The term “periviability” refers to the earliest stage of fetal maturity where there is a possibility of extrauterine survival. While consensus exists regarding the gestational age at which a neonate may be characterized as “viable,” there is ongoing debate about whether care at the lowest limits should be pursued given the significant risks (including death and severe disability) to those who receive resuscitation. As a result, although perinatal care practices are relatively consistent after 25 weeks' gestation, there is significant intercenter variation in management within the 22- to 24-week “gray zone” period—beyond what can be attributed to parental preference alone.[1] At this point, “there may be several legally and morally permissible goals of care,”[2] resulting in discrepant interventions between units around the world.[3]
[4]
[5]
[6] Recent data collected from multiple U.S. academic medical centers estimated 1-year survival rates of 30 and 56% for infants actively treated and born at 22 and 23 weeks, respectively,[7] although the University of Iowa published a study in 2020 demonstrating a 78% survival rate to discharge for infants born between 22 and 23 weeks' gestational age.[8] In Sweden, 22-week infants are routinely resuscitated,[9] and centers in Japan and Germany have also reported significantly higher survival rates for 22-week infants than most U.S. institutions.[10]
[11]
Studies examining such intercenter variation in neonatal intensive care unit (NICU) outcomes have hypothesized that “culture”—defined here as “the beliefs, norms, attitudes, and assumptions that are learned over time, shared by [an] organization's members, and operate usually at a subconscious level”[12]—can be one of many factors that impact patient care.[13]
[14]
[15] The exact relationship between unit culture, practice, and outcomes for periviable infants remains unclear. While one multicenter retrospective study was not able to describe an association between a culture of optimism and increased neonatal resuscitation following periviable delivery,[16] other studies have shown that pessimism, negative language, and value judgements about periviability from the medical team can adversely affect parental autonomy,[17] communication with families,[18] and clinical decision-making.[19]
[20]
[21]
Given these phenomena, the idea of institutional culture is worth exploring—how teams work together to care for periviable patients when families opt for resuscitation, what team members believe about the work they do, and the overarching moral consciousness of those within the institution regarding that work. Through examining the different components of culture, it may be possible to identify modifiable factors that can and should be targeted to improve clinical practices and ensure patient outcomes are congruent with their parents' goals.
Materials and Methods
Sample and Recruitment
Qualitative, semi-structured interviews were conducted at a large academic Midwestern hospital. All interviews were completed by the study's first author, S.Y.A. Recognizing the importance of interdisciplinary collaboration in the NICU, members of all five clinical groups caring for patients in our unit were initially approached and asked to participate: attending and fellow neonatal–perinatal medicine physicians, neonatal nurses, advanced practiced neonatal nurses, pediatric hospitalist physicians, and neonatal respiratory therapists. Shortly thereafter, our team realized the value of understanding the breadth of perinatal care culture and expanded to include obstetric physicians at our institution, specifically attending and fellow maternal–fetal medicine (MFM) physicians. The inclusion of these six clinical groups is reflective of ethical cultures research that has been previously published.[16]
Participants were recruited through flyers posted around the unit as well as emails sent from the research team to all eligible staff. All those who responded were invited to participate in the study with the goal of targeting at least five clinicians per role. This study was declared exempt under federal regulations by the Institutional Review Board of the Northwestern University, as was the amendment to include MFM providers.
Interview Protocol
All interviews began with a standard question: “Are you familiar with our institution's guidelines regarding perinatal counseling and resuscitation based on gestational age? If so, please describe them.” This was done to establish a shared awareness of current hospital guidelines for all interviewees. Participants were then asked a series of open-ended questions, which can be found in [Supplementary Material S1] (available in the online version). The interviewer expanded on specific topics based on interviewee responses.
Interviews ranged from 30 to 60 minutes in duration. All interviews were recorded, and participants signed consent forms prior to the discussion. A $10.00 Starbucks gift card was offered to each participant at the end of the session in appreciation for their participation. The purchase of gift cards was supported by divisional funds from the Lurie Children's Division of Pediatric Palliative Care.
Qualitative Data Analysis
All audio recordings were transcribed using Sonix automated translation software.[22] Digitally transcribed manuscripts were then reviewed by members of the research team for accuracy prior to analysis.
Inductive analysis was completed using the Dedoose software.[23] In phase 1, initial codes were generated by all members of the research team (S.Y.A., J.T.F., and N.H.) using study aims and a line-by-line review of two transcripts from each participant category (i.e., neonatal fellows, neonatal nurses, etc.). Coders were blind to each other's coding during this stage. Then, all team members (S.Y.A., JTF, and NH) met to discuss individual findings and develop a codebook. All differences were evaluated until either unanimity or a negotiated consensus was reached. Next, two members of the research team (S.Y.A. and JTF) worked independently and iteratively to code the remaining transcripts based on the newly established codebook, making changes when necessary based on emerging data. The codebook was revised by all team members (S.Y.A., J.T.F., and N.H.) over the course of several meetings, and all transcripts were revisited by each coder to ensure consistency and qualitative rigor. Code saturation was achieved with no new codes identified from the final transcripts. Each transcript was read 6 to 8 times on average by a team member (S.Y.A., J.T.F., or N.H.) during the analysis process. Once coding was complete, all team members (S.Y.A., J.T.F., and N.H.) met to review the final product and confirm intercoder reliability. Dedoose was used to further analyze the results and identify meaningful themes and correlations within the data.
Results
In total, 22 interviews were conducted. Participants included 3 MFM physicians, 5 attending neonatologists, 5 neonatal–perinatal fellows, 3 neonatal bedside nurses, 3 neonatal advanced practice nurses, and 3 pediatric NICU hospitalists. To ensure participant anonymity, no further demographics were collected. Only two individuals who responded to the recruitment inquiry did not go on to participate in the study—one advanced practice neonatal nurse and one neonatal nurse. Respiratory therapists declined the invitation to participate and did not provide a reason for this decision.
Across all transcripts, 2,805 code applications were completed utilizing 35 parent or subcodes ([Table 1]). Qualitative analysis of all codes revealed three central themes described by participants as contributory to institutional culture: Perception, Statements of Information, and Dynamic Factors. Within each of these themes, several subthemes were identified. [Fig. 1] depicts these findings as a conceptual model.
Table 1
Code application frequency by theme
|
Central theme
|
Subtheme
|
Sub-subtheme
|
Sub-sub-subtheme
|
Code applications (n)
|
|
I. Perception
|
432
|
|
I1. Attitudes and personal ethics
|
115
|
|
I2. Experience of moral distress and/or discomfort
|
280
|
|
|
I2a. Sense of arbitrariness
|
30
|
|
|
I2b. Desire for/lack of continuity
|
32
|
|
|
I2c. Impact of individual clinical experience
|
82
|
|
|
I2d. Influence of family goals and wishes
|
84
|
|
|
I2e. Institutional practice differences
|
52
|
|
|
I2f. Perceived lack of iteration/improvement
|
14
|
|
|
I2 g. Perception of infant discomfort
|
40
|
|
|
I2h. Prognostic and/or clinical uncertainty
|
25
|
|
I3. Perceptions of families regarding the child's clinical course
|
64
|
|
I4. Quality of life (QOL)
|
85
|
|
|
I4a. Participant defined “bad” QOL
|
30
|
|
|
I4b. Participant defined “good” QOL
|
34
|
|
|
I4c. QOL as defined by parents
|
38
|
|
II. Statements of information
|
225
|
|
II1. Academic curiosity
|
62
|
|
|
II1a. Desire for training and education
|
21
|
|
II2. Comparison to other diagnoses
|
16
|
|
II3. Institutional memory of patient outcomes
|
69
|
|
II4. General outcomes data
|
46
|
|
II5. Presence or absence of hospital protocols
|
73
|
|
|
II5a. Outside-hospital protocols
|
15
|
|
|
II5b. Prentice protocols
|
53
|
|
III. Dynamic factors
|
237
|
|
III1. Team dynamics
|
195
|
|
|
III1a. Influence of team role
|
74
|
|
|
III1b. Interdisciplinary collaboration
|
134
|
|
|
|
III1b1. Differences in communication styles between clinical disciplines
|
13
|
|
|
|
III1b2. Prioritization of maternal versus infant well-being based on clinical team role
|
8
|
|
|
|
IIII1b3. Role of palliative care team
|
4
|
|
|
III1c. Staff discussions surrounding periviability
|
34
|
|
III2. Time
|
55
|
|
|
III2a. Adjusting to change
|
34
|
Fig. 1 Qualitative analysis of all codes revealed three central themes described by participants as contributory to institutional culture: Perception, Statements of Information, and Dynamic Factors. Within each of these themes, several subthemes were identified. This figure depicts these findings as a conceptual model.
Perception
This theme consisted of cultural elements of an intuitive nature, based primarily on individual attitudes and insights. Subthemes include staff feelings of moral distress when caring for periviable neonates, the presumed experiences of NICU families, quality-of-life considerations, and personal ethics. Quality-of-life considerations were subdivided into “participant-defined” (i.e., “good” vs. “bad”) and “parent-defined.” “Parent-defined” refers to the patient's quality of life as determined by the ethics and beliefs of their parents, then subjectively interpreted and reported by the study participant in their interview. Staff feelings of moral distress were subdivided based on factors that participants described as contributory to feelings of anguish or constraint when providing care believed to be unethical or not in the best interest of the patient.[24] These included prognostic uncertainty, lack of clinical continuity, differences in institutional practices, a perceived lack of iteration or improvement in clinical practice over time, perceptions of infant discomfort, a sense of arbitrariness in patient outcomes or clinical decision-making, the influence of family goals and wishes, and the effect of experience on one's ability to cope and navigate through challenging situations.
Overall, participants were more likely to mention factors in the Perception theme (code application of n = 430) than factors in the Information category (code application of n = 225). Within Perception, staff discussion of moral distress comprised the largest subgroup (code application of n = 280), followed by personal ethics (code application of n = 115), influence of family wishes (code application of n = 84), and the effect of participant experience (code application of n = 82). Although moral distress composed the largest subgroup and was mentioned at least once in each participant interview, there was a wide range in frequency across interviews with a high of 27 code applications and a nadir of 4 code applications. References to moral distress both confirmed and refuted its significance, with one participant stating:
“I think where it gets hard is when you see the baby on all of these drips and you see the baby…so bloated from all of the fluids and you see their head so swollen and they're just fighting to live. And that fight goes on for days and days and days. I think that's where it gets really hard, because it's hard not to think, why? Why does this baby have to go through that?”
While another stated:
“I don't view it as moral distress. I think the older you get, the better you are able to bundle those feelings into a little side corner…if a family wants me to go hard, I'm going to go as hard for 22 weeker as a 26 weeker. That's just super easy…It's always sad when a lot of the babies die, right…but I don't think it impacts decision-making…I see the numbers, I talk about them every week. But it's not distress. I don't quite know how to quite give it the right nomenclature, but it's not distressful to me.”
Participants also described differences in moral distress based on team role, with a perceived increased risk of moral injury for neonatal nurses when compared with other clinicians, and a perceived decrease for clinicians with more seniority or decision-making power. Representative quotes include:
“I would probably say as an attending, I think I experience less moral distress…maybe because I kind of feel like I'm ultimately the decision-maker and the one who has to guide where we're going and the one who gets to or has to draw those lines in the sand…for example, [I may say] we're not going to do just compressions on this baby who's hypoxemic and that's why their heart will stop, whereas…somebody else in the care team may want to draw that line sooner. But until I make that decision it's kind of out of their hands.”
“It can feel more prevalent for the bedside nurse because typically the bedside nurse is there more often…so it feels more tangible, it feels more prevalent, it is what you are living and breathing for 12 and a half hours if you get out on time…You know, and consistency of care is important, so it's probably what you're living and breathing for 12 and a half hours for all of your shifts that week. So I think it can feel a little more present and a little harder to step away from I think as the bedside nurse.”
Knowledge of family goals was noted to both ameliorate and exacerbate moral distress, with some participants describing feelings of acceptance and others describing feelings of worry.
“I try to remind myself, that's not my decision to make…my definition of a good outcome may not be their definition of a good outcome. And that's ultimately what I'm there to support, the family's decisions, right? So, it's not on me.”
“Definitely if we are reaching the limits of what we can do, sometimes I have moral distress about continuing to support them. Usually, I feel like we can kind of align the medical team with the family and get to a point where we say, okay, the baby is suffering. We can kind of redirect. But I had one experience with 22-week twins that were born here a year and a half ago where the parents really wanted everything done. And we were just kind of hitting the point of futility with our access issues and how sick the baby was. And that was hard to watch and hard to be a part of.”
Statements of Information
This theme consisted of factors anchored in empirical elements such as existing data, specific patient cases, and institutional policy. Subthemes include pre- and antenatal hospital protocols for medical management at the limits of viability, mentions of outcomes data both locally and at other centers, the comparison between mortality and morbidity resulting from periviability versus other diagnoses (e.g., congenital anomalies), institutional memory of individual patient outcomes, and academic curiosity, defined here as the desire to learn from emerging scientific evidence. It is worth mentioning that while this theme is meant to represent more objective topics, our study relies on qualitative data and first-hand accounts. We therefore recognize that there is an inherent degree of subjectivity to the information presented by participants as “data.” The categorization of this theme as “Statements of Information” is meant to acknowledge that the authors are ultimately unable to differentiate these subjective interpretations of fact from simple errors or lack of knowledge.
Within the Statements of Information theme, institutional memory of patient cases (code application of n = 69) was mentioned in greater frequency than overall survival and outcomes data (code application of n = 46), mirroring the relationship between Perception and Statements of Information, with memory representing the more subjective and personal and survival data representing the more objective and impartial. Institutional memory was described as having a significant impact on culture. Representative quotes include:
“I think that it does affect the morale…if we see a 22 weeker on the board, we're like, Oh, no, I hope that they don't want a full resuscitation because everyone here has seen the ones that we've had recently, knows how sick they were, and the outcome…[it] definitely influences how people think about it here.”
“We've had really kind of sad outcomes for 22 weekers. And a lot of them have died in ways that I think the nurses and even the docs found uncomfortable to witness. There were 22 week twins who one of them died early in horrible respiratory distress. But the other died basically of lack of access…no one could get a PIV on him and so he was like 500 grams overloaded and just looked…it was like hard to look at him. I think the nurses found that very distressing. And we've just had a lot of kids die of overwhelming sepsis, but not in a quick way.”
Population-based survival data—specifically knowledge of more favorable clinical outcomes at other institutions when compared with the study institution—was discussed as both a source of moral distress and a catalyst for institutional change.
“We're several years behind other centers. So a lot of my colleagues have moral distress that we're resuscitating 22 weekers and there are places that are better…from their perspective, we as a unit are not trying hard enough to get better fast enough. So in ideal world, we would be going to [other institutions] every couple of months and like steal their protocols and implement, or do QI around the small baby unit more heavily than we currently are.”
“I went to [a conference] and it was a neonatologist talking about periviability and ethics bringing [survival] statistics from Japan…and it's like, okay, what does Japan do that we can learn from, you know? It's being done. So I think it's fascinating…I mean, surfactant changed the world. GBS prophylaxis changed the management of these infants. So there is still stuff to be learned.”
Population-based survival data was also referenced in the context of prenatal counseling, with differing perspectives regarding the utility of incorporating current literature when talking with families. Obstetric providers were more likely to report a preference for statistics and other objective metrics during these conversations in addition to parental goals or values when compared with neonatal providers.
“In order to have patient autonomy, it needs to be informed based on the best data that we have. And I think that's why having it be multidisciplinary is most important, because I think we all are biased in our counseling, from our perspectives on the maternal side versus your perspectives on the pediatric and neonatal side…I think a lot of patients will look to the Internet [and] that Internet data is not our institutional data. What is being done at some of those institutions that are on the forefront…are unique and not necessarily translational to our institution.”
“I noticed when neonatologists consult on patients, they kind of they start with broad strokes and they talk general. And sometimes I've wanted a little bit more specifics, more details. I struggle with that in my own counseling because patients want to know what's the risk of that? Sometimes we honestly just don't know. But there are some data out there in terms of survival and morbidity [that we could share.”
Dynamic Factors
This theme consisted of the relatively fluid factors of institutional culture that interact with both Perception and Statements of Information. There were two subthemes identified: time and team dynamics. Team dynamics was further divided into three sub-subthemes: perceived influence of team role on one's beliefs, interstaff discussions on the topic of periviability, and interdisciplinary collaboration, including differences in interdisciplinary communication styles, the role of the palliative care team, and maternal versus infant well-being.
The subtheme Time was discussed as a tool for cultural change at both the individual and institutional level. With the passage of time and introduction of new data or experiences into the clinical discourse, many participants reported an evolution in perspective on perviability.
“I don't know if it's necessarily training so much as having gotten older, which is a weird way to phrase it. But I think in residency, it was easier for me to be like, this is ridiculous. I can't believe anybody would do this. And now, whether it's that I'm out of training or a little bit older in life, I think it's easier for me to understand the parental drivers.”
“I think it is probably evolved. Back when I was a resident, I think 24 weeks was a relatively hard stop. And I don't know, I guess my personal views on periviability have changed. But I think the things that I have learned about are the things I had the time to really see play out…[a] 26 weeker that's 400 grams is totally different than that 23 week 400 grams…I do also think that a lot of the data and the advances in the NICU's abilities…it's different then it used to be.”
This in turn impacted approaches to medical management, creating space for innovation and new initiatives. As one attending said:
“I was involved with trying to develop the small baby team, [and] we got a lot of pushback early on like, well, we know how to take care of preemie babies…And then it was about 6 months later when some of the nurses that were really involved in the project said, hey, you know what? I think you're right. I think our 23 weekers are different than our 26 weekers, and maybe we should have different developmental protocols for the first 72 hours of life for those babies. So now, in conversations with the small baby group and even within our division, [things are different].”
The subtheme Team dynamics was described as foundational to the culture of the unit, with an emphasis on interdisciplinary collaboration between staff members and between specialties (n = 134). Several participants underscored the importance of a strong pediatric–obstetric relationship in the care of periviable infants, describing collaborative maternal–fetal management and joint counseling as a critical first step.
“I like that we offer joint counseling with MFM. I think that is absolutely invaluable when I've been a part of it. The conversations are just so much richer and so much better for the parents when we're both in the room together and we can have the conversation. And you just realize that, it's not usually that the MFM talks and then the neo talks. It's like the mom asked one question, the MFM answers it and then bounces it to the neo and bounces it back. So there are a ton of interplay.”
“I also appreciate that the discussion is a joint discussion with MFM because I think it's tough to define periviability from one discipline—just MFM's version or just the neonatology version.”
There were also references made to the seemingly inherent tensions that exist between the two specialties due to the interdependent nature of the maternal–fetal dyad.
“I guess where a lot of that stems from is reproductive health rights. Neonatology pushing the boundary of viability and the advances that they have made in postnatal care are wonderful…[but] it just feels like they're encroaching on reproductive rights. I don't necessarily think that those things should be tied, but I think that they are, if that makes sense…the law and medicine don't necessarily align, and I think that's where some of that struggle can be. Or at least the historical struggle.”
“It's hard when patients don't see their own health as the priority. As you know, patients will walk over coal hot coals and stand on their head if they thought it would mean a better outcome [for their baby]. So I think the most challenging part for me is when you're trying to emphasize how important their own health is, and that doesn't appear to resonate with them.”
With regard to interdisciplinary dynamics on the unit (i.e., neonatal physicians, nurses, respiratory therapists, etc.), several participants described the importance of clear communication as well as recognition of the skills and perspectives different team members bring to the care of periviable patients. Routine interdisciplinary discussions, debriefs, educational sessions, and working groups were highlighted as ways to maintain a positive unit culture.
Discussion
This exploratory, qualitative interview study shows that although the idea of “culture” is multifaceted, individual ethics, interpretations of quality of life, and those factors that contribute to or protect against experiences of moral distress play a primary role in how medical staff at our institution conceptualize the care of periviable infants. Protocols and population-based outcomes are a critical component of antenatal counseling and perinatal management;[25] however, the philosophy of this unit appears to be rooted in staff insights and attitudes more than empirical or objective data. These results are supported by other studies that have demonstrated the impact of personal, cultural, and social factors on clinical decision-making at the limits of viability.[26]
[27]
[28]
Study participants' descriptions of moral distress were particularly noteworthy and warrant further reflection. Although the concept of moral distress was referenced frequently throughout interviews, several participants denied experiencing moral distress themselves—some doing so while concurrently describing similar states of psychological disequilibrium. These results are supported by Prentice et al, whose longitudinal examination of clinicians caring for extremely low gestational age neonates found nuanced and heterogenous accounts of moral distress among team members, with the majority of clinicians acknowledging potential ethical challenges while not endorsing sustained or substantial moral distress.[24] Our study reinforces the complexity of moral distress that Prentice et al describes, demonstrating that while moral distress is an important part of understanding institutional culture and should be acknowledged for the purpose of preventing provider burnout and preserving moral integrity,[29] the experiences of staff members that shape institutional culture are often more dynamic than conventional definitions may imply. While individuals may find prognostic uncertainty, differences in institutional practices, or insufficient clinical continuity distressing, there may be room for the development of moral resilience through a recognition of clinical subjectivity. Moral resilience and the interconnected act of reframing challenging scenarios, acknowledges moral complexity while enabling clinicians to care for the patient in front of them.[29]
[30] We contend that an approach to periviability in the NICU built on moral resilience validates the emotionally demanding nature of the work while preserving individual integrity, preventing moral injury, and facilitating good patient care.
It is important to explicitly state that counseling and management at the limits of viability must be individualized and based on evolving data. Expectant parents vary in their values and attitudes toward obstetric and neonatal decision-making at these gestational ages[31]
[32] and may consider several legally and morally permissible options. High rates of morbidity and mortality remain a legitimate concern for infants born between 22 and 24 weeks (even in centers that routinely perform intensive care interventions), and in many cases, maternal health might be significantly affected as well. The preservation of bodily autonomy and parental self-determination mean that some families may choose perinatal hospice care when faced with a threatened delivery, while others may seek pregnancy termination. Thorough options counseling should include all paths available to pregnant persons, specific to the case in question.[33] If, however, a parental decision is made to pursue intensive care, clinicians then have an obligation to perform their clinical duties in a way that fully supports that choice—so long as there is no violation of the principles that guide professional ethical practice and the decision remains within “the zone of parental discretion.”[24]
[34] We should aim to achieve this goal through optimizing our clinical environments, taking innovative approaches to enhance all aspects of neonatal care, including unit culture.
As Lantos and Meadow wrote in their 2009 joint commentary, “protocols reflect values as well as facts, but values create facts.”[35] NICU clinicians vary widely in their prognostication, and use criteria to either initiate or withdraw life-sustaining therapies for periviable infants[36] based on what is often an overestimation of adverse outcomes for this population.[20]
[37]
[38]
[39] Prognostication by medical providers has therefore been described by some as a clinical intervention in and of itself, whereby the negative supposition of a certain outcome (e.g., nonsurvival) makes that outcome more likely through subtle changes in behavior or practice.[40]
[41] In this way, a culture in opposition to periviable intervention—either implicit or explicit—has the potential to generate a “self-fulfilling prophecy,” built on an expectation of failure and sense of “learnt helplessness”[42] that negatively impacts both patient survival and family experiences.[43]
[44] We believe that understanding the contribution of perception, statements of information, and team dynamics to culture is a necessary first step toward creating institutional change.
