Research has clearly documented the desire of many patients and family members for ongoing prognostic communication.1-6 Prognostic information is important to these patients and families because it is utilized to anticipate and prepare for predicted outcomes.7,8 Patients and families primarily hold physicians responsible for providing prognostic information.7,8 A majority of patients and family members who desire prognostic information prefer to be asked how much information they want in order to negotiate the content and timing of information.4,8,9
Nurses, physicians, and patients view honest, clear communication of prognostic information presented in a sensitive manner as an essential moral and legal right.9 Yet, physicians are often hesitant to provide this information.6,10-12 Physicians often avoid communicating an unfavorable prognosis to patients and families or provide overly optimistic or vague prognostic information in discussing death and dying.8,13-15 Many physicians provide this information only if asked and avoid this discussion by focusing on treatments.8,10 Nurses and physicians often underestimate prognostic information needs of patients and family members and overestimate their comprehension of this information.15
Research on actual prognostic communication practices in intensive care units (ICUs) has focused, primarily, on end-of-life family care conferences with physicians. Curtis and colleagues16 observed physicians and families in care conferences in 4 hospitals and delineated 2 descriptive frameworks: (1) focused on content of communication and (2) the other focused on style of communication and provider support to family members. Barton and colleagues17 analyzed end-of-life discussions in an ICU between physicians and families and identified 4 phases of communication. Of these, phase 2 was identified as of critical importance, because in this phase a listing of patients' medical problems by the physician implied a terminal status. Consensus with family members regarding this status was necessary in order to move to phase 3: decision making.17 Contradictions in communication between physicians and families in end-of-life conferences have been identified, as well as physician strategies to address these (Hsieh H. Deductive Content Analysis of End-of-Life Decision-Making in the ICU [unpublished dissertation]; 2004:1-179.)
Few studies have focused on critical care nurses' prognostic communication with families. A study by Shannon and colleagues18 analyzed critical care nurses' content and quality of communication with families immediately before, during, and after ICU care conferences. Liaschenko and colleagues19 conducted focus groups with critical care nurses and identified nurses act as "nodal points" for communication in ICUs and "integrat(e) large amounts of often disparate information into the big picture, thus making a meaningful whole of the patient's clinical situation."19(p227) Critical care nurses often work to "reframe" family members' perceptions of patient prognosis in order to achieve a realistic picture.20,21
There is a paucity of research on prognostic communication outside end-of-life family care conferences. Because communication occurring earlier in patients' trajectories frames and influences these discussions, an awareness of communication prior to end-of-life conferences is imperative. In addition, critical care nurses' actual prognostic communication in clinical practice has not received much attention. Because a majority of studies on this topic, to date, have focused on oncology patients, knowledge regarding communication with other types of patient populations is lacking. Thus, a study was implemented to shed light on these areas and peel away some of the layers of complexity of this topic.
METHODOLOGY AND RESEARCH PURPOSE
The highly technical environment in the ICU and the rapidly changing status of acutely ill patients cared for in this setting result in a complex, emotionally charged, and ethically laden milieu. It involves "complex human situations for which ethnography is most appropriate."22(p285) Longitudinal immersion in an ICU setting was deemed necessary to adequately capture the dynamic complexities of prognostic communication in a clinical setting between groups of people. Consequently, an ethnographic study was implemented with the purpose of describing prognostic communication of critical care nurses and physicians to family members of patients at high risk for death in an ICU.
SETTING
The setting was an open 22-bed adult medical/surgical ICU in a large urban, nonteaching hospital in a Midwestern region of the United States. This ICU was chosen partly because of its variety of patient populations, with a 74% medical patient population and 26% surgical population. The ICU admitted approximately 180 patients per month and averaged 10 patient deaths per month. Close to 100% of these deaths were following limiting life-sustaining treatments. Although it was an open ICU, a majority of patients with the highest acuity were followed up by the critical care physicians. Thus, this ICU was identified as an appropriate setting to study end-of-life prognostic communication.
Thus, this ICU was identified as an appropriate setting to study end-of-life prognostic communication.
STUDY METHODS
Sampling and Sample Inclusion Criteria
Purposive sampling was utilized in this study, because it is the most appropriate sampling method to meet ethnographic research goals.22 Nurse and physician key informants were identified by the researcher based on their ability and willingness to provide detailed information of their experiences of prognostic communication in this ICU. Discussions with key informants are a cornerstone of ethnographic research, to acquire detail on the research topic.22
The sample included critical care nurses and critical care physicians. Inclusion criteria for nurses included registered nurses providing care in this ICU for patients identified by study criteria as being at high risk for death. High risk for death was identified as medical diagnoses of (1) 2 or more body systems in failure, (2) sepsis, (3) neurological diagnosis associated with a mortality rate greater than or equal to 50%, (4)late stages of cancer, or (5) or persistent vegetative state. Because this study focused specifically on critical care providers, inclusion criteria for physicians included physicians with certification in critical care medicine or working on a fellowship in critical care medicine. Thus, physicians who did not specialize in critical care medicine were excluded.
Family members of patients identified as at high risk for death were also included in the sample. Family members were at least 18 years of age and able to speak English well enough to be understood by the researcher.
Samples
The sample included 7 of the critical physicians working in this ICU when the researcher was present and 3 critical care fellows who rotated through the ICU during the data collection period. All the attending physicians had specialty training and certification in pulmonary medicine, as well as having completed a fellowship in critical care medicine. Fellow physicians had completed 1 year of ICU fellowship training (see Table 1 for physician sample demographics).
Approximately 64 nurses were employed in this ICU during the data collection period. Because patients identified as at high risk for death were the most acutely ill patients in the ICU, many of the same experienced nurses cared for these patients. Thus, 14 ICU registered nurses were included in the sample (see Table 2 for nurse sample demographics).
The sample of family members included 20 family members (see Table 3 for family sample demographics). One family member was unaccounted for in the demographic information beyond sex, because of a missing demographic sheet.
Consent Process and Institutional Review Board Approval
A description of the study, including study objectives, methods, and confidentiality of data, was described to potential participants prior to consent. The consent form was reviewed and questions answered by the researcher prior to written consent and demographic forms being completed by the participants. Consent occurred for critical care physicians when the researcher first met the physician. With the critical care nurses, consent occurred when a nurse was identified as caring for a patient identified as being at high risk for death. Family members were consented once prognostic communication was anticipated or as soon as possible once this discussion was initiated.
The study was reviewed and approved by 2 institutional review boards. Both review boards authorized the researcher to review medical charts and record deidentified patient data.
Data Collection
The researcher observed communication in and around the ICU from critical care physicians and nurses to family members of ICU patients 12 to 18 hours per week for 4 to 5 days per week. This occurred from September 2008 through June 2009. Daily medical rounds and daytime family care conferences were observed throughout the data collection period.
The researcher asked the critical care nurses and physicians questions between patient care and during their breaks. Questions for providers included asking them to describe prognostic information they either received or communicated to another nurse, physician, or family member and their opinion of patients' prognoses. In addition, any significant trends or patterns identified by the researcher in the data were described to providers to elicit their feedback. Providers were asked whether these patterns were consistent with their experience, and why or why not. Data were compared between individuals, across groups, and with patients' medical charts to identify patterns in communication. Field notes were recorded in a notebook by the researcher while in the ICU and later typed into detailed manuscripts in another setting.
Approximately 5 months into data collection, 18 hours of formal interviews were conducted with nurse and physician key informants. This allowed time for the researcher to collect data based on observations and discussions on the unit and analyze these data for trends and patterns. Each nurse or physician informant was interviewed once, individually, in a private setting outside the hospital. All formal interviews were audiotaped and later transcribed verbatim. Interview questions were prepared ahead of time and focused on verifying and gleaning additional information regarding significant trends and patterns in the data.
The researcher continued to interview 5 additional nurses and 4 physicians until no significant new data were revealed. This was in addition to the original sample. Five critical care nurses were interviewed. Two nurses had 3 to 4 years of experience in ICU, and 3 had 15 to 30 years of experience. One of the nurses interviewed was the ICU assistant nursing manager, who frequently acted as day charge nurse. Four critical care attending physicians were interviewed, including the medical director of the ICU. All had at least 7 years of experience as a critical care specialist.
Data Analysis
Data analysis was concurrent with data collection, with three 2-week periods of time out of the field spent analyzing data. Based on Hammersley and Atkinson's23 recommendations for data analysis, events and processes were initially described while comparing patient cases to identify patterns in social practices (ie, prognostic communication) in varying circumstances. A "patient case" included all data pertaining to a specific patient regarding communication of prognostic information: (1) observations, (2) discussions, (3) interviews, and (4) information from the patient's medical record. Triangulation of this information provided a large scope of data from multiple perspectives. This aided in crosschecking inferences. Data were compared and contrasted throughout data collection to identify trends, patterns, and areas for follow-up. Areas for follow-up were brought back to nurses and physicians to obtain additional viewpoints and verify understandings.
RESULTS
Communication of Prognosis as Metaphor: "Painting a Picture"
The act of communicating prognostic information was described by critical care nurses and physicians as painting a picture. The prognostic picture communicated was the provider's prediction of future patient health outcomes. The purpose of painting this picture was to help family members "see" and thus "understand" the "whole picture" and anticipate and prepare for potential responses, including death.
Critical care nurse during an interview:
We do care conferences with the doctors[horizontal ellipsis] so the doctors can kind of paint the picture for them of what they think the outcome is going to be[horizontal ellipsis]
Critical care attending physician speaking to a critical care fellow about a patient:
The prognosis was poor yesterday, but we need to paint a slightly different picture today. Tell them he did have a big heart attack, but he's getting better.
Professional Roles in Communicating Prognostic Information
Critical care physicians and nurses perceived communication of prognostic information to patients and family members as a responsibility associated with the physician role. Nurses described "unspoken boundaries" regarding this communication, which precluded it from nurses' domain of practice. Physicians also did not recognize nurses as having a role in this communication.
Critical care attending physician during an interview:
Researcher: How would you describe your role in communicating prognostic information to family members?
Physician: As a physician you are responsible. You are the main person responsible. So, I think that's our job[horizontal ellipsis]. Who else would do it? Yeah, I view that's our job.
Critical care attending physician during an interview:
I think, as far as offering prognostic information, I don't think that's (nurses') role.
Critical care nurse during an interview:
There are "unspoken boundaries" regarding who can provide prognostic information, and I am afraid of "getting in trouble" if I communicate it to families.
Critical Care Physician Prognostic Communication
GENERAL DESCRIPTION OF PROGNOSIS AND GOAL OF COMMUNICATION
Critical care physicians varied somewhat in their description of prognostic information to family members. Some physicians utilized statistics to describe probabilities of outcomes and/or utilized metaphors, whereas others did not. Most attempted to individualize communication based on the family's current understanding and emotional level. All the physicians shared the same goal, however, which was to align family members' understandings of the patient's status and prognosis with their own.
Most attempted to individualize communication based on the family's current understanding and emotional level.
Critical care fellow physician during an interview:
Researcher: What is running through your mind when you discuss prognosis with family members?
Fellow: First, I assess what the family already knows and their perception of what is going on and what this means to them. If the family are physicians and nurses, then they know exactly what is going on since they "think just like me." But if not, you never know what they may be thinking. I act as a "buffer" and try to moderate perceptions, which are extreme either way, either overly optimistic or overly pessimistic[horizontal ellipsis]. I try to bring them to a middle ground, which is where my thinking is because this is where I am at[horizontal ellipsis]. It's my baseline and where I work from.
Critical care attending during an interview:
Researcher: How do you talk to families of patients who you think will die?
Physician: I vary the words depending on the family and the situation, but I use words that make sense, not medical words. I lead them down the path I want them to go.
Informal Communication of Prognostic Information
Physicians frequently communicated patient information informally to family members, either at the patient's bedside or in the hallways of the ICU. Generally, these conversations were spontaneous and short in duration. They often occurred during medical rounds, if family members were in or near the patient's room at that time.
Physicians usually began informal communication by expressing an overall perspective of how the patient was doing, using words such as "good," "well," "better," "worse," or the "same as yesterday." This was followed by current information regarding "clinical markers," specific data significant for the patient. Clinical markers varied from patient to patient, but included laboratory values, such as hemoglobin level or white blood cell count, chest x-ray results, or vital signs. A general prognostic statement often followed, especially if prompted by the family, such as when the patient might be extubated or transferred out of the ICU. This was almost always followed by a statement of uncertainty. Family members learned which information to focus on and what questions to ask based on role modeling of this informal communication.
Critical care physician speaking to a family member during rounds:
Physician: He's doing pretty well today. Just needs a little more time. (Description of clinical markers, including current level of support on the ventilator and blood oxygen level.) "Everything looks good[horizontal ellipsis]. So, he just needs more time. He needs more time to let his lungs heal."
Family: How much longer on the ventilator?
Physician: We'll turn down the oxygen, have him breathe more on his own, and see how he does. Probably by the end of the week. Every day is a different day around here, so you never know.
Ambiguity was prevalent in informal day-to-day physician communication of prognostic information to family members, as well as in response to researcher questions regarding patient prognosis.
Discussion outside a patient's room during medical rounds:
Critical care attending physician: He's finally showing some improvement.
Researcher (directed to the critical care physicians): So, do you expect this patient will survive?
Fellow (responded very quickly): No.
Researcher: You don't think he'll survive?
Fellow: Not necessarily[horizontal ellipsis]. He's not dead yet.
Critical care fellow speaking to the researcher while in the ICU about a patient:
Researcher: Do you agree with the decision from the care conference to trach (perform a tracheotomy) the patient and give her 2 weeks to get better?
Fellow: Yes.
Researcher: So, you think 2 weeks will make a difference?
Physician: It might.
Researcher: So, you think she may get better?
Fellow: Ahhh, I don't know. She's pretty fragile.
If there was significant negative prognostic information to be communicated leading to end-of-life discussions, this news was usually communicated formally in a private setting. Bad news perceived as less significant was usually communicated informally at the bedside. When this occurred, it usually followed a pattern of good news being communicated first, followed by the negative prognostic information. Uncertainty was almost always communicated immediately after prognostic information. Hope was sometimes expressed by physicians, but only if physicians wished to continue with aggressive cares and had not reached the point of wanting to withhold or withdraw life-sustaining treatments.
Critical care fellow speaking to family during rounds at the patient's bedside:
Fellow: Day by day he keeps getting better, so that's good news. There is an almost certain possibility that he'll lose some of his fingers and toes. It is too early to predict.
Family: How certain are you of this happening?
Fellow: It's not a guaranteed thing, 'cause nothing is guaranteed. But, I prefer not to say it won't happen and then it does. (The patient's wife starts crying.) There may still be more healing that may occur, and we'll let the extremities declare themselves before we decide what to do[horizontal ellipsis]. Since he's getting better, I want to lighten the sedation to see if anybody's home[horizontal ellipsis].
Family: That's certainly hard to hear.
Family: There's still some hope.
Fellow: Certainly, there's always hope.
Formal Communication of Prognostic Information
Significant negative prognostic information (eg, dying, terminal disease) was usually clearly communicated to family members for the first time in formal communication between ICU physicians and family members. Physicians were often prompted to initiate formal discussions by the bedside nurse or charge nurse, family member, or the ICU social worker.
Formal communication was usually scheduled 24 or more hours in advance and took place in a family waiting room in the ICU. The critical care attending and/or fellow physicians, patient family member(s), and ICU social worker and/or chaplain commonly attended. Sometimes the patient's bedside nurse also attended, but not always. The bedside nurses who attended the care conferences, which the researcher attended, did not actively participate in the formal end-of-life prognostic discussions. They observed instead. Formal care conferences were often preceded by a critical care physician note in the medical chart, which described the patient's poor prognosis and the need to "have a family care conference to discuss withdrawal and/or withholding life support."
The content of formal communication followed a general pattern. The discussion usually began with introductions, and then the critical care physician explicitly stated the goal of the conference: to determine what the patient would want if he/she was able to communicate. Next, the physician described, in detail, the patient's medical diagnoses, treatments, and health responses prior to and during hospitalization. Lay terms were often used instead of medical terms, or medical terms were used and explained by the physician. This was followed by the physician's explicit description of the patient's prognosis, often utilizing percentages to illustrate the probability of recovery. Indications of patient wishes for end of life, including conversations with family and friends and/or a health directive or living will, were discussed next. The family's view of the patient's previous quality of life was often solicited by the physician, as well.
The physician(s) answered family questions throughout the discussion and then toward the end, told the family "you need to decide where to go from here." If the family decided to withdraw support, the physician described how this would be implemented, and a time frame for withdrawal was negotiated between the family and physicians. At this point, the physician usually made a recommendation to change the patient's code status to a do-not-resuscitate (DNR). If a finite decision was not reached, or a family did not come to a decision that the physician agreed with, (although this disagreement was not expressed to families), another meeting was set up to be held 2 to 3 days later to reevaluate the situation.
What were noticeably missing from formal communication, which were prevalent in informal communication, were any ambiguous or "wishy-washy" statements or those expressing uncertainty or hope. This was especially evident in formal prognostic communication for which the physician had already decided prior to the discussion that it was time to withdraw life support. Although the communication in formal discussions was much clearer, as compared with informal communication, the word "dying" was never used in any of the formal discussions observed, and the word "death" or "dead" was rarely used. Euphemisms were utilized instead, sentences were structured in ways that did not include these words, or sentences were left "hanging" without completing them with one of these words.
Is a "Care Conference" a "Care Conference"?
One issue related to communication of prognostic information, which was confusing for the researcher, was there were many times when a discussion between providers and family members was scheduled ahead of time and identified by both nurses and physicians as a family care conference. Thus, the researcher expected the discussion to follow the pattern of formal prognostic communication just described. However, these discussions, instead, followed the pattern of informal prognostic communication. They were held at the patient's bedside or just outside the patient's room, standing up, without the chaplain or social worker, and without the detail and structure of content that formal communication followed. The researcher observed a "care conference," which a family member had requested, held in the doorway of a patient's room. Many people were walking past, while the patient's daughter attempted to discuss her desire to withdraw life support for her mother. Thus, it became apparent that the label of care conference could represent either formal or informal prognostic communication, even when end of life was the topic of discussion.
Critical Care Nurse Communication of Prognostic Information
Despite the lack of nurse and physician recognition of the nurse's role in prognostic communication, nurses played a significant role in this communication. Many nurses described waiting until the physician communicated prognosis initially with patients and families and then stepping in to "fill in the gaps" in physician communication. This involved interpreting physician communication for patients and families and providing more detailed prognostic information, if necessary. Nurses expressed specific concerns regarding ensuring that family members understood the potential short- and long-term implications of treatments on patients' quality of life.
Critical care nurse during an interview:
I explain things in layman's terms to families because when the doctors talk to families they have no idea what was said and don't want to admit it because they don't want to admit they don't understand. The physicians say they need a trach, but they don't explain what that means. Families don't ask question about long term.
Critical care nurse during an interview:
The nurse described a situation recently where a patient had an injury, which resulted in quadriplegia. The physician said to the woman, who was awake, alert, and able to mouth words, that he was going to put a "halo" on her and a feeding tube in her to give her food[horizontal ellipsis]. The nurse said she "couldn't stand all the medical terms" the physician was using because she felt the patient didn't understand what he was saying, so she said she stepped in and said, "Let me give this a try. '(Patient's name), do you understand that when the physician puts the halo on you will NEVER walk again and that when he puts the feeding tube in you will NEVER be able to eat food again?'" The nurse said, "The patient made a frown and lipped 'I don't want any of those things.'"
Critical care nurse during an interview:
There has been several times that the doctors will come in with the patient and talk about their adv[horizontal ellipsis] about their code status. And, OK, we're going to be a full code, and then they leave. And then I'll come in and say, "So you guys talked about it and so[horizontal ellipsis]," and I will just kind of reiterate, "If her heart would stop, we're going to pound on her chest, we're going to shock her, and they're like, 'Oh. And, and will that hurt?'" And I said, "It's going to hurt a lot." "Oh, well maybe we don't want that for our 85-year-old mom." And then they start to reconsider things[horizontal ellipsis]. They don't think about the breaking of the ribs and the shock that could hurt them and stuff like that. And so I think sometimes the way the doctors maybe word things when they're talking with families, too, the families don't get how painful that could be, or maybe that's not what the family member wanted[horizontal ellipsis]. I tell them, "I'm not trying to scare you or anything. I'm just telling you the reality of the situation, if it were to happen, this is what we're going to do or this is what it's like."
There were occasions when nurses did provide prognostic information, either directly or indirectly, prior to physicians. This usually involved situations where nurses perceived a high level of certainty regarding a patient's negative prognosis, especially if there was consensus with other nurses and/or physicians regarding this view and death was perceived as imminent. Both very experienced nurses and those with only 3 to 4 years of ICU experience described communicating this information.
Researcher: How do you work with family members who are not aware of the severity of the patient's condition?
Nurse: I try to bring them along slowly. I tell them what could possibly happen[horizontal ellipsis]. There was a patient I cared for last week, and all the nurses and doctors were talking at the desk about how this patient was going to die, but the surgeon gave him the option of surgery! I finally said to the family, "Look, he's either going to die with you all around, or he's going to die in the operating room." I said that because if I was them, I'd want to know.
Critical care nurse:
I had a patient who I was fairly certain would die, so I looked into one of the family member's eyes and leaned in very close to them and said, "He is very, very, VERY sick!" (The nurse leaned in within a few inches from the researcher and stared straight into her eyes while saying this.) The family didn't understand how sick the patient was, and when I said this, they suddenly jumped and their eyes widened[horizontal ellipsis]. They finally got it.
Researcher: Do you ever tell (family members) a little bit more, like I don't think they're going to survive?
Nurse: Yeah, if I think it's imminent, yeah. If I'm thinking it's 12 hours or a day, yeah. I'll say, "I don't know if they can survive today." Oh, absolutely. If they're on multipressors, and I can't get their pressure up. And a lot of times I'll tell them, "I don't think he can survive today, and if he does, I don't know what kind of damage his brain is going to have. He's had very low blood pressures, and something else you need to think about is that he may simply be surviving but how intact is he going to be at the end of this and would he want that?"
Nurses also played a critical role in preparing family members for the possibility of the patient's death. Thus, their communication was often structured with this purpose in mind.
Critical care nurse talking about a discussion she had with a patient's family member:
And (the patient's daughter) said, "Should I wait a little bit?" And I said, "We just drew labs for the dialysis to see how it's working, but I would guess that you're probably going to have to call your husband in and start to make some arrangements." And she was like, "OK, that's what I wanted to know," [horizontal ellipsis]instead of being blindsided in an hour when they couldn't get her mom's blood pressure up at all and (physician's name) had to come in.
Critical Care Nurse:
If I'm preparing them, then that final decision is a lot easier to make, than "Oh, he's been holding his own." Oh, oh, oh, oh, bam! We want to withdraw. "What? You're giving up? You're, you're not doing everything?" (raised inflection). But if they've been told all along, you know what? We're now doing this for him and he's not responding. We're now doing that for him and he's not responding to that either. These aren't good signs. But we're going to keep going. We'll see what happens tomorrow. Tomorrow comes, you know what? His white count is even higher. His kidneys are now[horizontal ellipsis] he's not making any urine. You know, this isn't good. OK, now they've got this much information. And then maybe the doctor will come in and say, "You know what? We've done what we can." Oh, yeah that's right. They have done this, and they did this today, and now they've done this. They're so much further down that road[horizontal ellipsis]."
Lastly, nurses assumed a central role in initiating code status discussions, either by discussing this topic directly with family members or requesting physicians initiate this discussion with families. Nurse concerns, which prompted these discussions, often focused on the futility of resuscitation for certain patients and causing more harm to the patient than good. Resuscitation, in these types of situations, was perceived by nurses as unethical and described as torture. Thus, nurses perceived themselves as being in an untenable position, because they recognized they would likely be the provider responsible for initiating cardiopulmonary resuscitation (CPR) in the event of respiratory and/or cardiac arrest. They also described feeling "caught in the middle," because of their sustained proximity at the bedside, which resulted in intimate knowledge of both patient and family wishes for code status, which sometimes conflicted with each other. Nurse concerns about code status were exacerbated by physicians' lack of follow-through. At times, code status discussion was delayed, previous discussions were not followed up on to identify family decisions, or the code status order was not processed according to hospital policy and procedure. Nurse facilitation of code status discussions significantly influenced patients' trajectories, because these discussions usually required prognostic communication and were often a stepping stone for other end-of-life discussions, including withholding other treatments or withdrawing support.
Very experienced critical care nurse:
(I)f I know the family who wants to go full board, and this is like, the patient is a frail, multidisease patient, that you know you're not going to want to code. And let's say they're on multiple pressors, and things are not going good. I may grab one of the docs and say, "You know what? Things are not looking good. You need to talk[horizontal ellipsis]." I may just put it in their face and just say, "You need to talk to this family. Do you really expect us to do CPR and everything on this patient?" And if they say, "Yep, I do." Then, OK, then that's their opinion, they don't want to talk to the family, that doesn't mean I won't go to the family and talk to the family about it, too.
Critical care nurse:
Nurse: (O)n a med/surg floor, those are things you really don't address; it's more like in critical care that code status needs to be addressed. I think, more often than not, it's brought up by the nurses. It's telling the doctor you need to talk to the family about code status. It's unfortunate. I mean, I'm glad that we do it, but I feel like physicians should be doing it more without us, you know, our little push to do it and stuff[horizontal ellipsis]. Sometimes, it will even be a family member saying, "You know, we talked before this, and they don't want this," and it's like, "Oh, did you talk to the doctor about this?" "No, we[horizontal ellipsis]." So, then we go back to the doctor and say, "Listen, this is what the family had told me. You need to go talk to them again and really get on the same page with them and get something definite figured out."
Interview with an experienced critical care nurse who works charge frequently:
Researcher: What about when it comes to code status?
Nurse: It's not addressed early enough.
Researcher: OK.
Nurse: It is not addressed early enough. That's all I've got to say about it[horizontal ellipsis]. Sometimes we're the ones who are the first times to even suggest it, and they're (patients and families) like surprised.
DISCUSSION
Many key findings from this study support previous research. These include physician and nurse communication focusing on "painting a picture" of the prognosis for families, attempting to frame or reframe families' understandings of prognosis to align with providers', and using communication to facilitate realistic goals and decision making and prepare families for the possibility of death.19-21
The content of formal communication in family conferences observed in this study aligns closely with the content framework described by Curtis and colleagues.16 This includes a detailed description of the patient's health history and hospital course, including treatments implemented and patient responses to treatments.17,21 This recitation by physicians in family conferences served an important purpose. It communicated indirectly to the family that the patient was dying or had little to no chance of recovery. This correlates with Barton and colleagues'17 description of phase 2 in end-of-life conferences, during which provider and family member consensus of prognosis and goals for care is necessary before decision making can proceed. Unwillingness of family to accept a poor prognosis can block further prognostic communication and defer hospice referral.17,24 This often leads to patients receiving aggressive treatments that providers may not feel are indicated, resulting in prolonged patient dying trajectories and iatrogenic suffering.24
Despite close alignment, some slight variation was noted between Curtis and colleagues'16 content framework and the pattern of formal communication observed in this study. In this study, physicians often set up care conferences after they had decided it was time to withhold and/or withdraw life-sustaining treatments. Thus, these end-of-life family conferences did not include statements of uncertainty or suggestions for additional treatments, which were included in the content framework described by Curtis and colleagues.16 Avoiding statements of uncertainty and discussion of treatment options served the purpose of assisting physicians in aligning families' perceptions of the prognosis with their own, by minimizing family members' hope for recovery. As a critical care physician in this study stated, "I lead them down the path I want them to go."
As noted by Norton and Bowers,21 the aligning of family members' understandings of prognosis with providers' can be interpreted in 2 ways. It can be interpreted as providers' efforts to facilitate informed consent or can be viewed as paternalistic.21 Providers in this study stated a desire to help families understand what is going on, in order to make appropriate, realistic decisions. Providers were very cognizant that they possess medical and clinical knowledge and expertise that most family members lack. Thus, their view was that their communication with families is an attempt to guide decision making so that decisions are made that are in the best interest of the patient. Although Slomka25 described communication and decision making in ICU as a negotiation between family and providers, this study suggests physicians control information in formal communication in order to guide family members to reach the same conclusion that the physician had already reached: the prognosis is poor, and it is time to withdraw treatment. Providers acting as gatekeepers to control both information and understandings of information have been identified by others.15,26 Achieving a consensus with family members by aligning family member and provider perceptions for patient care goals and the appropriate level of care is a recommendation for organizing end-of-life conferences.27
Achieving a consensus with family members by aligning family member and provider perceptions for patient care goals and the appropriate level of care is a recommendation for organizing end-of-life conferences.
Another recommendation for organizing end-of-life care conferences includes using the terms "dying" or "death" in these discussions.27 In this study, however, these words were avoided during formal communication at end of life. This may be interpreted as "accommodat(ing) the social taboo of not speaking of death directly."17(p17) However, it also can be interpreted as provider discomfort with discussing death and dying.15,24 Discomfort is plausible, considering the nurses and physicians in this study, as noted in other research, had little, if any, training in communication at end of life.10
End-of-life care conferences demonstrated shifting of responsibility for decision making to families. The statement "You need to decide" or "You need to make the decision" was consistently communicated from critical care physicians to family members during end-of-life care conferences. Use of the pronoun "you" rather than the pronoun "we" clearly places responsibility for decision making on family members' shoulders and takes responsibility from physicians. Sometimes physicians were willing to assume responsibility for providing a medical recommendation regarding withholding CPR, thus giving the family some guidance and support regarding what the physicians believed was the best course of action in the situation. Yet, recommendations for withdrawing support were never offered. There was no indication of collaboration in decision making between the family and the physician. Rather, the decision and thus the responsibility for the decision and its outcome were placed fully onto the family by the physician. This type of shifting of responsibility has been suggested as reluctance to assume the moral burden for death.25 Yet, shifting sole responsibility for end-of-life decision making to families, who are under extreme duress and lack understanding of medical and clinical knowledge, is unacceptable. Clearly, end of life is a time when families most need the support and guidance from physicians and nurses in the form of recommendations by providers, if families desire this, and shared decision making.27
Day-to-day prognostic communication, outside end-of-life care conferences, has not been the focus of much research. This study helped to fill this gap by describing this communication. A pattern of informal communication was evident, which focused on clinical markers identified by providers as important for a specific patient. These markers were analyzed over time to identify a general statement of how the patient was doing, using words such as "good," "better," "worse," or "the same." Any prognostic predictions were often delivered in an ambiguous manner and were peppered with statements of uncertainty, which differentiated it from formal prognostic communication in end-of-life care conferences. The prevalence of ambiguity and uncertainty in informal physician communication of prognosis suggests a tension experienced by providers regarding fulfilling their perceived responsibility of communicating prognosis with the uncertainty inherent in prognostication and perhaps hesitancy in assuming accountability for prognostic predictions. These informal discussions at the bedside set the foundation for formal end-of-life discussions in family care conferences.
An important consideration regarding the differences between informal prognostic communication at the bedside and formal communication relates to the needs of family members. Formal prognostic communication offers a structure that provides better support for families. This includes the privacy offered by communication in the family waiting room and the emotional support in the form of the social worker and chaplain. The structure of the content of formal prognostic communication, with its detailed, comprehensive recitation of the patient's health history and current status and discussions of patient wishes and quality of life, provides necessary support, as well. It serves to address potential gaps in family members' knowledge, assists in clearing up misunderstandings, and focuses the discussion. This, in turn, assists in aligning family members' understandings with providers' and facilitates decision making. A care conference communicated with an informal structure standing at the bedside is very different from a "care conference" communicated with a formal structure sitting down in a private setting with the social worker or chaplain. It is important for providers to recognize the differences between these 2 types of prognostic communications and consider family member needs in deciding which type is most appropriate.
Nurses described professional boundaries in relation to prognostic communication and hesitancy to communicate prognostic information, especially initial communication. Despite this perceived boundary restriction, nurses in this study demonstrated a significant role in prognostic communication. This included clarifying and interpreting physician prognostic communication for families, communicating aspects of quality of life and implications of possible interventions, and, occasionally, delivering prognostic information when it was delayed or not communicated. The perception of prognostic communication boundaries and the nurse's role in this communication has been described by others.28,29 Previous work is extended in this study, however, by identifying nurses' critical influence in initiating end-of-life discussions. This was often based on nurse concerns regarding addressing patient code status. Nurses either directly approached families themselves or prompted physicians to have this discussion. Discussion of code status was often the stepping stone for reviewing the patient's treatment plan and considering withholding and/or withdrawing treatments. Overall, nurses' prognostic communication work in this study was consistent with Liaschenko and colleagues'19 interpretation that nurses "manag(e) patients' trajectories of dying" in ICU.19(p229)
Considering the active involvement of nurses in prognostic communication in informal communication, their lack of participation in formal end-of-life discussions in this study is puzzling. The fact that the physicians did not solicit information from the nurses during family care conferences is also telling. Nurses possess intimate knowledge about patients and family members because of their ongoing proximity at the patient's bedside. Previous studies have verified nurses' feedback is often disregarded by physicians, despite their desire to be included in end-of-life decision making.30 The fact that nurses did not contribute to these conversations suggests that either they did not feel they had anything worthwhile to contribute, or they did not feel comfortable communicating in this setting. The lack of solicitation of nurse feedback from the physicians suggests physicians did not desire their input. It is plausible that because the physicians had often already made the decision to withdraw support prior to these discussions, physicians did not want to risk adding an additional voice to the conversation for fear that there might be inconsistency in the message being delivered to the family. Inconsistency in the message, including voicing statements of uncertainty, hope for recovery, or discussions of treatments, might confuse the family and prevent alignment of their understanding of the prognosis and goal for care with the physicians'. Yet, nurses' active involvement in decision making at end of life has been found to improve patient outcomes and increase nurse satisfaction.30 Thus, improving collaboration in communication at end of life between physicians and nurses has been recommended.30
Nurses described feeling "caught in the middle" in relation to issues of code status. They often possessed intimate knowledge of both patient and family wishes, which were sometimes in conflict with each other. There was frequently a delay between nurse requests for physicians to initiate or follow up on code status discussions and communication to the nurse of the code status decision and processing the order. This caused anxiety for nurses who recognized that they were most likely the provider who would need to initiate resuscitation in the event of patient arrest. Resuscitating a patient whom nurses perceived would not survive resuscitation or whose quality of life would be severely compromised if he/she did survive was perceived by nurses as "torturing" the patient. Thus, this was viewed as an ethical issue for nurses. As such, feeling forced to follow through with resuscitation in this type of situation on an ongoing basis might contribute to moral distress in nurses. Moral distress has been linked with nurses withdrawing from patients and families, nurse burnout, and turnover.31-33
LIMITATIONS
A limitation of the research is the scale of the study. The sample size was small with only 1 ICU utilized, consisting of a specific type of ICU from 1 specific type of hospital. This limits transferability of study data. Thus, further research is needed to verify study findings in other ICU settings.
A potential limitation includes the researcher's presence in the study setting distorting events and actions. However, experiences of researchers in ethnography suggest that "after a few days, the people of the subculture must focus their energy and attention on the usual events of the culture."22(p293) This was the experience of this researcher, as well. This suggests that reactive effects of a researcher's presence are likely minimal.
RECOMMENDATIONS
Based on the study findings, the following recommendations are offered:
Practice
(1) Nurses and physicians need to share responsibility for decision making and for the outcomes of decisions with families, especially at end of life.
Shared responsibility for decision making needs to be communicated to families starting with using the word "we" in relation to decision making, rather than the word "you." Providers should ask families if they would like a medical recommendation regarding decisions needing to be made and provide this if desired. Mutual sharing of information and ideas, working toward arriving at a consensus regarding the goal of care and how to achieve it, should be the aim of communication. The goal of care can include a smooth and comfortable transition to comfort care and a peaceful death.
(2) Nurses and physicians need to assess family members' needs for the privacy, information, and emotional support structured into formal care conferences.
The structure and support of an informal care conference at the bedside are very different from those of a formal care conference in a private setting. Discussions regarding negative prognostic information and family members' requests for discussing living wills and withdrawal need to be consistently structured as formal care conferences, which occur in a private setting and with the support of the chaplain and/or social worker. Time needs to be allotted for discussing the content included in these formal conferences, including the patient's history, treatments and responses, family concerns, patient wishes, and quality of life.
(3) Physicians need to be made aware that patient code status is a significant stressor for ICU nurses and strive for timely follow-up on code status issues.
This includes talking to the family to determine a decision regarding code status, communicating the code status decision to the nurse, and following all the procedural steps to process the order.
(4) Nurses should be encouraged to participate in formal care conferences with families.
A structure needs to be put in place, by collaboration between the ICU nursing manager and the ICU medical director, whereby nurse participation in formal care conferences is facilitated and not only welcomed, but also expected. In recognizing the need for consistent communication to families, communication between the physician(s) and nurse(s) should occur prior to the care conference to ensure there is agreement on the goal of care and treatment or nontreatment decisions and discuss any diverging opinions. Then, during the care conference, solicitation of the nurse's input should be structured into the format of the discussion with the family.
(5) More research needs to be conducted on the perception of DNR. Often families may feel the DNR is similar to giving up. In addition, further research is needed to discuss cultural or religious aspects that may impact decisions concerning end-of-life.
Education
1. Nurses and physicians need to receive educational content related to communication at end of life, focusing on communication with both patients/families and other providers (eg, discussing divergent views and resolving provider conflicts). This education should be integrated into the medical and nursing curricula in universities and colleges and provided as in-services. Content should include observation of expert communicators and role playing/simulations. Understanding and coping with emotions, including anger, frustration, and grief, both personal and those of patients/families, should also be included.
2. Physicians need educational content related to prognostication and communicating a poor prognosis, both in relation to timing, content, and style of communication. This education should be incorporated into medical school programs, residency programs, and fellowship programs and should include observation of expert communicators and role playing/simulations.
3. Nurses need educational content for caring for patients at end of life, preferably woven throughout multiple courses in nursing undergraduate curriculums or, if this is not feasible, as a separate course. This should be a required course and not an elective. The course should include frank small group discussions of personal fears related to death, dying, and mortality and discussion of cultural influences on death and dying, including how culture influences what can be said and what cannot be said in this regard.
Research
Repeating this study in different ICU settings would help to verify findings or identify contrasting findings. Intensive care units in a teaching hospital or a large urban level I trauma center hospital would be interesting options. Research focusing on how the "boundary" of prognostic communication is "learned" by nurses would shed light on this topic.
CONCLUSION
Nurses and physicians play complex, albeit complementary, roles in communication of prognostic information to family members. It is only by recognizing both nurse and physician communication patterns that a clearer picture of prognostic communication is revealed. Describing prognostic communication patterns of critical care nurses and physicians to families, both informal communication at the bedside and formal communication during care conferences, further illuminates this complex communication process.
Acknowledgments
The author thanks Dr Joan Liaschenko, PhD, RN, FAAN, for advising her dissertation research and the other members of her dissertation committee, Dr Ruth Lindquist, PhD, RN; Dr Susan O'Connor Von, PhD, RN; and Dr Karen Ho, PhD, at the University of Minnesota for their guidance and support of the author's work. The author acknowledges Sigma Theta Tau International, American Association of Critical Care Nursing, and Midwest Nursing Research Society for their financial support of her doctoral dissertation research, which resulted in this article.
References