ approach to iteratively review codes, collapsed codes into groups with related meanings and associations, and developed larger thematic categories.^24,26

All coauthors (C.R.B.; K.C.V.; L.B.W.; D.S.D., a bioethicist and pediatric EM physician; V.L.S., a leader in Washington disaster preparedness and an EM physician; and M.R.T., a bioethicist and critical care physician) reviewed exemplar quotations and themes and together developed the final thematic schema. We used Atlas.ti software version 8 (Scientific Software Development) to organize and store text and codes.

Among 41 clinicians and ethicists (mean [SD] age, 50.3 [11.4] years; 21 [51.2%)] women) who participated in 12 triage simulations, there were 5 Asian individuals (12.2%), 35 White individuals (85.4%), 1 individual with more than 1 race or ethnicity, and no individuals with other race responses, which included American Indian or Alaska Native, Black or African American, Native Hawaiian or other Pacific Islander, preferred to self-identify, or preferred not to say; there was 1 Hispanic or Latino individual (2.4%). Most participants were working in urban hospital settings (32 individuals [78.0%]), and the group had a mean (SD) 20.9 (11.5) years of health care experience (Table 1). In addition to data collected from simulation observation notes, we completed follow-up interviews with 21 participants to reach thematic saturation.

Three interrelated themes emerged from qualitative analysis: (1) understanding the broader approach to resource allocation, (2) contending with uncertainty, and (3) transforming mindset (Figure). Exemplar participant quotations for themes and subthemes are included in Table 2.

Participants in triage simulations strove to understand operational and ethical foundations of the triage process as a whole. They also strove to understand their roles and responsibility within this broader framework.

Participants could be uncertain or concerned about whether and how the triage process would be operationalized in clinical settings. They identified practical barriers to enlisting triage team members, such as limited staff and clinician availability, and questioned how triage and resource reallocation decisions would be communicated to clinical teams and implemented. They also anticipated that triage decisions and responsibility to implement these decisions may contribute to moral distress among clinical staff and highlighted the relevance of trust and buy-in among these clinicians, as well as strategies to support well-being.

Participants drew on their own experience and training in emergency response and bioethics to critically examine and question the design and ethical underpinnings of the triage process used in the simulation. Transparency about the triage process as a whole, including the process of development, operational features, and ethical justification, were seen as critical for fostering trust and cooperation from clinicians, triage team members, and the public. While participants may have disagreed with specific aspects of the triage approach, many participants also emphasized that they would and should accept the rules as defined to support consistency and fairness.

Participants agreed with the value of having a separate triage process to protect bedside clinicians, but they also anticipated that serving as a triage team member would be emotionally taxing and distressing. Many participants found it difficult to avoid feeling personally responsible for the tragic implications of the work. Several physician participants even described a professional obligation to accept ultimate responsibility for their decisions. Explicitly acknowledging that the triage team was only one component of a larger process of scarce resource allocation helped alleviate some of this moral burden. Clear institutional and state support was also seen as necessary to legitimize the triage team’s work.

Triage team members grappled with how to make decisions with limited clinical and contextual information about patient cases. They also struggled with ethically ambiguous features of these patient cases.

While triage team members recognized the need to limit access to extraneous patient information that may have distracted from their task or evoked biases, some found it challenging to determine prognostic estimates with the limited clinical and contextual data available to them. Critical care experience, specifically caring for patients with COVID-19, was seen as necessary for supporting pattern recognition. Nonetheless, participants sometimes struggled to develop a cohesive narrative or story for each patient or decipher one scenario among a range of possible clinical scenarios. Clinicians cited the value of quick visual assessments, or “eyeballing,” patients in their usual practice to glean intangible prognostic information. Absent a discussion with a patient’s clinical team, clinicians were sometimes uncertain how to interpret or how much to trust subjective data available to them (eg, a report of the patient’s clinical trajectory).

Many participants described a shift in the pace and approach to team deliberation over the course of the triage simulation as participants became more familiar with the process. Later in the simulation, participants were able to make prognostic determinations more rapidly and patients were discussed in the abstract, with less effort to fill in missing contextual details.

Multiple participants suggested that any ethical questions had been largely adjudicated during development of the triage process, leaving the triage team task itself relatively straightforward and dependent solely on clinical knowledge. However, others described deliberating over a range of ethical questions and uncertainties in the triage team simulation. While participants understood the triage team to be responsible for an intentionally narrow task of assigning prognostic categories, they grappled with other factors that were also felt to be ethically relevant to triage (eg, pregnancy status and duration of need for a limited resource). These features were sometimes used to nudge a patient case up or down in prognostic categories.

Clinicians found it helpful to refer back to ethical principles guiding triage to support critical reasoning through challenging patient cases. They appreciated input from ethicist team members to help clarify thinking and expose value conflicts.

Making important decisions with limited clinical information could feel irresponsible or neglectful of professional responsibilities to clinicians. Multiple participants described a default to select a more optimistic prognostic category when there was any residual prognostic uncertainty. Senior clinicians suggested that confidence resulting from years of clinical experience was valuable in making difficult decisions without becoming mired in details or indecision. Several identified the importance of simply acknowledging the unfamiliar and distressing nature of high-stakes decision-making with imperfect information to alleviate moral distress. While the triage process may not have been perfect, it offered a starting point for future iterative improvement.

Participants saw their clinical experience as valuable in making prognostic determinations. However, they also identified pitfalls of relying on their intuition and a need to substantially reshape their usual cognitive approach to patient care to better align with the unique task at hand.

Participants had been instructed to monitor for implicit biases, but they sometimes found it difficult or even contradictory to try to disentangle these biases from clinical judgment. They could be unsure of how to use information about patient characteristics, such as age and frailty, which were perceived to be associated with implicit biases, but also integral to prognostication.

Participants described active self-monitoring and explicit team discussions about how biases may shape decisions. The ethicist team member was especially active in monitoring for consistency and the association of implicit biases with outcomes of decision-making. Participants found the diversity of clinician and nonclinician team members’ perspectives to be helpful in guarding against individual biases and cultural assumptions within the medical profession.

Participants commented on how, in clinical practice, prognosis often constituted a more global assessment than physiologic survival, and this discrete outcome could be difficult to disentangle from considerations of quality of life. When considering patients with severe neurologic injury, some team members were confident in prognosticating about quality of life but felt unqualified to estimate likelihood of survival.

Understanding and making judgments grounded in the entire patient context, including patient goals and treatment preferences, was also seen as intrinsic to good medical practice. Participants sometimes caught themselves assuming an advocate role by arguing for the best conceivable prognosis for each patient case. Team members worked to segregate the narrowly defined triage team task from other clinical considerations through explicit self-reminders, team discussion, and monitoring by the ethicist team member.

While clinician team members recognized the need to adapt their usual cognitive approach, they also saw their work as an important opportunity to infuse humanity and compassion into the triage process and demonstrate respect for patients despite terrible circumstances. Some participants expressed concern about becoming too algorithmic or robotic as the simulation progressed.

When participants did not have preexisting relationships, they built trust in their teammates’ clinical skills and professionalism by observing them reason through cases. Participants recognized a potential tendency for more junior clinicians and ethicist team members to defer to senior clinicians and aimed to maintain an environment in which team members felt comfortable voicing opinions. Developing group consensus was also seen as a way of diffusing the moral burden of responsibility for a consequential task.

Clinician and ethicist team members described a role for the ethicist as a process facilitator and mediator. This team member often prompted clinicians to explain their reasoning aloud, which could motivate additional deliberation.

Some team members felt comfortable with triage after working in under-resourced countries where health care rationing was routine. Other participants identified familiar aspects of their clinical jobs that required an analogous mindset, such as prioritizing patients to receive immediate care in a busy emergency department. However, many participants also commented on how work as a member of the triage team was intrinsically different from their usual clinical role, and several team members suggested that participation in the simulation study itself had served as important practice and preparation for a real-world triage scenario.

Thematic analysis of triage team simulation observations and interviews with participants in this qualitative analysis revealed a clinically, operationally, and ethically complex process of triage team deliberation. During the COVID-19 pandemic, clinicians across a range of health care settings have struggled to adapt clinical decision-making to the realities of resource limitation.^1,2,13 While triage team participants were presented with a seemingly narrowly defined task focused on fair allocation of a scarce health care resource across the population, they nonetheless were reluctant to shift their focus entirely from patient-centered considerations and endeavored to maintain compassion for individual patients in the triage process. The teams’ task drew on participants’ existing skills and experience, but the task could also feel unfamiliar and even antithetical to participants’ professional values and required a transformation of their usual approach to decision-making.

Triage team members’ clinical experience, especially in critical care, was seen as important to support pattern recognition with limited information, comfort making decisions amid uncertainty, and acceptance of a morally weighty task. However, in addition to factual clinical knowledge, this work emphasized the relevance of how team members think and ways in which the deliberative process itself may support an effective team.^28-30 Patient cases presented to triage team participants were stripped of the usual rich contextual detail that informs unconscious pattern recognition and clinical intuition.^31,32 For this reason and because team members were explicitly asked to monitor for their own biases, participants were required to engage in a more analytical and self-aware approach to decision-making.³¹ Patient cases involving clinical and ethical uncertainty required critical thinking and reasoning.

Participation on hospital triage teams is likely to evoke ethical conflict and moral distress, not only because of the tragic circumstances under which these teams are deployed, but also because clinicians are asked to perform a task that may deeply conflict with their training and professional values (Table 3).^33-35 This tension is reflected in the broader experiences of clinicians working during the pandemic.³⁶ Moral distress has been associated with significant burnout and degraded mental health among clinicians during and before the COVID-19 pandemic and likely poses a challenge to preserving both triage team membership and the broader work force.^37,38

Insights from this work suggest opportunities to support triage team members and to improve the deliberative process. First, equipping team members with a broad understanding of clinical, operational, and ethical foundations of the entire process of resource allocation^17,39 may help them to understand their specific role and may help alleviate the burden of personal moral responsibility. Second, an ethicist or team facilitator may serve a valuable role in mediating and monitoring deliberation to ensure that the team operates under consistent principles and goals.⁴⁰ Diversity of perspectives and a trusting team environment may also be associated with improved active discussion and carefully considered decisions. Finally, our findings suggest that advance training should include cognitive approaches to bias reduction⁴¹ and basic bioethical reasoning. While it may be difficult to fully prepare clinicians for the reality of participating in triage,¹⁸ practice in simulations may help team members apply and develop skills. Simulated practice may also support clinicians in being more willing to volunteer^42,43 and able to quickly adapt to an unfamiliar role.²¹

Our study has several limitations. First, this high-fidelity simulation was conducted among clinicians embedded in an ongoing health care emergency, but the experience of participating in a true triage scenario may differ. The roles and composition of triage teams may also differ among scarce resource allocation plans,¹² and the process should be adapted to unique institutional and regional contexts. Mirroring the racial composition of physicians in Washington state,⁴⁴ the majority of participants in our study were White, and findings may not reflect the perspectives of clinicians with different racial backgrounds.

Thematic analysis of observation and interviews with participants in multi-institutional triage team simulations suggests that there is value but also challenges to the triage team model as a component of scarce health care resource allocation. These findings highlight the need for triage team members to have extensive clinical experience and other expertise, such as bioethics training. Education in ethical and operational foundations of scarce resource allocation and experiential training, such as triage simulations, may help prepare team members to perform a challenging task in advance of clinical deployment.

Accepted for Publication: February 28, 2022.

Published: April 18, 2022. doi:10.1001/jamanetworkopen.2022.7639

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Butler CR et al. JAMA Network Open.

Corresponding Author: Catherine R. Butler, MD, MA, Division of Nephrology, Department of Medicine, University of Washington, 325 Ninth Ave, Seattle, WA 98104 (cathb@uw.edu).

Author Contributions: Dr Butler had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Butler, Webster, Diekema, Gray, Sakata, Tonelli.

Acquisition, analysis, or interpretation of data: Butler, Webster, Diekema, Gray, Tonelli, Vranas.

Drafting of the manuscript: Butler, Vranas.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Butler.

Obtained funding: Butler.

Administrative, technical, or material support: Butler, Gray, Sakata, Tonelli.

Supervision: Butler, Diekema.

Conflict of Interest Disclosures: None reported.

Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs or the US government.

Additional Contributions: The authors would like to acknowledge Josh Edrich, MPH (Northwest Healthcare Response Network), for his contributions to study coordination. They also thank the study participants, whose efforts made this work possible. The contributor and participants were not compensated for this work.