“The Most Difficult Conversations We Have”: Resident Experiences With Pediatric Neuroprognostication

ABSTRACT

Objective

Pediatric neuroprognostication is often complicated by uncertainty and embedded in high‐stakes medical decisions. For child neurology residents, participation in neuroprognostication can represent key learning experiences but also comes with emotional and ethical challenges, which poses a risk of moral distress. This study sought to describe contextual influences and impacts on residents participating in neuroprognostication.

Methods

One‐on‐one, semi‐structured virtual interviews were conducted with senior (postgraduate year 5–6) child neurology residents. Deidentified transcripts were iteratively analyzed to develop themes and generate a narrative of resident experiences. Initial themes were shared with participants for additional feedback. Final themes were synthesized to characterize contextual influences and the emotional and professional impacts of participating in neuroprognostication.

Results

Fifteen residents from 12 institutions across 10 states participated, representing a range of program sizes and geographic locations across the United States. Qualitative analysis identified five themes: (1) Serious neurologic pathology and decisions with high stakes, highlighting neurology teams' roles in major medical decisions; (2) Navigating uncertainty and hope, describing the difficulty of limited data; (3) Systemic barriers and challenges to prognostic communication, underscoring team‐ and institution‐level factors; (4) Negative emotional impacts, with concerns of developing emotional exhaustion; and (5) The importance of neuroprognostication, emphasizing the gravity of this responsibility and distress when residents felt ineffective in this role.

Conclusions

Residents describe neuroprognostication as an emotionally salient experience carrying a strong sense of responsibility. This narrative highlights opportunities to support residents in training, improve prognostic communication, and address potential threats to resident well‐being.

1. Introduction

Estimating and communicating neurologic prognosis are central to the care provided by child neurologists. Complicating this effort, neuroprognostication frequently entails significant uncertainty [1] and is often embedded in high‐stakes medical decisions at times of emotional distress for families [2]. These conditions can have substantial emotional impacts on clinicians caring for children with neurologic disease [3, 4, 5], which may be heightened for residents early in their careers. For many residents, prognostication conversations are key learning experiences, and yet residents' experiences of these formative encounters are not well described. Describing the features of these experiences that shape residents' perspectives and approaches to prognostication may offer insights for resident education and highlight challenges and opportunities for strengthening prognostic communication.

Effective prognostic communication—providing clear information, taking time to answer questions, and conveying difficult news sensitively—is associated with improved patient care and caregiver outcomes, such as trust in clinical teams and decreased decisional regret or anxiety [6]. However, effective prognostication of neurologic outcomes is complicated by challenges such as variable disease trajectories, multidomain impacts (motor, cognitive, sensory), long‐term caregiving needs, and fragmented care systems [1]. For child neurologists, prognostic uncertainty is amplified by inherent variations in neurodevelopmental trajectory, limited evidence to support estimating prognosis [7, 8], complex pediatric ethical considerations [9], existential concerns like “out‐of‐order” loss (where a parent is caring for a sick child) [5], and a rapidly evolving therapeutic landscape that is fundamentally altering the natural history of individual diseases [10]. Addressing these factors requires individual competence with managing uncertainty and serious illness communication, and also depends on the broader healthcare and psychosocial context.

Neuroprognostication, due to its inherent uncertainty, can represent a “disorienting dilemma” associated with transformative learning experiences [11, 12, 13]. However, the challenges associated with neuroprognostication also pose ethical dilemmas for residents [14]. If inadequately addressed, such dilemmas could represent a threat of moral distress (where an individual is unable to act in accordance with their ethical beliefs because of institutional constraints), such as concerns about providing ongoing aggressive treatments when one does not believe that it is in the best interest of the patient or witnessing low quality of patient care due to poor team communication [15, 16]. If left unaddressed, repeated exposures to morally distressing events can lead to a “crescendo effect,” [15] contributing to compassion fatigue, burnout, and the intention to leave one's job [17]. This ultimately can lead to moral injury: an enduring psychological, spiritual, behavioral, social, and emotional harm to one's conscience as a result of experiences that conflict with their values or beliefs [18]. Studies of neurology [19] and pediatrics [20] residents using standardized items related to moral distress have identified that residents frequently report experiencing moral distress in caring for patients with neurologic illness. However, given the nature of the items on these scales, whether and how participation in prognostication may interact with resident moral distress is not clear.

We conducted qualitative interviews with senior (postgraduate year [PGY] 5–6) child neurology residents on their experiences with neuroprognostication. We previously reported results from these interviews focused on the educational experience of residents, including how they conceptualized and reported learning about neuroprognostication and the gaps and opportunities that they identified to inform future educational interventions [21]. Here, we focus on the contextual factors and the emotional and professional impacts that contribute to resident experiences of neuroprognostication. Specifically, we sought to understand (1) how residents describe experiencing the process of neuroprognostication, (2) what challenges they encounter in prognostication, and (3) how residents perceive participating in neuroprognostication impacts them professionally and emotionally.

2. Methods

2.1. Study Design

This exploratory qualitative study was conducted using a constructivist grounded theory approach [22] to construct an interpretive account of how senior child neurology residents experience neuroprognostication, including contextual factors and emotional sequelae. Aligned with a constructivist stance, we approached resident narratives as situated and shaped by clinical context, training role, and interpersonal dynamics, and treated resulting themes as an interpretive account formed through iterative, inductive analysis. We report methods in accordance with SRQR [23] and COREQ [24] guidelines. Additional details regarding the methods have been published separately [21].

2.2. Research Team and Reflexivity

Interviews were performed by a male, PGY‐5 child neurology resident with interest in neuropalliative care (G.L.L.). A peer interviewer was used to support rapport and facilitate candid discussion of potentially emotionally and professionally sensitive experiences. To reduce potential social desirability bias, participation was voluntary and confidentiality was emphasized. Additionally, a standardized, piloted semi‐structured interview guide was used across interviews. The interviewer had formal training in health professions education methods through a residency scholarly concentration and additional certificate training.

The research team included clinical research coordinators (B.D.W. and S.D.B.), pediatric palliative care physicians with expertise in qualitative methods and prognostication (J.A.M., D.B.M., and J.N.B.), a mixed‐methods researcher focused on pediatric serious illness (C.E.W.), and a child neurology attending with a background in bioethics (C.J.W.). The team met regularly to discuss emerging interpretations, interrogate assumptions, and refine the analytic frameworks guiding coding and thematic development.

2.3. Participants and Recruitment

Eligible participants were PGY‐5/6 child neurology residents enrolled in Accreditation Council for Graduate Medical Education (ACGME)–accredited programs. We focused on senior child neurology residents because their training stage afforded both repeated exposure to neuroprognostication and greater opportunity to reflect on and describe how they approach these conversations. We recruited a convenience sample of residents identified in a prior survey‐based study [25] via email in August 2024, followed by purposive snowball sampling to enhance variation by geographic region and program size through March 2025. Residents from the interviewer's home institution were excluded to avoid role conflicts. When the interviewer had prior acquaintance with a participant (e.g., through conferences), this relationship was disclosed during the consent process, and the voluntary nature of participation was reiterated. Participants confirmed review of the research information sheet prior to interviews, provided electronic consent, and received a nominal $50 cash incentive for participation.

2.4. Sample Size

We estimated a sample size guided by the information power framework [26] with a goal of achieving thematic saturation [27]. Given our focused study aim (child neurology resident experiences with prognostication), specific participant group (senior child neurology residents), and anticipated high‐quality interviews (piloted guide, trained peer clinician interviewer, and professional transcription), we estimated that 12–18 interviews would be sufficient. Analytic meetings began after the first two interviews to discuss emerging concepts and to inform subsequent interviews. A codebook was iteratively developed and formalized after 11 interviews (described below). Recruitment concluded after 15 interviews when new interviews added minimal novel content and themes were stable across cases, suggesting that thematic saturation had been achieved.

2.5. Data Collection

We developed a semi‐structured interview guide (Appendix S1) through iterative research team meetings, including designing questions to explore five key moral distress domains (complicity in wrongdoing, lack of voice, wrongdoing associated with professional values, levels of root causes, and repeated experiences) [16, 28]. The guide was piloted with four recent child neurology residency graduates (within 2 years of completing training), informing modifications, including additional example prompts and a specific working definition of neuroprognostication (“the process of utilizing available information to estimate the expected course of a neurologic disease or injury, likelihood of a cure, functional outcomes, and life expectancy”). The final guide included 11 questions, including one on the impact of participating in neuroprognostication conversations and 5 aligned with the moral distress domains.

The updated guide was used for audio‐recorded one‐on‐one Zoom interviews, which were professionally transcribed using a HIPAA‐compliant workflow (TranscribeMe!, San Francisco, CA). To help maintain rapport and interview flow, and because we utilized a single interviewer, field notes were not systematically collected as part of the study protocol. Following the interviews, participants were asked to complete a brief Qualtrics survey that included basic demographic questions. We did not conduct repeat interviews. Transcripts were not returned to participants.

2.6. Data Analysis

Demographic data were summarized descriptively. Transcripts were deidentified and managed in Dedoose (Sociocultural Research Consultants LLC, Los Angeles, CA) to support team‐based coding. Analysis proceeded concurrently with data collection using iterative coding and constant comparison, with analytic meetings beginning after the first two interviews. To develop a working codebook, three analysts (G.L.L., B.D.W., and S.D.B.) independently read the first 11 transcripts for familiarization and performed line‐by‐line coding on a subset (n = 4) to generate initial codes and candidate themes. The team met regularly to discuss themes, with an audit trail of coding decisions and category revisions.

To enhance rigor and reduce coder drift, the analysts independently reapplied the working codebook to a subset of transcripts (n = 6, including the initially coded plus two additional transcripts), discussed discrepancies, and clarified boundaries between codes. This represented 40% (6/15) of the final transcripts, consistent with guidance that at least 10% of transcripts are group‐coded to enhance analytic rigor [29, 30]. The first author (G.L.L.) coded the remaining transcripts (n = 9) with ongoing memoing and team discussions to discuss emerging themes and disconfirming cases and assess for conceptual sufficiency. A summary of themes and preliminary interpretations was shared to participants who indicated interest in providing anonymized feedback with standardized member checking questions [31]. Final themes were synthesized to characterize the contextual influences and emotional and professional impacts of participating in neuroprognostication for residents.

2.7. Standard Protocol Approvals, Registrations, and Patient Consents

The study protocol was reviewed and approved by the Iinstitutional review board at Stanford University (IRB protocol #75477). Informed consent was obtained from all participants. Deidentified transcripts were stored on secure servers with restricted access.

3. Results

3.1. Participant Demographics

Fifteen residents participated in interviews, including 7/19 (37%) of the residents invited from a previous survey‐based study [25], with 8 additional residents recruited via purposive snowballing. Interviews lasted a mean of 46 min (range 31–61). The mean age of the participants was 32.5 ± 2.7 years. Eleven (73%) participants identified as female. Participants represented a diversity of program sizes (average class size 4.4 ± 2.1, median 4, and range 2‐8) and geographic locations (12 programs across 10 states in all 4 US census regions).

3.2. Experience of Participating in Neuroprognostication

Qualitative analysis identified five themes describing residents' experience of participating in neuroprognostication: (1) serious neurologic pathology and decisions with high stakes, (2) navigating uncertainty and hope, (3) systemic barriers and challenges to prognostic communication, (4) negative emotional impacts, and (5) the importance of neuroprognostication. The major themes and sub‐themes are summarized in Table 1.

Table 1.

Overall themes and subthemes.

Overall themes	Sub‐themes
Serious neurologic pathology and decisions with high stakes	Influencing major medical decisions. Importance of creating shared mental models. Frequent negative connotations. Seeing the big picture.
Navigating uncertainty and hope	Difficulty of inexact science. Recognizing prognostication as an ongoing process with evolving long‐term outcomes. Practicing humility and listening. Wanting to keep hope alive. Making space for prognostic discordance.
Systemic barriers and challenges in prognostication	Lack of time. Lack of provider continuity. Concerns about inequities in care. Navigating discordance among teams. Working in a medical hierarchy.
Negative emotional impacts	General negative feelings. Frustration and distrust. Developing emotional exhaustion.
The importance of neuroprognostication	Gravity and heaviness. Privilege and responsibility. Distressing experiences in prognostication.

3.2.1. Theme 1: Serious Neurologic Pathology and Decisions With High Stakes

When asked to reflect on neuroprognostication experiences, residents frequently detailed situations involving serious neurologics pathology and decisions with high stakes (Table 2). For instance, many residents described viewing the purpose of neuroprognostication as influencing major medical decisions: “A lot of times, neurology is the team that the ICUs will lean on to be like, ‘Well, if they can recover from their brain injury, then it's worth pursuing everything else’” [Resident 8]. As another resident noted, this perception adds a significant sense of responsibility: “It's just really hard to be put in that position where a very big life‐changing, as in life‐ending or life‐sustaining, decision is made because families, I think, really take our word for what it is” [Resident 15].

Table 2.

Serious neurologic pathology and decisions with high stakes, additional quotes.

Sub‐theme	Representative quote
Influencing major medical decisions	I don't want to be the person telling you to not put this kid on ECMO because he had a stroke … I don't think … that it should be within the world of neurology to make this decision about that if that's what the family thinks and if they think unlike from the other [medical teams]. [Resident 5]
	One of the questions that the PICU really needed help with for us was helping mom to understand his prognosis so she could make the best decision about intubation. [Resident 11]
Importance of creating shared mental models	I feel so much relief once I find out the family's on a similar page or has the same idea of what's going on. [Resident 8]
	The intended outcome of [prognostication] is for the family to be equipped to understand a little bit about what's next in the future and oftentimes to make decisions. [Resident 7]
Frequent negative connotations	I haven't really thought about this before, but neuroprognostication is not inherently negative or positive. But I feel like when I'm asked [to give a prognosis] it often means “Tell the bad news.” [Resident 8]
	When they put in a consult for prognostication, what I interpret it as is they have high suspicion that there's a poor prognosis, and they want the neuro team to confirm their suspicions. [Resident 3]
Seeing the big picture	What they want to know is … will their child be able to have a life of dignity and comfort? [Resident 11]

This emphasis on major medical decisions underscored the importance of creating shared mental models. As one resident summarized, “The ideal effect of neuroprognostication is to provide clarity and to synthesize information in a way that's useful to families” [Resident 10]. Reaching a shared understanding shaped whether residents experienced the conversations as successful or distressing: “Feeling like you and the parents are leaving on the same wavelength of what the next day, week, month, year is going to look like is something that makes you feel better … when those disconnect … that is the hardest” [Resident 5].

Residents also highlighted that neuroprognostication has frequent negative connotations—where they perceive the request to provide prognostication as necessarily entailing delivering bad news. As one resident shared, “When I first heard neuroprognostication, my head goes to those very hard or terrible news or to very uncertain prognoses” [Resident 11]. Notably, at the end of an interview, another resident highlighted that despite their initial negative associations, they realized that neuroprognostication occurs in all facets of neurologic practice: “Reviewing through my experiences, I'm kind of amazed by how many times, big and small, I've had to do it, whether I realize it or not. Even giving an epilepsy diagnosis or an autism diagnosis, I realize there's some degree of neuroprognostication” [Resident 4].

Finally, a subset of residents extended their descriptions of the purpose of neuroprognostication to encompass seeing the big picture, including trying to address questions about how a particular injury or disease will affect a child's personhood. As one resident described, “There, the questions are, ‘How much of my child have I lost? Will they return to what they were before?’” [Resident 10]. These broad, fundamental questions further emphasize the high stakes associated with neuroprognostication, where residents felt tasked with connecting medical uncertainty to concrete implications for function, relationships, and quality of life.

3.2.2. Theme 2: Navigating Uncertainty and Hope

Our analysis identified several subthemes associated with navigating uncertainty and hope (Table 3). One common sentiment contributing to uncertainty is the difficulty of inexact science: “I wish we had beautiful data for prognosticating in peds, but we don't, right? Most of the time, we just don't have such concrete outcome data” [Resident 11]. Another resident emphasized that existing data can also be limited: “Those buckets are also quite broad. It's just like, ‘Oh, 80% likelihood that you have some degree of cerebral palsy.’ But we're not saying whether … they have difficulty with using their hand or they're completely hemiparetic” [Resident 9]. One resident directly linked uncertainty to less effective conversations: “The less certain I am, the poorer they go” [Resident 1]. Another described that “I think we have to hedge. I think it's just a necessary evil because we don't know” [Resident 3].

Table 3.

Navigating uncertainty and hope, additional quotes.

Sub‐theme	Representative quote
Difficulty of inexact science	We can't say for certain this is what kids are going to do because kids' brains can recover and do well and in ways that we don't expect. And that's what's truly remarkable about kids' brains. But it means they can't tell you exactly what things are going to look like. [Resident 5]
	I think what's really hard is when we get asked to do it for patients where we don't have a ton of data. I think there's a good amount of literature in our cardiac arrest patients. But being asked to prognosticate for other things, I think that's where it gets a little bit more tricky. [Resident 15]
Recognizing prognostication as an ongoing process with evolving long‐term outcomes	A lot of times in the first conversation, you only get through the headline of what you're going to share … giving people time to process that and having another conversation once they're ready to hear more details or answer their questions because people can't think of questions in that acute moment a lot of the time and making that conversation just as formal and important as the first one. [Resident 7]
	Three years down the line, the way that you interact with your child just might look different, and you're going to find meaning in how you interact with that child, even if that looks different from what you thought it would look like three years ago and X number of years ago. [Resident 14]
	Sometimes it just takes a few visits. Sometimes it takes backing off. I think really having to read the family and kind of where they are in terms of that kind of grief. [Resident 2]
Practicing humility and listening	We need to listen more. We need to listen 10 times more than we speak … if we just listened to what the family's questions were about the future, then we could probably answer things in their terms and in their words. [Resident 8]
	You have to be putting them first. You're sharing medical information, but you're sharing devastating information to a human being about their family member. It's going to feel awkward for me … it feels devastating for them. Awkward is a very minimal feeling compared to devastated. [Resident 12]
Wanting to keep hope alive	Naming the uncertainty is really helpful … I acknowledge the uncertainty and acknowledge that there is some hope, that these things could happen. And so not losing that hope is usually something I try to address with families. [Resident 7]
	No matter what you say to a family, you're always going to leave room for the child being able to surpass that because there's always the anomaly … it's very easy to hedge with giving neuroprognostic information because families are hopeful. [Resident 10]
Making space for prognostic discordance	I think that you can give them all the information, and it may not be for another five years that they're really able to understand that.… You have to just accept that they're at where they're at, and they're the ones that have to really experience all of it. [Resident 10]

Several residents identified that an element of managing uncertainty is recognizing prognostication as an ongoing process with evolving long‐term outcomes. This included naming the importance of being thoughtful updaters: “Neuroprognostication is not a one‐time event. It's something that evolves over time. And so our job is to deliver the information, answer those questions, but also to follow along and revise that kind of conversation as we go along” [Resident 14]. This also reflected recognition of limitations in traditional functional outcomes: “You could read about a kid who is wheelchair‐bound … is G‐tube fed and has a tracheostomy. And that could sound like a really bad outcome. But then if you meet the child and they're actually a really happy kid and the family has really positive interactions with them, and they're communicating with them in their way, for them, that could be a very positive outcome” [Resident 7].

Another sub‐theme was the importance of practicing humility and listening while engaged in prognostication. One resident reflected that “I would have wished that I had infused more of that humility of uncertainty earlier because I think it's helpful. Again, I think it's being honest and transparent—I think is something that elicits more trust, and it can give you more credibility” [Resident 9]. Another resident reflected on considering how they respond to family disagreements about prognosis: “If [the family] aren't in agreement with the prognosis, it's usually like, to what end? … I try to understand kind of where that resistance is coming from and then meet them where they're at” [Resident 2].

Residents also described wanting to keep hope alive. For instance, one resident recounted a mother's description of a prior neurologist—whom she labeled “Dr. Death” for conveying what the family perceived as an overly pessimistic prognosis—and reflected on how that characterization still shapes their own approach to balancing realism with hope: “I have her voice in my head still of understanding of like, from my perspective, everyone wants to be hopeful … how do you provide hope and balance that with the reality of, we see these kids when they're in the hospital and not doing anything and trach vented in 5 years … how do you paint a picture of that? Because that, I think, is an important part of our job, but also there are kids, [a] couple [of] percent that look fine” [Resident 5].

Finally, residents shared the ways in which they focus on making space for prognostic discordance between clinicians and families. One resident described a situation where “[the patient's mother] was like, “I don't really care what your data says. My kid's making all of these steps in the right direction, and so I'm not going to give up on her.” And so I was like, “That's totally your prerogative to do” [Resident 15]. Another resident noted that disagreements about prognosis do not always reflect a lack of knowledge: “[Families] know. They don't share the belief, but it's not like they don't understand. And so my job isn't to crush somebody's—to change their belief system and their faith and their hope” [Resident 11].

3.2.3. Theme 3: Systemic Barriers and Challenges in Prognostication

Residents described multiple systemic challenges that had negative impacts on their experiences with prognostication (Table 4). One resident named challenges associated with a lack of time: “I often feel like the systems are failing me, and I wish I could have another follow‐up appointment to talk more with them [the families] and answer questions that are coming up and ensure their understanding” [Resident 8]. Another example included a lack of provider continuity: “Something out of our control is really turnover of teams. I don't know how to solve that one, but in terms of building trust, it's really hard to see the faces switch every week” [Resident 4]. Residents also described concerns about inequities in care based on psychosocial or demographic factors such as language barriers: “The ‘pleasant’ families, the ones that are personable, perhaps also English speaking, all of those factors make things go more smoothly … sometimes we give better care to the ‘nice’ families because it's easier” [Resident 10].

Table 4.

Systemic barriers and challenges to prognostication, additional quotes.

Sub‐theme	Representative quotes
Lack of time	I wish every conversation we could do in a conference room or every conversation we could have a very controlled situation and a structured time and not have to leave because the pager is going off or things like that. [Resident 12]
Lack of provider continuity	Having continuity is probably the best because I think you need time to build that trust, and you can't really build it after a week of just appearing. [Resident 4]
	It's hard with the lack of continuity … if I had not gone off service, maybe that conversation would have been able to be had earlier in our shared course. [Resident 9]
	It's very difficult talking with families when you say … that's just their brain's current background. But today, it's a Tuesday [with a new attending], so now we're going to aggressively treat it like status [epilepticus]. [Resident 12]
Concerns about inequities in care	I think it's definitely challenging if there is a sense that I get that a family is—or health literacy might be more challenging. Those are the ones that are harder, I think, sometimes because they might not understand exactly what we're trying to say. [Resident 15]
	I wish I could say that the prognostication discussions go equally as well when there are differences in culture or language barriers.… But I just truly feel like I'm not as effective when I'm working with those limitations. [Resident 8]
Navigating discordance among teams	If they're going to put their ideas forward anyways, then it feels like, “Why are they including yours if they're not going to respect those thoughts?” And so it can be frustrating. [Resident 1]
	“Some people were saying like, ‘Oh, gosh, comfort.’ And then others were pushing for more, ‘Let's try a more aggressive approach to our treatments …’ trying to navigate that could be a little bit whiplash feeling.” [Resident 11]
	We get pretty frustrated and when the messaging is a little bit discrepant … I think the frustration can come in like, “If you knew this, why do you need us?” sort of thing or, “If you were going to say this all along, why do you need us to back you up?” or, “Why did you say this to the family when you knew you were going to call us and we might not agree?” sort of thing. [Resident 15]
Working in a medical hierarchy	Sometimes I have to be the one that will teach some of the attendings a more family‐centric way of having these discussions. And that can feel a little challenging because there's this inherent hierarchy. [Resident 8]
	That was another one where it feels bad, where if it feels like the attending had to step in to say something, then you automatically feel like—you are like, “Oh, shoot. Did I say something wrong?” [Resident 15]

Other challenges included inter‐ and intra‐team dynamics. For instance, a common difficulty was navigating discordance among teams. One resident explained that “it can be frustrating when they want us to support this narrative that they have that we might not agree with … do we have to backtrack something that was already said? And then does that create more distrust?” [Resident 15]. Similarly, residents named challenges working in a medical hierarchy, especially in supporting trust and autonomy: “Our goal as [residents] is to be the main person … when you're telling them something on Sunday and the very next day you're telling them something different because your attending has a different idea … it makes you feel like you are less knowledgeable and less accurate to families” [Resident 1].

3.2.4. Theme 4: Negative Emotional Impacts

Residents consistently described prognostication as an emotionally salient experience (Table 5). Several described general negative feelings: “I've never been asked like, ‘What does it feel like?’ I can use a bad word. But yeah, I think it feels bad either because you weren't clear or it feels bad because you just ruined this family's hopes and dreams” [Resident 15]. Another resident described how the association of neuroprognostication with bad news limits their ability to share their experiences: “I can tell almost nobody about what I do in my job, which is a lot of neuroprognostication because you even say the first sentence and people are upset and you're like, ‘Oh, we haven't even started yet’” [Resident 7].

Table 5.

Negative emotional impacts, additional quotes.

Sub‐theme	Representative quote
General negative emotions	It feels not good of course, because you sit there and feel like you're the bad guy in the situation, and you're the one taking something away from them. [Resident 12]
	I still feel inadequate. I think it's still hard for me if someone puts me on the spot and asks me for level of certainty based on what I've seen. I always feel like I need to apologize for myself for all the things that I don't know. [Resident 10]
	I hate not being helpful … it doesn't seem like an unreasonable question … to not be able to give that answer definitively is really tough. [Resident 11]
Frustration and distrust	Frustration: We do all this training … We are all working with the understanding that we're providing the best care to our patients. And then to have that doubted by a family member, I think, inherently creates this tension. [Resident 8]
	Frustration: I do feel a sense of frustration, I guess, if that's the word, that we're giving them our evidence, and this is what we could expect. But I understand they're in a stressful situation and need time to cope. But it does make me feel uneasy when they're making decisions based on that false hope rather than the science. [Resident 3]
	Regret: I think those conversations can go as well as they possibly can, but I still think you're going to be left feeling like you didn't maybe do everything. [Resident 10]
	Fear of consequences: It's those ones where it's like, you make the wrong call, they withdrew care, but they could have had this very fulfilling life. Those are the ones that I think are always the hard ones and the ones that you are always like, “Was that the patient that was going to actually do really well or—” [Resident 15]
Emotional exhaustion	I think sometimes you know in advance, this case is going to be horrible. And so I'll build up a wall before I've even seen the child or the family for the first time. [Resident 7]
	I feel much less emotional than I did. I remember the first family meeting, I was like, well, one, nervous about things, but also really upset because it was a poor prognosis. But it feels much more routine now than it did. [Resident 14]

Several negative emotions like frustration and distrust arose in discussions. Residents named specific episodes of frustration (“It can be frustrating in the moment that—it might feel like families don't understand or they can't see” [Resident 10]), regret (“It took quite a few conversations … to really say the word 'encephalopathy.' Reflecting back, I wish someone in the team would have done it a little bit sooner because exactly what had happened was distrust” [Resident 4]), and a fear of consequences (“You learn that despite what your goal is to give answers, oftentimes hedging can be safer than not … you don't want to be the physician that tells the parent that [bad news] unless you know for 100% certainty” [Resident 1]).

Finally, residents described developing emotional exhaustion when engaging in prognostication: “I feel like sometimes … there ends up being almost an emotional dissociation or like emotional distancing … over time, it can remove the emotionality or humanness of being a provider” [Resident 13]. One resident attributed this to having participated in many prognostication conversations: “Sometimes it can just seem like we're having these discussions again and again. It becomes kind of like a routine when it shouldn't be a routine … I don't want to say you lose emotion about it, but I don't know” [Resident 6].

3.2.5. Theme 5: The Importance of Neuroprognostication

Despite the challenges that they described, residents also explained how important participating in neuroprognostication is (Table 6). Many residents described that there is a gravity and heaviness to prognostication: “There's just this gravity to certain things in life, in these existential crises that people face … you can't have those conversations routinely and not think about just the meaning of life and all of that too because I think that all gets pulled up” [Resident 10]. This contributed to residents explaining that they felt that being asked to prognosticate is a privilege and responsibility, even when the prognosis is guarded: “It's a terrible feeling to give that horrible news. But I do think that it is an act of love to our patients to do that because ultimately, families want that information, they want to know that it helps them to feel some power and some control of the situation” [Resident 11]. As another resident put it, “They are often the most difficult conversations we have. And I think the ones that sit with me the hardest, of coming to a diagnosis, can provide families relief” [Resident 5].

Table 6.

Importance of neuroprognostication, additional quotes.

Sub‐theme	Representative quote
Gravity and heaviness	This is the only time that the parent is going to hear that first time of like, “Okay. Your kid has a brain injury. There's going to be a disability.” And if you kind of say the wrong things, it can often sort of echo in their ears for a long time. [Resident 8]
	It's hard to go from one patient room with a child who is neurologically devastated from some injury, and then your next clinic's patient is a kid with recurrent febrile seizures. For both of those families, how they respond to that situation is the same. I think that's a challenge inevitably for the physician, right? [Resident 9]
Privilege and responsibility	It's a privilege that we get to be the people to share these types of information and give prognosis [Resident 12].
	I really value these conversations because I think it really helps families understand what's going on with their kids … it's sad a lot of the time, but it can bring them a lot of peace because it often brings them a lot of answers that they need to make decisions. [Resident 7]
	I think in a strange way, sometimes I feel … almost honored that someone is trusting us to have these conversations and to really help a family navigate through a very difficult situation. [Resident 4]
Distressing experiences in prognostication	I feel a little bit stuck because I want to be able to meet the family in terms of where they're at and what kind of care that they want, but I'm not able to do it in the terms that they want … that can affect me in a personal way, too, because I want to be there for those families in a way that I can't be. [Resident 2]
	We want to be validating. But when you're put in a position that feels very unnatural where you can't be validating and you can't be that person that provides a lifeboat, I think that that is really difficult. [Resident 10]

Residents' sense of responsibility stood in tension with examples of distressing experiences in prognostication. One resident shared an experience where they felt that a prognostic conversation was ineffective: “That felt like a failure … one, you feel bad because you alter people's understanding about the health care field and create sort of a basis for distress … but then you feel like you failed the child, that they're not getting the care they need” [Resident 4]. Another resident described distress when asked to prognosticate in a non‐ideal setting: “I tried to bring that up at one point and say, like, ‘Can we at least have some consideration before we ruin someone's life today about how we do it?’ And they're like, ‘Yeah, that sounds lovely. We don't have time for that.’ I was like, ‘Oh, no’” [Resident 11]. Notably, these types of distressing experiences were specifically described in only 6/15 interviews (40%).

4. Discussion

Our analysis provides a narrative of how child neurology residents experience neuroprognostication. Residents describe neuroprognostication as a critical task often involving serious neurologic disease and high‐stakes decision‐making. They highlight that skillful prognostic communication requires balancing uncertainty and hope. They also identify systemic factors that can erode their ability to communicate effectively: limited time, poor continuity, health inequities, and discordant messages among health care teams. Residents describe these factors as contributing to negative emotions and, over time, emotional exhaustion. Still, they describe feeling privileged to contribute to what can be an existential moment in patients' and families' lives. This sense of responsibility is juxtaposed with examples of distressing experiences when they felt unable to fulfill this role. Overall, this narrative description highlights influences at multiple levels, from individual‐level concerns to specialty‐wide considerations, that impact how residents experience neuroprognostication. These factors in turn represent opportunities to support residents engaged in neuroprognostication and promote effective prognostic communication.

At an individual level, residents identified navigating uncertainty and hope as a central component of their neuroprognostication experience. As such, strategies targeted at managing prognostic uncertainty are critical [1]. For instance, specific educational strategies have been proposed for narrating, framing, and simulating uncertainty in medicine [32]. A recent study in pediatric oncology that included patients, caregivers, and oncologists identified four key recommendations to help manage hope and uncertainty in prognostic discussions (prioritize honesty; provide leadership, guidance, and expertise; provide compassionate, individualized support; and acknowledge inherent uncertainty as a space for hope) [33]. A framework specific to neurology provides approaches for both formulating a prediction (anticipating, anchoring, tailoring, and de‐biasing) and communicating “what to expect” (engaging, informing, responding, and framing) [34]. This can also include developing “medical forecasting” skills as a more formal, structured, and data‐driven approach to estimating prognoses [35]; importantly, forecasting also includes frequent updating with recognizing prognostication as an ongoing process. Individual‐level supports can also include developing understanding and skills to make space for prognostic discordance between patients/families and clinicians (e.g., recognizing differences between discordant beliefs and understanding [36] and responding to expressed hope for a miracle [37, 38]). The fact that residents are trainees emphasizes the importance of effective oversight by supervisors that supports them in a manner limiting emotional harm; we previously described how residents identify effective modeling and graded autonomy as key educational contributors to learning to prognosticate [21]. Consolidating these frameworks and strategies represents a key component of future curricular development in neuroprognostication.

At a team and institutional level, residents identified multiple systemic barriers and challenges in prognostication. The idea that skillful prognostic communication is a clinical intervention that changes patient care outcomes [39] emphasizes the need to protect the time, space, and other resources necessary for these conversations. Qualitative interviews with parents of infants with neurologica conditions about their experience with family conferences named “prioritize attendance of consistent and supportive team members” and “take responsibility for calling regular meetings” as opportunities for clinical teams to support continuity and improve the patient/family experience [40]. The experience of discordance among teams underlines the need to prioritize a structured approach to family meetings [41], including preparation that considers appropriate teams and venue and ensures team consensus as part of the pre‐meeting steps [42]. These pre‐meetings can also serve as a chance to delineate roles among trainees and attending physicians to reduce the potential impact of working in a medical hierarchy. Continuing the development, implementation, and assessment of family‐informed approaches to neuroprognostic communication (e.g., the ALIGN framework [42] or BRIDGE tool [43]) represents strategies that may support creating shared mental models—which residents identified as one of the central purposes of neuroprognostication. Finally, dedicated debriefing models have been developed for teams to use to reduce distress after difficult experiences (e.g., “4As”: Ask‐Affirm‐Assess‐Act [44], “3D”: Defusing, Discovering, and Deepening [45]). Debriefing can also be implemented at an institutional level to address concerns that can contribute to moral distress (e.g., Schwartz Rounds [46, 47]).

At a specialty‐wide level, a critical need is addressing limitations in existing neuroprognostic outcome data [2, 7, 8] that contribute to the difficulty of inexact science. For instance, clinicians frequently predict worse functional outcomes than actual outcomes [48], a pattern more pronounced for physicians early in their training [49]. One interpretation of the theme that residents associate neuroprognostication with frequent negative connotations is that this perspective reflects a pessimistic prognostic tendency. Such pessimism may contribute to a “self‐fulfilling prophecy,” whereby treatment decisions guided by prognostic estimates shape outcomes that then populate the evidence base used to generate future prognoses [50]. Moreover, many common neuroprognostic outcome measures have been criticized as not correlating well with long‐term difficulties, creating artificial dichotomies, and not reflecting quality of life [51]. For instance, a study with families of children born prematurely found that only 12% of parents with children clinically classified as having severe neurodevelopmental impairment agreed with this classification [52]. Identifying and describing outcomes that reflect information important to families are important steps in improving prognostic outcome data [51, 53, 54]. Finally, residents emphasized that neuroprognostication has evolving long‐term outcomes; reframing prognostication as longitudinal anticipatory guidance across inpatient and outpatient settings may help mitigate negative connotations and better align with families' evolving informational needs.

Notably, while negative emotions related to prognostication were prevalent across participants, distressing experiences suggestive of moral distress arose in only a subset of interviews. This may reflect limitations of the current study, including that the initial hypothesis overestimated or misattributed the risk that neuroprognostic conversations generate moral distress. That is, prognostic conversations may be emotionally salient without consistently meeting key features of moral distress (i.e., perceived constraint from acting in accordance with one's values). Future work explicitly eliciting morally distressing experiences—whether involving prognostication or not—may better characterize threats of moral distress in child neurology training. Alternatively, our focus on senior residents may have selected participants with greater clinical confidence, communication skills, and tolerance for uncertainty; longitudinal or comparative studies involving junior trainees could help characterize the development of these skills over training. Another possibility is that, while working among clinical teams and within medical hierarchies presents unique challenges, these contextual factors also serve as buffers against moral distress by placing ultimate responsibility for prognostic framing with attending physicians, warranting comparative study of resident versus attending prognostic experiences. Finally, other methodological limitations include social desirability bias potentially leading to reluctance to disclose morally distressing events, the need for additional interviews despite reaching thematic saturation, or interview guide questions that may have de‐emphasized values conflicts over emotional experiences. Nevertheless, these experiences emerged in 40% of interviews, underscoring the need to address prognostication‐related challenges to support resident well‐being and career longevity.

In conclusion, the emotional salience and sense of responsibility that child neurology residents express when describing participating in neuroprognostication underscore the importance of these encounters as transformative learning experiences. Our narrative account of these experiences highlights the individual‐, team‐, institution‐, and specialty‐level challenges faced by residents. Future work, including developing and studying educational interventions to support resilience and neuroprognostication skills, implementation of family‐informed approaches to neuroprognostic communication, and addressing limitations in existing prognostic outcome data, represents opportunities to not only support residents in training but also promote effective prognostic communication that can impact patient care.

Author Contributions

Grant L. Lin: conceptualization, investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, formal analysis, project administration, data curation. Beatrice D. White: investigation, formal analysis, writing – review and editing. Shelby D. Burk: investigation, formal analysis, writing – review and editing. Jessica A. Moriarty: conceptualization, writing – review and editing, supervision, methodology. David B. Mauser: conceptualization, writing – review and editing, supervision, methodology. Claire E. Wakefield: conceptualization, writing – review and editing, validation, formal analysis, supervision, resources, investigation, methodology. Courtney J. Wusthoff: conceptualization, software, supervision, writing – review and editing, methodology, resources. Justin N. Baker: conceptualization, investigation, funding acquisition, writing – review and editing, validation, methodology, formal analysis, software, supervision, resources.

Ethics Statement

The study protocol was reviewed and approved by the institutional review board at Stanford University (IRB protocol #75477). All participants provided electronic consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File

Data Availability Statement

Anonymized data not published within this article will be made available by request from any qualified investigator. The data that support the findings of this study are available from the corresponding author upon reasonable request.