We used a modified nominal group technique (NGT), a consensus methodology widely used in emergency medicine, to identify high priority ethical challenges relevant to the implementation of precision medicine in the emergency medicine setting and arrive at a consensus framework [24-31]. This study followed the COREQ (Consolidated Criteria for Reporting Qualitative Research) reporting framework for qualitative studies [32].

NGT is a 4-step process in which stakeholders generate and then prioritize questions or research ideas (Figure 1) [33]. NGT was well-suited for our research question as our aim was to generate a comprehensive set of ethical challenges which could then be organized and ordered. This is in contrast to other frequently used consensus techniques such as the Delphi technique, which are designed instead to reach a consensus opinion on one particular question [24]. NGT has previously been used successfully to identify research priorities in the field of emergency medicine [25,27,34-37].

A NGT session group size is recommended at greater than 6 participants, with the ideal between 9 and 12 [34]. As precision emergency medicine is an emerging field, we sought participants who had familiarity with the clinical practice of emergency medicine, including an understanding of the types of data that are collected throughout a patient’s journey and the ways in which data is used for clinical decision-making. Participants were recruited through email to the Society of Academic Medicine Ethics Committee listserve; the email invitation also asked recipients for referral and sharing to others who may be interested. A quota sample of the first 14 respondents to the email were registered as potential participants in the NGT group to meet the ideal group size and allow for scheduling conflicts. In total, 12 emergency medicine physicians who reported working knowledge of precision medicine topics, as evidenced by publication records and participation in precision medicine conferences or similar activities, joined our NGT session. Participants represented a spectrum of subspecialties (education, ethics, pediatrics, diversity, equity, inclusion, and informatics) across various stages of training (resident, early-, mid-, and late-career) and practice settings (county hospital, community hospital, academic center, and integrated managed care consortium) (Specialty, Level of Training, and Practice Setting included in Table S1 in Multimedia Appendix 1) to provide a range of emergency medicine practice perspectives.

Participants took part in the NGT session during the American College of Emergency Physicians annual conference in Boston in October 2021. The process was moderated by the lead researcher of the group (CR), and notes were taken by 2 separate researchers (ES and MG).

This study was approved by the Stanford University Institutional Review Board (IRB #62687); participants gave written informed consent before taking part in this study. All data were collected anonymously. Participants received a US $125 gift card in compensation for their time.

In this study, we have identified critical ethical challenges facing the implementation of precision medicine in the unique practice environment of the emergency department. The growing use of vast amounts of genomic, clinical, and lifestyle data to guide individualized therapies holds immense potential to improve emergency care. However, the “anyone, anywhere, anytime” mandate of emergency medicine presents distinct hurdles for the deployment of precision approaches that rely on the timely availability of complete patient information in a fast-paced and high-acuity setting. Using a modified nominal group technique, we engaged a diverse panel of emergency physicians to develop a consensus framework mapping key ethical challenges along the continuum of precision medicine implementation, from data acquisition to use in clinical practice to monitoring of downstream consequences. Crucially, this framework highlights the need to proactively identify and mitigate potential pitfalls to best realize the promise of precision medicine in the ED while promoting equitable access and benefit for the diverse patient populations we serve. Our findings provide a framework to guide research, practice innovation, and policy aimed at the ethical translation of precision medicine to the front lines of emergency care.

A recurring theme in our NGT discussion was “whose values are we accounting for?” [40]. Participants acknowledged the sometimes conflicting needs of the individual versus those of the system. Providers may recommend medications that prolong survival, but side effects may be undesirable or cost prohibitive for some patients [41]. The definition of high-quality care envisioned by precision medicine may not align with the individual goals of patients. There are multiple stakeholders driving the discourse of precision medicine—health care providers, health care insurers, regulators, and industry—the least empowered of which is the patient [40]. There were many questions raised regarding challenges to patient autonomy and the ambiguity of who is the ultimate decision maker in precision medicine and how that might change over time, as the needs of the many may be at odds with the needs of a few.

This tension maybe at odds with the ethical principle of autonomy [42]. The definition of respect for autonomy requires that patients are able to make their own decisions while incorporating their own values. This potentially extends to the use of their data and specimens, but the limits of patient autonomy remain under debate [43-45]. For example, the appropriate approach to consent for enrollment in biobanks that start in emergency departments remains under debate [46,47].

Ensuring that patient autonomy is maintained can be particularly challenging in the emergency care context. In the emergency department, providers make decisions with patients in a time-constrained setting, particularly compared with those historical clinical environments of precision medicine, such as oncology, which may occur over weeks. Not only does this limit time for discussion, it may also create a sense of external pressure to make a decision [48]. In such time-sensitive situations, patient personal information and preferences may be disclosed and decisions required under duress, thus also implicating privacy [49].

Privacy includes not only the protection of one’s identity and personal information, but also the ability to make choices without them being disclosed to others [50]. Another challenge to the patient’s privacy is that the practice of precision medicine in the emergency department will require massive amounts of data transiting among and across institutions and devices. Ensuring the safety and security of this protected information is a major hurdle for precision medicine initiatives, and the abovementioned time-sensitive nature of the emergency department encounter makes it necessary to share this data in real time, which is unique from other medical specialties [51-54]. In addition, when patients know that the information they provide to the system will be used to make recommendations for themselves or others, it may have downstream effects. For example, alcohol use and history for the allotment of liver transplant is one area that may leave patients in a moral quandary, unsure whether to be open with their behavior but fearing the effects of those disclosures [55,56]. Patient perceptions of the trustworthiness of health care providers and institutions will impact the success of any scientific or research endeavor, particularly among historically disadvantaged populations [57].

It is no surprise that our consensus framework includes justice as a key area for research in precision medicine ethics as it grows in the emergency department and in the prehospital setting [58-60]. Emergency physicians carry the distinct duty under federal law to provide care to anyone who presents for emergency care [61]. This duty is widely recognized to rest on a fundamental moral responsibility to provide care “promptly and expertly, without prejudice or partiality” [49]. At first it may seem that precision medicine would help to limit personal biases which might affect care, using data as opposed to gestalt or local practice variation. But structural biases may encode results and outcomes on patients that might inadvertently limit the benefits of the precision medicine model. When they are present in the initial development of the model, such as lack of representation and personal biases, affects the validity of results used to make targeted recommendations [62].

Justice, or more specifically with respect to medical ethics, distributive justice, emphasizes that resources are to be distributed fairly, equitably, and appropriately. This can be heavily hindered if providers rely on precision models to make recommendations without a critical appraisal of how well the results fit the patient in front of them. Even if done well, the outputs of precision medicine may not be available to all [63]. Precision medicine which is ultimately only accessible to a few but built on the data of many, would be an affront to the growing field of social emergency medicine [59]. In oncology, an area with a much longer history of precision medicine, 78% of patients and providers reported concerns regarding the accessibility of precision medicine [64].

As noted by participants, challenges in implementing precision medicine in the emergency medicine context occur anywhere from when data is collected (or not collected) to the downstream effects of using precision medicine interventions. In emergency medicine, providers stand prepared to address a range of conditions at any time. We often must rely on incomplete patient histories (eg, due to the severity of illness or mental status changes frequently encountered in emergency medicine) and limited access to previous medical records (eg, our patients may be uninsured, underinsured, and without access to routine primary care providers) [65]. Precision medicine approaches will need to be deployable in a timely manner and maintain their precision with only the limited patient-specific information available to emergency physicians.

Furthermore, even when data are available, the actualization of precision medicine in the care environment may serve only to perpetuate systemic flaws in the health care system if that data is inaccurate, biased, or exclusionary from the start [7,66]. A 12-year review of the National Hospital Ambulatory Medical Care Survey (NHAMCS) illustrates this risk: Black patients were 10% less likely to receive immediate or emergent Emergency Severity Index (ESI) scores and to be admitted, but 1.26 times more likely to die in the emergency department or the hospital than White patients [65]. Using such historically biased data to make precision medicine recommendations about resource allocation could amplify these disparities. While precision medicine aims to deliver individually tailored care, it still requires categorizing patients into discrete groups by shared characteristics—a process that carries its own implications and potential for bias [4,67].

Additional logistic concerns for the actualization of precision medicine in emergency departments, include costs, legal issues, and patient education. Efforts made to collate data from emergency care settings to be used in precision medicine research may result in products or treatment pathways unaffordable to these patients [68-70]. This may be mitigated through insurers [71], but the emergency department is a safety net, and as such this approach may not benefit the most vulnerable patients while those who are insured may find that their plans have little incentive to offer this option [72]. Research has shown that the cost of precision medicine is a concern shared by patients and providers alike [63,64].

Even when payment is not an issue, a data- and machine-supported approach may not be conducive to truly informed consent given the limitations of adequately educating emergency department patients regarding how precision medicine and its associated research works [63,73]. Beyond the challenge of consenting in the emergency context, providers must also consider issues regarding disclosing incidental information and the potential changes associated with the use of data over time which interconnected to other growing datasets. Stewardship of these data, defined as “the responsible management of something entrusted to one’s care” [74], will be critical to maintaining trust of patients and communities as this science evolves [57].

Discussion in this NGT process highlighted the need to actively maintain a holistic view of the patient without being overly reductive to these categories. Currently, only a limited amount of data is collected with respect to health measures or lab studies, and even less is recorded for social determinants of health. Future work should focus on systematic collection of environmental and social factors in emergency care settings, including neighborhood-level data on environmental exposures, community resources, and barriers to care access. However, this data collection must be done thoughtfully to avoid reinforcing biases—for example, by ensuring community engagement in data collection methods, using validated measurement tools that have been tested across diverse populations, and analyzing social factors as modifiable system-level variables rather than immutable patient characteristics. In emergency medicine specifically, this could include developing standardized screening tools for social needs that are feasible in acute care settings, creating automated systems to gather community-level data that do not burden individual patient encounters, and establishing partnerships with community organizations to better understand local health contexts. These approaches could help create more comprehensive precision medicine models while maintaining emergency medicine’s core mission of providing equitable care to all patients.

The use of a NGT and participant diversity are key strengths of this study. However, several limitations warrant discussion. While NGT aims to limit the biases associated with other consensus models, expertise bias is still a potential limitation, as all participants were aware of and knowledgeable on this topic beforehand. While this expertise strengthens the confidence in the working knowledge of the topic, it can also lead to anchoring. We took several steps to limit its impact. NGT, by way of its structured format, is meant to limit the influence of any one dominant individual in discussion [39]. In addition, we included a diverse spectrum of emergency physicians from residents to attending physicians with varying years of experience, as well as emergency physicians with a variety of subspeciality training.

Self-selection bias should also be considered, as participants volunteered in response to an email invitation to the Society of Academic Medicine Ethics Committee listserve. This recruitment method may have attracted participants with stronger interest in ethics and precision medicine, though their engagement likely enhanced the depth of insights generated during the NGT session.

Finally, though the number of participants was within ideal range for an NGT [24], the relatively small group size remains a limitation. Future research could benefit from additional strategies that would incorporate a higher number of emergency physicians and experts from outside of the field which might offer perspective to our work.

Realizing the promise of precision medicine’s “right person, right place, right time” must be balanced by the context of the “anyone, anywhere, anytime” of emergency medicine. Our work has illuminated a range of concerns for the development of precision emergency medicine, and provides a focused framework for the most pressing ethical considerations along the continuum of data acquisition to implementation. This framework may be used to direct research and innovation toward addressing these challenges in the emergency medicine setting.