Techniques to Communicate Better With Parents During End-of-Life Scenarios in Neonatology

The study took place in Sainte-Justine hospital, a tertiary care mother-child hospital affiliated with the University of Montreal, with ∼4000 deliveries a year and a 67-bed NICU.

We used similar methods reported in our previous studies.^11,12  Participants performed a videotaped simulation using a high-fidelity manikin assisted by a Neonatal Resuscitation Program (NRP)–trained provider (neonatal nurse or respiratory therapist). They were informed that 2 actors, playing the role of the parents, would be present at the resuscitation and to interact with them as they would do in practice. A term neonate would be born after an urgent cesarean delivery for fetal distress. They had 3 minutes to prepare for the birth. The manikin was programmed to remain pulseless despite resuscitation. Standardized actors were trained and followed specific guidelines (Supplemental Information). Immediately after each simulation, standardized debriefing occurred.¹¹ 

To be representative of the diversity of clinicians who participate in resuscitations in our institution, we aimed to recruit 30 NRP-trained participants, with varying backgrounds and experience.

To represent a wide sample of perspectives, we aimed to recruit 20 evaluators with various interdisciplinary backgrounds, including bereaved parents.¹³  Each video was viewed by all evaluators, including the 2 standardized actors.

Two registered NRP instructors independently evaluated resuscitations using the NRP performance checklist,^7,12,14  and their scores were averaged.

The evaluators scored videos on a scale of 0 to 10 for (1) overall performance, (2) communication with the parents during resuscitation, and (3) communication with the parents after resuscitation. They were asked to justify their scores with 2 open-ended questions:

1. Identify up to 3 aspects that made interaction with parents optimal.

2. Identify up to 3 aspects of interactions that could be improved (or should be avoided).

Finally, each was asked to rate the top 10 communicators.

A list of the most reported positive and negative themes invoked by evaluators was developed. Of those, behaviors that could be easily observed were identified. All the videos were then watched by an independent evaluator (blinded to the previous scores and to the study protocol) who examined whether each participant had displayed these behaviors. The frequency with which the top 10 communicators demonstrated each of these behaviors was compared with the remaining participants.

For statistical analysis, the participants were separated in 4 different categories: junior residents, senior residents, neonatologists and fellows, and transport team personnel. The appropriateness of these groupings was tested using the intraclass correlation coefficient. Differences between groups were tested by using nonparametric statistics.

All answers to the open-ended questions were analyzed by using thematic analysis^15–17 : themes and subthemes were developed by 3 independent investigators who coded 50 randomly chosen evaluations. Themes and subthemes were then defined, and 50 more evaluations were coded by 2 investigators. After 85% coding agreement was reached, the coding themes and subthemes were finalized, and the remainder of the evaluations were coded by 1 investigator (Supplemental Information). Comparisons between quantitative data (evaluation on a scale of 10; being a top communicator or not) and answers to open-ended questions were also performed.

Participants consented to being observed, videotaped, and having their videotapes evaluated. They were assured these were anonymous and would not influence their clinical evaluations. Evaluators were informed about the sensitive nature of the study and signed a confidentiality agreement, agreeing to keep the content of the videos and evaluations confidential. Recruiting bereaved parents was ethically challenging and involved particular care (see Supplemental Information). The protocol was codeveloped with a bereaved veteran resource parent (M.S.). The protocol was approved by the Institutional Review Board of Centre Hospitalier Universitaire Sainte-Justine.

Thirty-one participants were recruited: 15 pediatric residents (6 junior [postgraduate year 1 and 2] and 9 senior residents [postgraduate year 3 or 4]), 5 neonatal fellows, 3 neonatologists, 3 neonatal nurse practitioners, and 5 transport and resuscitation team providers (2 nurses and 3 respiratory therapists).

The videos were analyzed by 21 evaluators: 3 involved in the simulation (the 2 standardized actors and the assisting clinician), 6 members of the neonatal team (1 neonatologist, 1 fellow, 2 nurses, 2 neonatal nurse practitioners), 2 obstetrics providers (a maternal-fetal medicine specialist and a nurse), 1 pediatrician, 3 allied health professionals (social worker, psychologist, and respiratory therapist), and 6 parents. Nine potential parent participants were contacted (Supplemental Information): 8 answered the invitation, 1 declined, and another did not confirm his participation. Six participated: 4 mothers and 2 fathers. Five had experienced a neonatal death, and the sixth had an infant who faced ongoing complex medical problems and had had 3 extensive resuscitations.

The technical resuscitation “NRP scores” were between 69% and 100%. The transport team and neonatal nurse practitioners scored slightly higher than the other groups (mean score: 88% vs 81%; P = .04) (Table 1).

Each video was examined by 21 evaluators, leading to 651 evaluations. In all groups, communication scores were lower during the resuscitation (Table 1). Neonatologists, fellows, and the transport team and neonatal nurse practitioners had higher scores than residents for communication after the resuscitation (Table 1).

In each of the 651 evaluations, 3 positive and 3 negative evaluator comments were identified, for a total of 3906 items, which were subject to thematic analysis.

Although evaluators were not asked to score communication before the resuscitation, 16 evaluators reported (in open-ended questions) examples of positive and/or negative interactions between providers and parents during the 3-minute preparation time (Table 2). At least 1 of 3 positive main themes were invoked by all evaluators. Good communicators (1) introduced themselves in simple language, (2) asked the name of the infant and used it, and (3) prepared the parents for a potentially difficult situation (Table 2).

When identifying positive interactions during the resuscitation, evaluators invoked 6 themes (between 2 and 6 per evaluator) (Table 3). Good communicators (1) recognized the presence of the parents and allowed the father to approach the bedside; (2) used the name of the infant; (3) prepared the parents for death in a stepwise fashion; (4) used the words death, dying, or dead; (5) remained calm; and (6) made a clear decision to stop the resuscitation without asking parents’ permission (Table 3).

When identifying positive interactions, evaluators invoked 8 themes (between 3 and 8 themes invoked per evaluator): (1) clearly and unambiguously stating the infant had died, (2) avoiding medical jargon and metaphors related to death, (3) making a clear statement that this was not the fault of the parents, (4) listening to the parents and providing silence, (5) providing proximity (either between clinician-parent, mother-father, clinician-infant, and/or parents-infant), (6) speaking about what happens after a death, (7) offering support from hospital staff and parents’ loved ones, and (8) offering future support (Table 4).

Each of the 31 evaluators identified 10 top communicators, for a total of 310 votes. The top 10 communicators were more likely to be neonatologists or neonatal fellows (P = .035). There was broad agreement about the top 10 communicators, 8 of whom were scored in the top 10 by 81% of evaluators; 2 others were scored in the top 10 by 62%. One or more of the following 5 words were more likely to be used in the positive comments for top 10 communicators: “empathy,” “compassion,” “warm,” “human,” or “trust” (P < .001). Notably, when these words were used in the negative (“lacking empathy,” “not compassionate,” etc), participants were never in the top 10.

In the list of all possible positive interactions, 21 behaviors that can easily be observed were identified. For example, “sitting down” and “using the infant’s name” were identified. In contrast, “empathy” or “compassion,” which are subjective, were not. Then, all the videos were reviewed by an independent reviewer to examine whether these behaviors were present (Table 5); 90% of the top 10 communicators adopted at least 20 of these 21 behaviors. Importantly, such behaviors were strongly associated with the use of the words “empathy,” “compassion,” “warm,” “human,” and “trust” in the open-ended evaluators’ comments (P < .001; use of 1 of the 5 words). Providers who displayed all 21 behaviors had evaluations >9 out of 10 by all evaluators.

Communication scores and rankings given by neonatal and non-neonatal providers were generally similar. Actors gave significantly higher scores compared with the others (P = .04), and parent evaluators gave lower scores compared with provider evaluators (P = .045). Discrepancies of >2 points, on our scale of 10, between evaluations occurred in 19% of the 651 evaluations. The 6 participants who had discrepant scores usually had lower scores from parent evaluators, who had different perspectives from providers and actors. Discrepancies were mainly related to the language participants used related to death. For example, parents had problems understanding some sentences, or thought they were insensitive: “your infant was born without a heart” or “her heart never came back.” Some of the language commonly used in the medical field was judged to be negative by all parent evaluators, for example, “letting the infant die” or “allowing natural death.” Other discrepancies were less common. For example, 1 of the participants placed the hands of the father on the infant early in the resuscitation, eventually showed him how to provide cardiac massage, asked him to provide it, and went to speak to the mother for 20 seconds, informing her that resuscitation would be stopped and that her infant was dead and would be in her arms soon. All parents evaluated this participant favorably (scores >8), even using this as an example of a positive interaction. In contrast, provider evaluators were critical and gave low marks, using this example as something to avoid.

This is the first study to evaluate communications skills during a simulated end-of-life scenario, with interdisciplinary participation and evaluation. The goal was to identify core behaviors that could be easily taught and emulated. All participants were adequate in their technical and communication skills, and they were NRP trained and often participated in resuscitations, some rarely as team leader.

The evidence regarding how to communicate with families before a critical delivery or when a child becomes unstable is scarce and generally not addressed in resuscitation manuals. A resuscitation team may only have seconds to communicate with parents, but our results show that this time can be used effectively. Similarly, there is scarce evidence regarding communication with families during resuscitations. Although parental presence remains controversial,^18–24  the mother is inevitably physically present at birth (although sometimes under general anesthesia) and the father is often there. The communication scores during the resuscitation were the lowest. Indeed, this task is complex because it requires multitasking by the code leader. Ideally, a provider would be dedicated to do this task. Several institutions have developed specific training guidelines for such a family support role.^25,26  Realistically, in these rare situations, the additional personnel (when they exist) are generally diverted to help with the resuscitation. In this study, it was possible for more than half of the participants to communicate well (scores >80%), using 6 core behaviors. This took a total time of less than a minute and was feasible while leading a resuscitation.

Optimal communication after death comprised many elements. Not surprisingly, some of the behaviors identified in this article been described in the literature²⁷ : for example, guidelines on how to give bad news^28,29  or how to speak to parents about life-and-death issues.³⁰  Yet, none of these publications include all the core behaviors described in this empirical investigation.

Neonatologists and neonatal fellows had higher evaluations on average, suggesting that experience is important to improve skills. On the other hand, some junior residents with limited clinical experience were highly scored. Some seem to have “natural skills” for these interactions. To those who have this natural talent, our results may seem infantilizing: “introduce yourself,” “use the infant’s name,” “clearly state that the infant has died,” or “sit down.” Interestingly, when we asked the top communicators to describe what they did, they were mostly unable to inform us clearly, they “just did what they usually do.” But even “natural communicators” have bad days or days when these situations happen at a bad time, for example, at the end of a long call. In these situations, these core behaviors can be used as a checklist to make sure parents feel well treated. For those who start their training or for whom these interactions are stressful or come less naturally, it is possible to be “good with parents” by adopting some simple behaviors and avoiding others because these are perceived acutely by parents and classified consistently by professionals (Fig 1). Institutions can also have an impact on these core behaviors, for example, by including the name and/or sex of the child during “time-outs” before an urgent delivery or by making sure there are chairs available in all rooms where resuscitations and/or deaths occur.

It is not rare to hear that a compassionate provider is one with a good heart and that intentions are what counts. Some may consider such behaviors as less virtuous when they are part of checklists. Unfortunately, empathy and good intentions may not be enough. Caring clinicians can inadvertently display behaviors that are judged to be insensitive, such as leaving the dead infant alone or not using his or her name. Clinicians can display behaviors that ensure families will feel well cared for in these difficult moments, even if sometimes it may not “come from the heart.”

Although evaluators generally agreed, in some cases, parent evaluators disagreed with nonparent evaluators. We should remember the language we use to speak about death can be unclear. Although the actors were lay people, they were probably too involved with the medical world to realize some sentences made little sense or could be insensitive to families. Parents’ criticisms mainly had to do with jargon related to death (“born without a heart”), but some were about elaborate vocabulary, such as “allow natural death.” This vocabulary evolves in academic circles, without stakeholder involvement. We were also sometimes surprised by parental evaluations, for example, judging favorably the inclusion of the father-actor providing cardiac massage. This is a reminder that we should remain humble and curious and always include stakeholders in these kinds of investigations. The inclusion of parent evaluators in this study gave us unique insights. It was also a complex endeavor. Bereaved stakeholders are often not included in research or clinical initiatives because of risks to them. Although they are indeed a vulnerable population, excluding them from such activities omits an important and often complementary perspective.

This study has several limitations. It was performed in a single center, and results may be different in other centers. On the other hand, because we examined the perspectives of diverse evaluators on various resuscitators and communicators, which lead to >600 evaluations and almost 4000 examples of positive and negative interactions, the main core behaviors would probably be similar. These core behaviors are a skeleton on which to build additional skills with practice and training. Cultural sensitivity and flexibility in their application is also important. For example, in some cultures, the infant is not named before birth. This was also a simulation study, and it is unknown whether these scenarios would exactly translate to a real-life experience. The evaluations were also based on high-_fidelity videos, which leads to limitations. On the other hand, investigating these resuscitations in real life would be almost impossible.

The simulations were also limited to 1 scenario. We used this scenario because it had previously been safely used with trainees.¹¹  However, many of these core behaviors could be applied in other circumstances that are more frequent within the NICU,⁴  such as when an infant with respiratory failure becomes bradycardic and eventually dies, with or without extensive cardiopulmonary resuscitation. Developing these “hybrid” technical communication simulations was also appreciated by trainees. These could be developed for other situations. Neither the NRP nor the Pediatric Advanced Life Support textbooks include a score sheet related to interaction with parents. Some of the core behaviors identified in this study could be considered objective behaviors to validate in future studies.

This study describes core behaviors that are easy to teach and learn. “Empathy,” “compassion,” and “trust” were all words associated with these behaviors. The results show that concrete actions may be taken to make it possible for all clinicians to interact in a way parents can feel well treated, even during complex resuscitations, even on bad days.

 
• NRP

  Neonatal Resuscitation Program