Who Is Attending? End-of-Life Decision Making in the Intensive Care Unit

Abstract

Purpose

Traditional expectations of the single attending physician who manages a patient's care do not apply in today's intensive care units (ICUs). Although many physicians and other professionals have adapted to the complexity of multiple attendings, ICU patients and families often expect the traditional, single physician model, particularly at the time of end-of-life decision making (EOLDM). Our purpose was to examine the role of ICU attending physicians in different types of ICUs and the consequences of that role for clinicians, patients, and families in the context of EOLDM.

Methods

Prospective ethnographic study in a university hospital, tertiary care center. We conducted 7 months of observations including 157 interviews in each of four adult critical care units.

Results

The term “attending physician” was understood by most patients and families to signify an individual accountable person. In practice, “the attending physician” was an ICU role, filled by multiple physicians on a rotating basis or by multiple physicians simultaneously. Clinicians noted that management of EOLDM varied in relation to these multiple and shifting attending responsibilities. The attending physician role in this practice context and in the EOLDM process created confusion for families and for some clinicians about who was making patient care decisions and with whom they should confer.

Conclusions

Any intervention to improve the process of EOLDM in ICUs needs to reflect system changes that address clinician and patient/family confusion about EOLDM roles of the various attending physicians encountered in the ICU.

Introduction

This research is part of a larger ethnographic study of end-of-life decision making (EOLDM) in adult intensive care units (ICUs).¹ Ethnographic research is a type of qualitative research that has been proposed for ICU studies to “understand complex social phenomena.”² Ethnographic research is designed to address issues of culture, the understanding of which Pronovost and Vohr identified as crucial to improvement of health care.³ The goal of the research reported here was to examine the role of ICU attending physicians in different types of ICUs and the consequences for clinicians, patients, and families in the context of EOLDM.

Our research protocol stated that we would only interview patients or family members “with approval of the attending physician.” We used the phrase “the attending physician” as though it needed no definition, and we assumed that one person could be identified as the attending. This traditional understanding of the attending physician as a single accountable person is reinforced in hospital policies (e.g., certain types of orders must be signed by the attending), in hospital forms (where a space is to be filled in naming the attending), and in everyday communication in ICUs. It is also a common model in primary care, more familiar to families, to have a single personal physician or other primary care provider who manages and coordinates all aspects of a patient's care. This expectation is currently being reinforced in primary care by the new patient-centered medical home model, where one of the emphases is on selection of a single personal clinician for coordination and continuity across providers and across settings.⁴

As we determined from whom to seek permission to interview patients and families, we realized that the attending physician was rarely a single individual, rather it was a role filled by multiple physicians at different times and, sometimes, simultaneously. Frequently, it was impossible to identify a single attending physician responsible for EOLDM. In this article, we assess the role of “the attending physician” in four adult ICUs and the consequences of role complexities for clinicians, patients, and families, particularly in the context of EOLDM.

Materials and Methods

Design and setting

The larger ethnographic study was focused on EOLDM in the ICU and a description of different ICU cultures for EOLDM.¹ The study took place in four separate units within a single >700 bed academic health care center in upstate New York: a medical ICU (MICU), a surgical ICU (SICU), a cardiovascular ICU (CVICU), and a burn/trauma ICU (BTICU). The four units were studied sequentially by a team of researchers.

Three units had intensivists as medical directors (one internist, one pulmonologist, and one surgeon); the CVICU had a cardiologist as medical director. None of the units were closed (single intensivist attending mode for all admissions), although intensivists managed much of the daily care of patients, particularly in the MICU where members of teams of intensivists or neurologists were the attendings for almost all patients. For more complete information on methods, please see our previous publication on the larger study.¹ The focus for the present analyses was ICU attending physicians and EOLDM.

Participants

The study received human subjects approval from the institutional review board; all interviewed participants were informed about the study and consented to participate. Participants were health care clinicians, patients, and family members. We conducted 46 interviews with 30 physicians (7–13 per unit), 60 interviews with 48 nurses (12–18 per unit), and 13 interviews with 10 other care providers (3–5 per unit; 6 social workers, 2 ethicists, an ICU pharmacist, and a chaplain). We also conducted 34 interviews with 38 family members and 4 interviews with 4 patients. The age range for physicians was 27 to 45 (36.4±5.9), for nurses 26 to 54 (39.3±7.0), for social workers and other providers 31 to 59 (43.9±11.7), for family members 30 to 88 (65.8±12.6), and for patients 67 to 84 (75.3±6.9). See Table 1 for race, ethnicity, and gender.

Table 1.

Ethnicity, Gender, and Race

	Physicians, n=30	Nurses, n=48	Other providers, n=10	Family members, n=38	Patients, n=4
Hispanic	0	0	1	0	0
White	23	48	10	34	3
Black	0	0	0	2	0
Asian	7	0	0	2	0
American Indian	0	0	0	0	1
Male	27	5	1	16	0

Methods

Methods included participant observation and semi-structured interviews related to ICU EOLDM generally and about specific patients identified as near the EOL, and collection of relevant artifacts. An interview guide was developed based on the literature. It was modified during the course of the study based on ongoing data collection and analysis.

Participant observation focused on patients nearing the EOL, who were identified by querying nurses and physicians, and by attending daily rounds. Once we had been on units for a week or two, nurses and physicians often spontaneously told us about such patients without prompting. For the semi-structured interviews, patients and/or their family members were invited for interviews only with permission of the current attending physician. At times individual family members or clinicians were interviewed more than once if the EOLDM process was extended. At times interviews took place before or after family meetings that we also observed. Sometimes, two or three persons were interviewed together as seemed suitable, for example, a brother and sister jointly planning for their mother.

Early in the data collection process it became apparent that we needed to explore the meaning of the term the “attending physician,” as questions about that role and its meaning grew out of early participant observations and interviews. We then incorporated that focus into subsequent observations and added a question about it to interviews. This is in keeping with iterative ethnographic methods, which begin with broad observations followed by more focused ones.⁵

The original research team consisted of seven nurse data collectors with advanced education collectively in sociology, anthropology, and nursing. Their clinical backgrounds included critical care, psychiatric, cardiac, emergency, and gerontologic nursing. The analysis team for this report included six of the data collectors plus a medical anthropologist (NP) and an intensivist with expertise in ICU EOL (TP).

Trustworthiness

Four issues are traditionally involved in establishing trustworthiness of qualitative research: credibility, transferability, dependability, and confirmability. Credibility can be established via several approaches to triangulation of data-constant comparison processes, which were used by the multiple researchers involved in this research as data were collected over time and across units; multiple approaches to data collection, which were employed in this study (extended observational approaches through participant observation, semi-structured interviews, and artifactual data); and a variety of participants (data from several different groups of clinicians and families). We used researchers with varying backgrounds and perspectives in data collection and analyses.^6,7

Transferability relates to the relevance of the data beyond the narrow context in which it is produced and can be established by providing sufficient background context for the study to allow readers to judge its relevance to their own situations.⁷ The context for this research has been provided in the Methods section. Literature included in our Discussion section provides reinforcement of the transferability of these data.

Dependability is reinforced by use of overlapping methods,⁷ which have been described above.

Confirmability is established through providing sufficient detail of study methods for reader evaluation and triangulation to reduce investigator bias. Background information is provided about the expertise of the researchers involved. A description of the study methods/procedures for evaluation by readers is also provided as well as a reference to a more extended methods description in the initial manuscript published.¹ Limitations of the study with regard to this particular set of analyses are identified in the Discussion section.

Procedure

Field-note data were collected daily from observations of rounds and other unit activities for approximately 25 hours a week for 7 months on each unit for a total of >2,800 hours from 2000 through 2004. Field notes were transcribed and checked for accuracy. Data also included 157 interviews and observation of 22 family meetings (2–8/unit). Interviews and family meetings were audiotaped, transcribed, and checked for accuracy. All transcriptions were entered in to a software program for data management (ATLAS.ti Version 6.2; Scientific Software Development GmbH, Berlin, Germany).

Data analysis

In the larger study, data analysis⁸ was conducted simultaneously with data collection to develop an understanding of the unit cultures. We developed a sense of how behavior was patterned and generated a list of codes applied to meaning units in the data. Codes were combined into categories and themes⁹ related to EOLDM. In ethnographic work data are illustrated by quotations.

For this article, data contained in three codes and their eight related subcodes were used to purposefully access and analyze relevant data to develop an understanding of the role of the attending physicians in EOLDM. These were:

• 1. Decision making:

  • 1.1 Involvement of attending

  • 1.2 Involvement of outside physician

• 2. Influence:

  • 2.1 Clinician

  • 2.2 Outside professional

• 3. Role:

  • 3.1 Attending

  • 3.2 Attending/control of decisions

  • 3.3 Intensivist

  • 3.4 Primary care clinician

Results

ICU attending physicians

The term “attending” could be used simply as Pronovost and Vohr described it, to refer to “senior-ranking physicians,” as opposed to, for example, residents.^3(p22) However, an intensivist provided a window into the definition of the attending physician most likely to be problematic for staff and particularly for families in the ICU setting: “What attending means is that you're the [physician] ultimately responsible for the person.” Although this appears conceptually simple, it was challenging to identify who was in the attending physician role for a given ICU patient at a particular time.

There were differences by ICU and by type of patient concerning who might serve as an attending physician. The medical intensivists, a group with common backgrounds in pulmonary/critical care, were attendings for the majority of patients admitted to the MICU. However, neurology patients were attended by neurologists, and a very few patients were attended by their primary care clinicians. In interviews and observations, the term “the attending [physician]” was used at times for the intensivist, for the community physician, or for a subspecialist who had provided outpatient care for a patient's chronic illness, demonstrating the complexity and the potential for confusion related to the attending physician role in the MICU. Even physicians were confused. A neurologist commented:

Now she [patient] was not on the cardiac team, but they transferred her to the medical team. So, a whole new set of interns and residents came on board, and they were like, “What's going on, who's making decisions?”…In fact, come to find out that ultimately she was transferred to my service.

On one of the surgical units, the primary intensivist was an internist. The roles of medical director/intensivist and the attending physician were closely aligned, but the attending role was always shared with the operating surgeon, who remained the hospital-designated attending physician for almost all surgical service patients. The attending surgeon was described by an intensivist as the person who performed the surgery, with a sense of responsibility for outcomes. Major decisions about treatment were the responsibility of the attending surgeons. An intensivist working with surgical patients described the attending relationship with the surgeons this way in response to a question about who was the patient's attending:

The way it works for patients in this unit is that it's joint care between the surgeons and myself.…I try to maintain that relationship and work with them and not have it come down to who's in charge. On paper they are the attending…but usually…we can work together to take care of patients.

When asked who was the attending, SICU nurses generally named the intensivist medical director, whom they saw making decisions, as the patients' attending. However, if the questioner was then asked, “But what about the surgeon?” they usually said something such as, “Oh yes, that is the attending surgeon,” distinct from the attending intensivist.

The BTICU attending role resembled that in the SICU except that the intensivists were surgeons or anesthesiologists. On the CVICU, there were different attending groups for medical, precardiac transplant, and postsurgical patients.

Multiple attending physicians

Multiple physicians could serve in the attending role simultaneously. An intensivist, asked to identify “the attending” referred to “the admitting attending, the operating attending, and the consulting attendings.” A nurse manager was asked if the “attending of record for the patient, making the decisions, is different than the ICU attending” and responded, “It could be both, but we always have to go back to the attending physician.” A SICU nurse reflected, “It seems funny to have two attendings, and they are both talking to the family, and they're not talking to each other.”

Rotations

Regardless of who qualified to be in the attending physician role, the role generally was temporary and was held sequentially by different physicians over time. Services (e.g., neurology) developed schedules to manage rotating attending coverage. Documentation about who was currently filling the attending physician role was challenging to find. Data collectors repeatedly noted that they had difficulty discovering who was the attending. Some clinicians also expressed confusion, including a nurse who complained that she could not identify the attending when she was transferring a patient out of the ICU.

In field notes a medical resident noted some of the decisional consequences of the rotation schedule: “He says this [complexity of changing attending physicians] is really difficult with so many attendings. Even in the time he has been on the unit (2.5 weeks) there have been different opinions.” On the other hand, rotations were seen as necessary by physicians. A neurologist commented that being on-service “really is draining after 2 weeks.”

The ICU attending role in EOLDM

The ICU attending role around EOLDM varied by type of unit and patient. Medical intensivists generally were comfortable with discussion of limitation of treatment. The nurses agreed that intensivists should take the lead with families; as one nurse said, “physicians get the formal ball rolling.” Two MICU nurses also commented on how much better than the community physicians or subspecialists (other than neurologists) the medical intensivists were at managing EOL discussions.

A medical intensivist articulated some of the differences between medical and surgical physicians in dealing with EOL discussions:

[Surgeons] hang onto the number of times that they've been hopeful and the outcome has been much better than expected. So I think it's in large ways looking at a different side of the same coin. They say, “Well, true, the vast majority of people are going to not do well but there is a small percentage that will surprise us”…and that's one way to present it to [a patient's husband]. Another way [the intensivist's way] to present it to him is that “there are very, very few people that do well, and her age and her injury and how she's doing right now, it's unlikely that she'll get back to a very good quality of life.”

A medical intensivist working with surgical patients noted that the presence of attending surgeons made his work different from working in the MICU. He believed surgeons' sense of responsibility for the outcome after surgery made it more difficult for them to deal with EOL issues. “Most of the time the internist [or intensivist] doesn't feel…palpably culpable for the turnout, whereas the surgeon does.” A medical intensivist indicated that if the family of a surgical patient came to him to request limitation of treatment, he always called the operating surgeon to give the surgeon “a chance to talk them out of it.” An anesthesia intensivist attending in the SICU said that the surgeons were “reluctant to give up” and “didn't want to give responsibility to others.” He expressed frustration with times when a family wanted a patient to have a do not resuscitate (DNR) order but believed he could not institute the order because only the surgeon could do so. EOL decisions could become more problematic when the attending surgeon was not available. In two of our research cases EOLDM requested by family was delayed because the operating surgeon was out of town, and the “covering” surgeon did not wish to make withdrawal decisions.

There was also variation among surgeons in their views toward EOLDM involving their patients. One surgeon clarified his involvement in EOLDM as differing depending on whether the patient had elective surgery, in which case they were expected to do well and rarely needed EOLDM, and emergency patients, where “there's not quite the emotional bond between family and patient and physician.” In the latter case EOLDM was easier for him. Another surgeon also commented that the EOLDM process “is very dependent on the attending [intensivist] who's on for that week…There are subtle shifts in the medical…psychosocial management [of the patients] depending on the attending.” Similar comments were made by nurse practitioners (NPs), social workers, and a pharmacist. On the other hand, an intensivist who worked with surgeons indicated that he adjusted his care and interactions according to which ICU-based surgeon was currently in the attending role: “See, the unwritten rule is that…we know that Dr. So and So wants us to do everything. We also know that Dr. B doesn't want everything.”

The staff nurses in the surgical units thought that an intensivist medical director should initiate discussions about EOLDM. A surgeon told us he believed the intensivists did not encourage or welcome surgeons' inclusion in EOL discussions; however, this may have been more difficult because surgeons were less often present. A social worker said, “The surgeons are just a lot less available to families…being in the OR all day…The family either has to be here very early in the morning or be willing to come in at 5 and sit and wait until they appear…That communication piece…is one of the biggest problems with any kind of decision making.”

On the BTICU the attending role alternated among surgeons and anesthesiologist intensivists for the surgical patients, and, as in the SICU, intensivists worked with the operating surgeon. At EOL the operating surgeon generally took precedence in decision making. On the CVICU the role of attending at EOLDM for medical cardiac patients resembled that in the MICU, and for surgical cardiac patients it resembled that in the SICU. A single nursing staff had to adjust to different medical subcultures in supporting families around EOL situations.

The ICU attending role and timing of EOLDM

In cases where a community physician (co-)attended and a patient with a poor prognosis was failing, the intensivists called to ask if the community physician wanted the intensivist to discuss EOLDM with the patient or family. They usually agreed. According to one intensivist, “The [community] attendings aren't here physically…It's much easier for us…Probably 90% of the time we get the [DNR] forms filled out even if we're not attending.”

Attending physicians initiated discussions about limitation of treatment at different times in different types of ICUs. In the MICU, discussion of EOLDM by the attending physicians generally began early in a patient's ICU admission and included multiple parties, including patients, families, and other care clinicians who were not physicians. Family meetings to address goals of care, prognosis, and possible decisions about limiting treatment were arranged when members of the healthcare team began to express concern about the potential for a poor outcome.

Discussions about limitation of treatment for surgical patients generally took place later in a patient's illness trajectory, not uncommonly in a crisis atmosphere. A number of times during our study, families were called emergently to make a withdrawal of treatment decision when the patient was likely to die within hours, even with continued aggressive treatment. The major exceptions to this pattern occurred when families initiated the topic of limiting treatment. In such cases, the intensivist would notify the surgeon, and discussions would begin. Surgeons rarely initiated treatment withdrawal discussions.

The ICU attending role in EOLDM and nursing staff

Interactions between physicians and nursing staff in EOLDM varied by type of unit and type of patient. For medical patients, nurses felt empowered to ask questions on rounds, about “ultimate goals of treatment” or about prognosis. For surgical patients, staff nurses most commonly went to a NP or care coordinator to have them raise the issue. One NP said: “We're kind of a bridge to the physician…because the nurses definitely don't go to the physicians and say we have end of life issues…Usually a nurse will go to one of the nurse leaders, or one of the NPs to express their concern.”

Nursing staff members were aware of variations in particular physicians' attitudes and behaviors about EOLDM. Nurses on units told of biding their time for a rotation to occur so that they could bring up the topic with a different attending physician who was more open to considering alternatives to continued aggressive care.

The consequences of the ICU attending role for family involvement in EOLDM

Most families' knowledge of physicians was based on relationships with community physicians. They were rarely knowledgeable about the network of complex relationships surrounding the attending physician role in the hospital and ICU settings. Family members often were confused about who was in the attending role and frustrated by trying to participate effectively in EOLDM with multiple “attendings.” The husband of a patient said, “At home you have a primary, you see one doctor. When you come up here, you get a group, and it is kind of disconcerting sometimes when you first come in to know what is going on.”

Figuring out who the attending[s] was [were] from the multiple physicians involved in care, and sequential filling of the attending role added to uncertainty among families about who was making health care decisions for their loved ones, especially EOL decisions. When the interviewer asked, “Did they [physicians] tell you who they are?” the husband of a patient replied, “Sometimes they do, sometimes they don't.” Some families attempted to figure out on their own who was in charge. A family member said, “I look them over and see who was the oldest one, and then I can figure out who was boss [the attending].”

Anger over lack of continuity was visible to data collectors, but seldom noted by clinicians. From field notes about a son: “The problem right now is how his father is getting bounced around…from one doctor to another. This is hard on the family. When the family asks how the patient is, the care clinician is not familiar with [him].”

A nurse leader noted that the continuity of care issue was “Huge. Huge. And it's very hard for families. You don't really get to know anybody…You have a doctor for one week, and you're just starting to think, ‘OK, this guy's got a grip on things. I can trust him.’ And somebody else comes in and he's gone.” A patient's husband complained about the delay in meetings to inform him about his wife's condition because of rotations, “I want to start to get information as often as is needed. I don't want to wait another 2 weeks to have another meeting [because attendings keep changing].”

Discussion

There have been many studies of attending physicians and EOLDM. There are studies demonstrating their power to make decisions guided by influences other than patient or family wishes.^10–12 A case has been made that intensivists are better qualified than other physicians to manage EOLDM.¹³ Variations in physician practice have been identified.^10,14,15 Our study differed from these in focusing on the complex issues associated with the attending role and the impact of the complexity of the role on other clinicians, patients, and families especially during EOLDM. We had observations of and interviews with many types of physicians and their involvement in EOLDM.

Over a quarter of a century ago, Orsher^16(p52) identified the classical responsibilities of the attending physician: “Medicolegally, academically, and socially he is responsible for the kind and quality of care given to patients.” However, MacIntyre^17(p58) expressed concern with the prevailing societal view of the attending as a single physician: “The physician who actually examines and prescribes for the patient…is very likely to change from occasion to occasion…The notion of the attending physician as the patient's freely chosen physician at whose instance and under whose direction other doctors operate is indeed a piece of mythology.” One could argue that this is especially the case in acute care hospitals, at a time when there is a movement to revitalize such a role in primary care.

The multiplicity of physicians who may be involved in care has been critiqued as leading to an “erosion of the sacred trust” between physicians and patients¹⁸ and confusion about who is in charge of patient care.¹⁹ In a recent study of acute care medical patients asked to name “the inpatient physician in charge of their care,” 75% of them could not identify anyone, indicating that they did not know whom to contact for explanations or requests.²⁰ Recently in a study of the last 3 days of life of patients who died in a large teaching hospital, most providers caring for them reported knowing their patients for <24 hours.²¹

The findings that operating surgeons often retained responsibility for EOLDM for patients on whom they had operated echoed concerns raised by Buchman et al.²² about surgeons' goals in dealing with their patients who were likely to die. Surgeons' difficulties acquiescing to the impending deaths of their patients, and confusion about which attending (surgeon versus intensivist) would assume responsibility for EOLDM led to delayed EOL discussions for surgical patients compared with medical patients. Guidelines that are being developed for identifying patients who would benefit from palliative care in surgical ICUs are being developed.²³

The results of this study give some insight into why ICU staff on occasion, but especially patients and families, cannot identify from whom they should seek explanations and to whom they should express concerns in ICU situations. The attending is not an individual but a role that is held by different individuals sequentially over time or by multiple physicians at the same time. Identifying who is acting in the ICU attending role for an ICU patient at a particular time is rarely simple, frequently frustrates even experienced nursing and physician staff, and complicates EOLDM in particular.

The complexities of physician-physician attending relationships, and thus physician-patient relationships, were opaque to family members. Their perception of the attending as an individual physician was not congruent with how attending responsibilities are operationalized in the contemporary ICU. Medical professionals acknowledged that the way the role of attending was implemented in ICUs created multiple “hand-off” issues that affected continuity of care, but they underestimated the deleterious effect of attending practices on patient and family-centered care giving.

With multiple “attending” physicians involved in care at EOL and different rotation schedules by specialty, families experienced a constant turnover of physicians to approach about EOLDM. This was especially distressing to families trying to come to terms with the potential loss of a family member, who were being asked to make difficult decisions in a complex and unfamiliar situation, as has been noted by other researchers.^24,25 Families have identified helpful and unhelpful behavior by physicians.¹² Changing physicians requires families to learn to work with multiple physicians and their varied helpful and unhelpful behaviors.

These data may be considered limited because of their age and because the research took place in a single institution. However, problems with rotations, which physicians identified as necessary, and continuity are not unique to the ICUs studied or to the study hospital. The data are unique, in that they were collected prospectively and include a wide range of observations and interviews of processes unfolding over time with multiple participants on multiple units. They focus on human interactions, not technologies, and current literature suggests they are not dated. Focusing on a single institution allowed collection of data while holding constant the institutional culture and revealed many cultural differences among units in management of EOLDM.

Conclusions

System changes to address the multiple problems related to hand-offs, coordination, and continuity in this aspect of care need consideration. Solutions lie in seeking creative changes in hospital systems of care to address the communication, coordination, and continuity problems generated by the complexities of attending roles in acute care, and especially around EOLDM in ICUs, including evaluation of the effects of these changes on families' experience of care.

Closed ICUs have been proposed to improve EOLDM. Although moving to closed units, where a single team of intensivists attend, might be assumed to diminish some of the confusion observed, in the MICU, with only two teams acting as attendings, there was still confusion. The use of rotations means a better way of communicating with patients and families is needed. Some study has been made of the difficulty associated with altering patterns of intensivist staffing^27,28 to address issues such as care continuity, but the study of outcomes has been limited to measures such as hospital and ICU mortality, length of stay, intensivist burnout, job distress, and work-home life imbalance, but the work has not focused on the consequences of these changes for families.

Since this study was initiated, palliative care teams have been added as a consulting service in many hospitals. Palliative care consultation may offer improved communication, coordination, and continuity in working with families and patients in EOLDM.^29,30 However, palliative care represents another group of consultants whose physician members may shift into the attending role in some circumstances and who also may rotate the role, adding to an already complex, confusing situation for patients and families. The effectiveness of their involvement will depend, in part, on their ability to negotiate the attending physician subcultures in EOLDM as well as deal with their own communication, coordination, and continuity issues on behalf of patients and families. Technical guidelines for ICU palliative care have been developed and should be of assistance.³¹

Changes require that clinicians “examine how culture affects the system and structures within which we do our work…[as] culture influences…how we treat our patients.”^3(p39) Recently, several authors have proposed that one cultural change, increased involvement of ICU nurses in EOLDM and palliative care, would improve the process.^32,33,34

Whatever systems of care are or whatever changes are made, improving families' understanding of how various physicians relate to each other and to families around patient care is an important part of the work that needs to be done. Health care clinicians need to be sensitive to these issues and aware of limitations in family members' knowledge of a complex system with multiple interrelationships.

Better, more up-to-date documentation, perhaps in the electronic medical record, and communication of that information to families would be helpful. A beginning step is to achieve clarity and accountability among the professional care givers, and then to make sure families know the appropriate person[s] to communicate with by posting the current patient attending and nurse caring for the patient where families can see it, and to update this information as as changes are made.

Acknowledgments

The study was conducted at the University of Rochester, Strong Memorial Hospital, Rochester, New York. The research team thanks the patients, families, and staff on the units, who generously participated in this study.

A previous version was presented at Royal College of Nursing Research Conference, Harrogate, England, 2011. This work was supported by NIH grants, RO1NR04940 and R15NR012147.

Author Disclosure Statement

No competing financial interests exist.