Nurse-physician perspectives on the care of dying patients in intensive care units: Collaboration, moral distress, and ethical climate*

Transcription

1 Nurse-physician perspectives on the care of dying patients in intensive care units: Collaboration, moral distress, and ethical climate* Ann B. Hamric, PhD, RN, FAAN; Leslie J. Blackhall, MD, MTS Objective: To explore registered nurses and attending physicians perspectives on caring for dying patients in intensive care units (ICUs), with particular attention to the relationships among moral distress, ethical climate, physician/nurse collaboration, and satisfaction with quality of care. Design: Descriptive pilot study using a survey design. Setting: Fourteen ICUs in two institutions in different regions of Virginia. Subjects: Twenty-nine attending physicians who admitted patients to the ICUs and 196 registered nurses engaged in direct patient care. Interventions: Survey questionnaire. Measurements and Main Results: At the first site, registered nurses reported lower collaboration (p <.001), higher moral distress (p <.001), a more negative ethical environment (p <.001), and less satisfaction with quality of care (p.005) than did attending physicians. The highest moral distress situations for both registered nurses and physicians involved those situations in which caregivers felt pressured to continue unwarranted aggressive treatment. Nurses perceived distressing situations occurring more frequently than did physicians. At the second site, 45% of the registered nurses surveyed reported having left or considered leaving a position because of moral distress. For physicians, collaboration related to satisfaction with quality of care (p <.001) and ethical environment (p.004); for nurses, collaboration was related to satisfaction (p <.001) and ethical climate (p <.001) at both sites and negatively related to moral distress at site 2 (p.05). Overall, registered nurses with higher moral distress scores had lower satisfaction with quality of care (p <.001), lower perception of ethical environment (p <.001), and lower perception of collaboration (p <.001). Conclusions: Registered nurses experienced more moral distress and lower collaboration than physicians, they perceived their ethical environment as more negative, and they were less satisfied with the quality of care provided on their units than were physicians. Provider assessments of quality of care were strongly related to perception of collaboration. Improving the ethical climate in ICUs through explicit discussions of moral distress, recognition of differences in nurse/physician values, and improving collaboration may mitigate frustration arising from differences in perspective. (Crit Care Med 2007; 35: ) KEY WORDS: end-of-life care; intensive care units; collaboration; ethical climate; moral distress; nurse-physician relationships Intensive care units (ICUs) are places where the sickest of patients receive the most technologically sophisticated care that medicine can offer. Because these units care for people at the brink of death, they are also places where patients, families, and the health professionals caring for them struggle with decisions about the appropriateness of aggressive care (1). Despite *See also p From the University of Virginia Schools of Nursing and Medicine, Charlottesville, VA. Supported, in part, by intramural funds from the University of Virginia Schools of Nursing and Medicine and the UVA School of Nursing Alumni Association, Charlottesville, VA. The authors have not disclosed any potential conflicts of interest. Copyright 2007 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: /01.CCM D advances in critical care, a relatively high proportion of patients admitted to ICUs will die before leaving the hospital. In certain subgroups, such as patients with cirrhosis (2), those requiring prolonged mechanical ventilation (3), and certain cancer patients (4, 5), the mortality rate exceeds 50%. To look at it from another direction, one recent study suggested that approximately 20% of people in the United States die during or shortly after an ICU stay (6). Thus, although ICUs are dedicated to saving lives, they are also places where a significant proportion of patients will die. In an ideal world, terminally ill patients would not undergo aggressive, futile interventions at the end of their lives. Unfortunately, even with the best prognostic models, we are unable to accurately predict whether individual patients will survive their stay in an ICU (7). For this reason, most commentators agree that improving the quality of care at the end of life is an important part of ICU care (8 11). In the ICU setting this has many components, including improving communication with dying patients; facilitating decisions concerning the use of life support; reducing pain, dyspnea, anxiety, and other symptoms; and supporting families. Despite the increasing attention paid to this topic, evidence both from Europe and the United States suggests that we could do better. End-of-life practices are variable and may depend more on local physicians beliefs and practice patterns than on the characteristics of their patients (12, 13). Patients symptoms are often not well managed (14 19), and families needs for information and support are incompletely met (20 23). Many position statements on improving end-of-life care (EOLC) in ICUs have affirmed the multidisciplinary nature of this 422 Crit Care Med 2007 Vol. 35, No. 2

2 complex and sensitive task (8 11, 24). Although these statements acknowledge that improving nurse/physician collaboration is crucial, they offer few concrete suggestions. More detailed discussions of collaboration may be found in the nursing literature (25 27). Nurse/physician collaboration has been shown to improve a variety of ICU patient outcomes (28 30) and may be particularly important in EOLC for several reasons. For family members wrestling with concerns about their loved one, the nurse is the most visible resource for information and education. Since nurses spend more time at the bedside than physicians, they may be more aware of patients symptom experiences and are often privy to invaluable information about the concerns and values of patients and their families. Finally, although physicians are responsible for prescribing medications for symptom control, writing do-not-resuscitate orders, and the like, nurses are primarily responsible for implementing any EOLC plan. Unfortunately, several studies have noted conflict between nurses and physicians concerning EOLC in ICUs (31 33), including concerns about overtreatment, communication, and quality of life (34 37). The nursing literature has documented distress and burnout among ICU nurses resulting from these concerns (38, 39). The purpose of this pilot project was to explore the perspectives and experiences of physicians (MDs) and registered nurses (RNs) working together with dying patients in ICUs and in particular to examine the relationships between moral distress, ethical climate, RN/MD collaboration, and satisfaction with quality of care. The influence of ethics variables, particularly perceptions of moral distress and ethical climate, is only beginning to be studied in the context of EOLC. Moral distress differs from a moral dilemma, a situation in which the practitioner is conflicted about the correct ethical choice. Moral distress occurs when the practitioner feels certain of the ethical course of action but is constrained from taking that action. This phenomenon has been identified and studied in the nursing literature (38, 40 42) but has only recently been described in medical literature (43, 44). Corley (42) hypothesized that increased moral distress would be related to lower nurse satisfaction and decreased quality of care, but these relationships have not been studied. In addition, although it is widely recognized that organizational environments affect caregiving practices, there has been little study of the ethical dimensions of these environments (45 47). Ethical climate has been defined as the organizational practices and conditions that promote discussion and resolution of decisions with ethical content (48). It may be a salient variable in EOLC, given the complex ethical decisions that often attend such care. METHODS Study Design. We developed and pilot tested a survey to study RN/MD perspectives on EOLC in ICUs, using a combination of existing instruments adapted for this study and investigator-designed questions. The study was approved by the authors institutional review board and the review boards at the two study sites. Informed consent was implied by respondents return of the questionnaire. Data on variables associated with RN/MD collaboration in end-of-life ICU care were gathered from RNs and attending MDs in adult ICUs in two institutions. Survey development began with two focus groups, one with eight ICU RNs and one with 12 medical residents and fellows at the authors institution. These discussions helped identify salient factors affecting collaboration from each discipline s perspective. Study variables were selected based on focus group comments, their importance in prior literature, and our clinical observations. In response to site 1 participant feedback, changes were made for the site 2 survey, as noted subsequently. In this article, we present data on moral distress, characteristics of the ethical environment/climate in the ICU, communication with dying patients, satisfaction with quality of care, collaboration, and selected clinician demographic and setting characteristics. Measures. One challenge in designing the survey was identifying instruments appropriate for both MD and RN participants. Most instruments have been developed from the perspective of one discipline, and some had items that could not be adapted for use with another discipline. Furthermore, most instruments reviewed needed to be shortened to minimize overall survey length. In all cases, scale adjustments were made to focus survey items on EOLC considerations and to use items that were appropriate for both RN and MD respondents. The following instruments were adapted to measure the variables of interest in this study. Moral Distress. Corley s Moral Distress Scale (49) was adapted for use with both RN and MD participants. The original 38-item scale was reduced to 19 items. Scale items describe situations that could engender moral distress; subjects rate both the frequency and the level of disturbance (intensity) that the situation causes on a scale from 0 (never occurred/not disturbing) to 4 (occurred very frequently/greatly disturbing). To measure current level of moral distress, we multiplied the frequency and intensity scores for each item. Each item product of frequency and intensity ranged from 0 to 16. These products were added to obtain a composite score. Using this scoring scheme allowed all items marked as never experienced or not disturbing to be eliminated from the score, reflecting actual moral distress. The Cronbach s internal consistency reliability of the 19-item scale (using the product score for each item) was.83 (MDs.81; RNs.85). Ethical Environment/Climate. In the first site sample, 11 items from McDaniel s (50) 20-item Ethics Environment Questionnaire (EEQ) were used to measure perceptions of support in the work environment for addressing ethical dilemmas. The Cronbach s of the shortened EEQ in this pilot was.89. For reasons of cost, this scale could not be used with the second sample. Olson s (48) Hospital Ethical Climate Survey (HECS) addresses this same concept, although Olson used the term climate rather than environment. In addition, EEQ items focus on the overall organization, whereas the HECS blends items relating to the specific unit with organizational items. The HECS was shortened from 26 to 15 items to measure how RNs and MDs perceived the ethical climate of their work setting; items analogous to the EEQ were used as possible. The final HECS scale in this study had a Cronbach s of.88. In both scales, higher scores denoted more positive perception of ethical environment/climate. Since the EEQ and HECS do not contain the same questions, we only provide within-site comparison rather than across-site comparison. End-of-Life Communication. Because of strongly expressed opinions by some RN and MD participants in our focus groups, we asked several specific questions about end-of-life communication. These items were not treated as a scale but were analyzed individually. Satisfaction with Quality of Care. This variable was measured by four items that focused on respondents assessment of the quality of care on their unit and their ability to give excellent medical and nursing care. Items were, Overall, I am satisfied with my ability to give excellent medical/nursing care where I work, Nursing care on this unit is generally excellent, Medical care on this unit is generally excellent, and, I am satisfied with my ability to influence the quality of patient care in my practice area. The Cronbach s for these items was.86. Collaboration. After literature review, the instrument developed by Hojat et al. (51) was selected as most amenable for use with both RNs and MDs. This instrument was originally tested with medical and nursing students, so some items were not appropriate for practicing professionals. Seven items from this scale were used in this study, and two items focusing on ethical decision making and MDs valuing of nursing judgments were added. Items were worded to make them specific to the respondent s unit. Because RN/MD communi- Crit Care Med 2007 Vol. 35, No

3 cation is central to collaboration, four communication items were added, three from Shortell et al. (52) and one written by the investigators to focus on communication related to dying patients. The resulting 13 item measure had a Cronbach s of.89. Statistical Analysis. SPSS version 11.0 was used to analyze the data. Descriptive statistics were calculated, and differences between sites and between MDs and RNs were analyzed with one-way analyses of variance. Because all the measures are based on ordinal scales and some measures were skewed, we report bivariate associations by Spearman s rank-order correlation coefficients. Table 1. Sample demographics MDs Site 1 RNs Site 1 RNs Site 2 (n 90) Age, yrs, mean (SD) 45 (8.14) 38 (8.46) 39 (8.32) Clinical experience, yrs, mean (SD) 19.2 (8.52) 11.2 (7.79) 12.5 (8.01) Years in current position, mean (SD) 7.4 (7.64) 7.3 (6.26) 7.2 (6.27) Subspecialty board certified (MD only), % 86 RN education, % Diploma 9 26 AD BSN MSN 0 5 RESULTS Settings and Sample. The first survey site was a 631-bed community hospital in rural southwest Virginia. The institution had seven ICUs, including cardiac/ vascular surgery, general surgery, neurotrauma, general medicine, coronary care, neonatal, and pediatric ICUs. All RNs in direct patient care (n 280; 56% responded) and attending MDs, both onsite (n 60; 57% responded) and in the community (n 53; 15% responded) who admitted patients to the ICUs, were surveyed; 198 surveys were returned for a 50.4% return rate (this number included staff from neonatal and pediatric areas; results not reported here). The final adult ICU sample included 106 RNs and 29 MDs. Factors that improved the response rate at this site included our ability to pay $25 for each returned questionnaire, an excellent site champion who encouraged participation and coordinated data collection, and strong administrative support for the study. This institution had no palliative care service. The second site was a 481-bed university-affiliated hospital in urban eastern Virginia. Seven adult specialty ICUs were surveyed, including coronary care, general surgical, burn/trauma, neuroscience, cardiovascular/thoracic, pulmonary/vascular surgical, and vascular ICUs. Ninety-four of 280 surveys were returned for a 33% response rate. Data collection was complicated at this site by multiple factors; for example, we were unable to pay participants for completing the survey. Our site champion did not have dedicated time set aside for this project; administrative interest in the study was also less than at site 1. Only four of 30 MDs participated, making it impossible to compare MDs and RNs within this institution; these MDs were therefore dropped from further analyses. There was an active palliative care service at this second site. Both institutions have medical residency programs. Table 1 compares the demographics of the MD and RN samples. were primarily white (90%) and male (86%); the RNs at both sites were primarily white and female (91%). Nursing respondents differed in ethnic background (17% of the RNs at site 2 were Asian or African- American compared with 2% at site 1; the remainder at both sites were white) and education (49% held BSN or MSN degrees at site 2 compared with 29% at site 1). Approximately 87% of participants at both sites were Christian. Moral Distress. MDs and RNs significantly differed in their perception of moral distress, with RNs experiencing more moral distress than MDs (p.001). (Missing data resulted in lower sample sizes for the composite scale scores.) Although site 1 RNs had higher scores than site 2 RNs, the difference was not statistically significant (p.125). As seen in Table 2, the range of scores was quite wide for both groups. In looking at the responses to the moral distress scale items, we noted some interesting patterns emerging. In general, the same situations evoked moral distress in both MDs and RNs. The most distressing situations for both groups involved caregivers feeling pressured to continue aggressive treatment in situations where they did not think such treatment was warranted. There was no significant difference between site 1 RNs and MDs in their overall rating of the level of disturbance the situations would cause them (RN mean [SD 9.56] vs. MD mean [SD 11.06], t 1.51, p.139). However, RNs perceived these morally distressing situations as occurring more frequently than did MDs (RN mean [SD 9.56] vs. MD mean [SD 6.99], t 5.23, p.001), and this difference accounted for most of the difference in scores between groups. Table 2. Moral distress a Group Mean Score (SD) Range (n 89) Site 2 RNs (n 79) (33.74) b (33.22) b (22.59) b a Scale range 0 304; b p.001 between RNs and MDs; not significant between RNs. Table 3 lists the highest frequency times intensity scores and relative rankings for RNs compared with MDs. In all situations, the RNs at the first site had the highest moral distress scores. MDs had significantly lower scores on all items except for placing a gastrostomy tube in a severely demented elderly person for this situation the MD mean was higher than the nurse means at either site, and MDs ranked this the fifth most distressing situation. One of the largest differences between MDs and RNs was the situation, Let medical students perform painful procedures on patients solely to increase their skill. MDs did not see this as morally distressing (mean score.98), but RNs did (mean score at site 1 was 5.26; at site 2, it was 4.21). The last question on the Moral Distress Scale asks, Have you ever left or considered quitting a clinical position because of discomfort with the way patient care was handled in your institution? At site 1, we reworded this question to refer only to EOLC: Have you ever left or considered quitting a clinical position because of discomfort with the way end-oflife care was handled in your institution? With this narrow wording, 1% of the RNs (1 of 104) said they had left a position. However, 23% had considered leaving their institution. When this question was 424 Crit Care Med 2007 Vol. 35, No. 2

4 Table 3. Top six situations of moral distress a Situation Follow the family s wishes to continue life support even though not in patient s best interest Initiate lifesaving actions that only prolong death Continue to participate in care when no one will make the decision to pull the plug Order/follow orders for aggressive treatments that are unnecessary for terminally ill patients Work with physicians/nurses who are not as competent as care requires Assist a physician (or another physician) who in my opinion is providing incompetent care Put a gastrostomy tube in a severely demented elderly person who is a no code a Range for each item Table 4. Nurse and physician responses about end-of-life communication site 1 Site1RN(n 106) Site 1 MD Have you urged physicians to get a patient/family to agree to a DNR order so that futile therapy could be withdrawn? changed in the site 2 survey to the original wording, the response was startling. Even though this group of RNs had somewhat lower overall distress than their site 1 colleagues, 17% of the site 2 RNs had Have you been urged by nurses to get a patient/family to agree to a DNR order so that futile therapy could be withdrawn? Do physicians in your unit withhold information about diagnosis and prognosis from patients and their families? Have you withheld information about diagnosis/prognosis from your patients or their families? Have you been asked by a physician not to discuss information about treatment options? Have you asked a nurse not to discuss information about treatment options with a patient? 12 3 Have you experienced frustration with the way physicians communicate with you about your patient s end of life care? Site 2 RNs (n 88) Has the nursing staff on your unit voiced frustration with the way you communicate with them about your patient s end of life care? 75 0 Are you satisfied with the way physicians Are you satisfied with the way nurses communicate with their terminally ill patients? communicate with their terminally ill patients? DNR, do-not-resuscitate. Values are percent reporting occasional/frequent occurrence. Crit Care Med 2007 Vol. 35, No. 2 Mean (SD) Rank Mean (SD) Rank Mean (SD) Rank (4.70) (4.39) (3.48) (4.83) (4.61) (3.85) (4.84) (4.96) (3.67) (4.72) (4.67) (2.18) (5.03) (3.68) (2.83) (4.30) (3.82) (2.62) (3.97) (3.58) (3.29) 5 left a position, and 28% had considered leaving, representing almost half (45%) of the respondents. Overall, RNs with higher moral distress scores were more likely to have left or considered leaving a Table 5. Ethical environment scores, site 1 a Group Mean Score (SD) Range 30.8 (6.17) b (4.49) b a Scale range 11 44; b p.001. Table 6. Satisfaction with quality of care a Group Mean Score (SD) Range Site 2 RNs (n 89) (2.5) b,c (1.88) c (2.10) b 9 16 a Scale range 4 16; b p.005; c p.001. position than RNs with lower scores. These findings together with the significant relationship between moral distress and satisfaction (discussed subsequently) indicate that the experience of moral distress may be a powerful factor in RN turnover. By comparison, only one MD (3%) had considered leaving a position, and none had left. End-of-Life Communication. Additional responses about communication with dying patients revealed interesting differences (Table 4). Whereas RNs and MDs at the first site similarly reported RN pressure to obtain do-not-resuscitate orders, there were significant disagreements in other areas of end-of-life communication. Notably, withholding diagnostic/prognostic information, nurse frustration with physician communication, and satisfaction with patient/ physician communication were areas of disagreement between RNs and MDs. Ethical Environment. As noted, the first survey measured ethical environment using 11 items from the EEQ (50). The second survey measured ethical climate using 15 items from the HECS (48). RNs rated the ethical environment as worse than did their MD colleagues. Table 5 displays the ethical environment scores of MDs and RNs at the first site, where the difference was significant. Satisfaction. Differences were also seen among the groups on satisfaction with care quality (Table 6). RNs at the first site rated their satisfaction with quality of care lower than did their MD colleagues (p.005); 425

5 Table 7. Collaboration a Group Mean Score (SD) Range Site 2 RNs (n 87) (6.86) b (5.28) b (4.92) b a Scale range 13 52; b p.001 for Site 1 RNs compared with other two groups. they were also less satisfied than RNs at the second site (p.001). Collaboration. Overall, RNs rated collaboration within their teams less favorably than did MDs (Table 7). The RNs at the first site rated their collaboration significantly lower (p.001) than did the second site RNs or site 1 MDs, and their range of scores was greater. Correlations Among Variables. Different patterns were seen in the relationships among variables, as shown in Table 8. For site 1 MDs, collaboration was correlated with satisfaction with quality of care and ethical environment. Some of the lack of statistical significance of other correlations may be due to the lack of power that occurs with a small sample size; for example, there was a negative relationship between satisfaction and moral distress (r.36), but this was not statistically significant. Although the RNs at both sites showed similar patterns in the relationships between collaboration, satisfaction, ethical climate/environment, and moral distress, there were some differences (Table 8). Moral distress was negatively associated with ethical environment for RNs at the first site, but this was not significant for RNs at the second site. Comparisons of RNs With High and Low Moral Distress. Because there was wide variability in the RNs moral distress scores, comparisons were made between RNs with high moral distress (top 33% of scores, ) and those with low moral distress (bottom 33% of scores, ) on key variables of interest using RN data from both sites. RNs with high moral distress had significantly lower satisfaction scores (F 2, , p.001), lower ethical climate/environment scores (F 2, , p.001), and lower collaboration scores (F 2, , p.008) than did RNs with lower moral distress. There were insufficient numbers Table 8. Correlations Variables Satisfaction Moral Distress Ethical Environment Collaboration.80 (p.001).01 NS.51 (p.004) Satisfaction NS.57 (p.001) Moral distress NS Ethical environment 1.0 Site 1 Nurses Variables Satisfaction Moral Distress Ethical Environment Collaboration.64 (p.001).16 NS.51 (p.001) Satisfaction (p.001).47 (p.001) Moral distress (p.001) Ethical climate 1.0 Site 2 Nurses (n 87) Variables Satisfaction Moral Distress Ethical Climate Collaboration.63 (p.001).22 (p.05).59 (p.001) Satisfaction (p.003).53 (p.001) Moral distress NS Ethical climate 1.0 MD, physician; RN, registered nurse; NS, not significant. of MDs with high moral distress scores to permit a similar comparison. DISCUSSION In exploring the factors that relate to collaboration between MDs and RNs caring for dying patients in ICUs, we found that both groups of professionals experienced moral distress. This phenomenon is a known hazard of the nursing profession, because RNs are frequently responsible for implementing a plan of care they have had little input into formulating. Attending MDs in this pilot study, however, also experienced moral distress. It is striking that the majority of the items listed as the highest sources of moral distress by both MDs and RNs had to do with pressure to continue aggressive treatment in situations where they did not think such treatment was warranted. This represents a change from the beginnings of the bioethics movement, when ethical conflicts (including the cases of Karen Ann Quinlan, Paul Brophy, and Claire Conroy) generally involved situations where the medical team insisted on aggressive care despite patient or family requests to withdraw it. Although both MDs and RNs clearly feel distressed at continuing aggressive treatment for terminal patients, RNs reported more moral distress than MDs in these futile care scenarios. For MDs, moral distress did not correlate with other study variables. RNs with high moral distress scores, however, reported a more negative ethical climate/environment in their ICU, lower satisfaction with quality of care, and less collaboration with MDs than RNs with low moral distress. Almost half (45%) of the RNs had left (17%) or considered leaving (28%) a position because of moral distress; RNs with high moral distress were more likely to have left or considered leaving a position than those with lower scores. Reviewing her previous studies (some unpublished), Corley (42) reported that 15% of RNs in her initial 1995 study said they had left a position, whereas 26% reported leaving in a 1999 sample. These findings demonstrate that RNs experiences of moral distress need to be recognized and addressed in ICU settings, particularly given the nationwide continuing shortage of RNs. However, our data suggest that moral distress may be an important area for further study for MDs as well, as significant correlations with satisfaction and other variables may be revealed with a larger sample size. When faced with similar futile care situations, why do RNs experience more moral distress than MDs? RNs may feel more intensely distressed at giving care they perceive as unwarranted because they have less impact on end-of-life decision making than MDs (32, 33, 53). This lack of input may increase frustration with the decisions that are made (32, 54). Furthermore, the RN is constantly confronted with the reality of patients suffering, whereas MDs do not spend as much time at the bedside. In this study, however, the difference in total moral distress scores was driven mainly by the RNs perceptions of the frequency of distressing situations: RNs and MDs were equally distressed by giving futile care, but RNs perceived that this happened more frequently than did 426 Crit Care Med 2007 Vol. 35, No. 2

6 the MDs. MDs generally have other practice sites or activities that provide a break from the stress of the ICU, and this ability to leave the ICU and engage in varied off-unit activities may influence their perception of the frequency of distressing situations. For RNs, the same dynamic may work in reverse; that is, even a few distressing situations may lead to a globalized sense of frequency because the RN cannot leave the situation as easily. Studies conducted in Europe (43, 55) and Australia (33) have shown that ICU RNs perceive care to be futile more frequently than do their MD colleagues. Frick et al. (55), in a study in Switzerland, examined this issue in depth and found that the longer patients stayed in the ICU and the higher their acuity, the more RN/MD disagreement about prognosis and direction of care emerged, with RNs being more pessimistic and more likely to consider withdrawal of care than MDs. Nurse-physician disagreement occurred in 63% of cases where the patient eventually died. The moral burden RNs feel in having to implement care that they believe to be contrary to the patient s best interests could be one explanation for these findings. Nursing concerns with EOL communication were clearly demonstrated in our findings (Table 4). Almost half of the RNs in site 1 reported that MDs in their ICU withheld information about diagnosis or prognosis from their patients occasionally or frequently. Most of these RNs (75%) reported frustration with the way MDs communicated with them about EOLC. Although selection bias may be operating here, as the MDs who responded to the survey may tend to be more interested in EOLC and better communicators with RNs and patients, only 10% of the MDs responding believed they withheld information on a regular basis, and none of the MDs surveyed reported that an RN had expressed frustration to them. This latter finding raises the question of whether the RNs had not clearly expressed their concerns, whether the MDs had not heard the RNs frustration, or some combination of both. Whereas RNs were generally satisfied with the way MDs communicated with their terminally ill patients (72%), MD satisfaction with RN communication was even higher (97%). These discrepancies raise questions about MD/RN communication patterns that deserve further study. Other differences between the MDs and RNs at site 1 are notable as well. Nurses overall reported a lower satisfaction with the quality of care (p.014), perceived their ethical environment as worse (p.001), and perceived less collaboration (p.001) than did MDs. For site 1 RNs and those with higher moral distress scores, moral distress correlated negatively with perceptions of ethical environment; this was not true for MDs. Negative perceptions of ethical climate were correlated with intentions to leave either their position or the profession in a recent study of RNs in Missouri (56). Again, the setting-bound nature of RN practice may make the quality of the ethical climate more important in daily practice than it is for MDs. Differences between RNs at the two different sites are also intriguing. Site 2 RNs were more satisfied with the quality of care at their institution than were RNs at the first site (p.001). Moral distress was significantly related to ethical environment for the site 1 RNs but unrelated to ethical climate for the site 2 RNs. Data from other studies and an integrative literature review showing the variability in decisions to forgo life support among different hospitals and different units (12, 13, 58) support the idea that unit cultures differ in important respects. Our findings bear interesting similarities to those of Thomas et al (37). In studying teamwork, these authors noted that RNs perceived more difficulty speaking up if they saw a problem with patient care, felt they had too little input into decision making, were less satisfied with conflict resolution, and were less positive about RN/MD teamwork than were MDs. Shannon et al. (58) studied MD, RN, and patient perceptions of quality of care and patient satisfaction, finding significant differences among provider/patient groups and between units. Physicians rated their perception of quality of care higher than did RNs, and there were high correlations between perceptions of collaboration and appraisal of unit quality of care among MDs and RNs. These differences are consistent with those found in our study. Differences in perspective between RNs and MDs may be due to differences in their position responsibilities, status/ authority, gender, or training and/or the differences in medical and nursing cultures (37, 54). The bedside RN is more continuously confronted by the patient s suffering, whether caused by medical interventions or by the disease process itself. The attending MD, however, must sign the order to withdraw life support and thus feels the burden of responsibility for such decisions and for the possibility that withdrawal of care may prematurely end the life of a patient who could have been saved. Indeed, experiences with patients who unexpectedly improved in response to aggressive therapy are powerful examples used by MDs to justify aggressive therapy in difficult cases. The Frick et al. (55) study underscores this. Although RNs, who were more pessimistic about prognosis and more likely to favor withdrawal of care, were also more accurate in predicting prognosis than MDs, they favored withdrawal of life support in some cases where the patient survived and went on to have a good quality of life. Thus RNs may tend to focus on the suffering of the many, whereas MDs are more concerned with the survival of the few. We view this as an inevitable tension given the differing perspectives of these disciplines and one that deserves explicit discussion in caregiving teams. In this study, the finding that RNs and MDs agreed on the most distressing situations provides some common ground for beginning these discussions. Both MD and RN perspectives are valuable and both are needed to help patients and families navigate the complex terrain of EOL decision making. In fact, most interventions that have proven successful in improving end-of-life decision making in ICUs (57, 59, 60) involved RNs in structured ways such as mandating their involvement in family meetings, holding multidisciplinary case reviews, or adding palliative care teams with a strong nursing component to the ICU environment. These successful interventions seemed to have two things in common: they improved communication between caregivers, and they brought nursing and physician perspectives together with those of patients and their families. These efforts may be the exception rather than the norm, however. Although virtually all the position papers on improving EOLC in ICUs have acknowledged the importance of nursing, few have made specific suggestions about how to improve collaboration. Since our findings indicate that MDs tend to believe that they are collaborating well, even when their nursing colleagues disagree, general exhortations to collaborate will not improve the interactions between RNs and MDs around EOLC. Specific in- Crit Care Med 2007 Vol. 35, No

7 terventions need to be developed and tested to accomplish this goal. In the institutions we studied, for example, our data suggest that increasing opportunities for structured communication between RNs and MDs about EOLC with special attention to the experience of moral distress might mitigate conflicts arising from differences in perspective. Improving the ethical climate in ICUs through explicit discussions of moral distress, recognition of the values differences between the two professions, and giving attention to improved collaboration may mitigate frustration arising from differences in perspective and improve the quality of patient care. Limitations. Any conclusions derived from these findings are preliminary, as this was an exploratory project. Limitations of this study include the lack of MD participation at site 2 and analysis challenges resulting from changes in the survey made after our experience with data collection at site 1. On the other hand, the MDs and RNs in these ICUs had a relatively low turnover (they had worked together on average for 7 years) and had similar religious and ethnic backgrounds. Since these factors would seem to minimize differences, the amount of disagreement we saw was notable. Cross-sectional, correlational studies cannot assert causal relationships. For example, it is not known whether improving a unit s ethical climate will improve MD/RN collaboration, or vice versa. Further study is needed to uncover the complex relationships among these variables. CONCLUSIONS In this study, RNs experienced more moral distress than MDs and perceived less collaboration than their MD colleagues. They perceived their ethical environment as more negative, and they were less satisfied with the quality of care provided on their units than were MDs. Provider assessments of quality of care were strongly related to perception of collaboration. This is consistent with data from other studies (37, 58). By definition, collaboration does not occur unless both parties agree it is occurring. Because the perspectives of both RNs and MDs are needed in EOLC, as elsewhere in ICUs, attention must be paid to how we build and maintain teamwork. Furthermore, our data point to the importance of moral distress and the ethical climate in ICUs as variables worthy of further investigation. Given the variability in perspectives and experiences seen in this study and others, it is clear that unit and institutional cultures vary widely. Examination of these issues in larger medical centers, as well as institutions with more staff turnover and ethnic variability among providers, could reveal wider differences than those seen here. Much of the research done to date has been conducted in single sites. Multiple-site research is needed to better understand the complex interplay among ethical factors and collaboration. Knowledge of this interplay can then lead to intervention studies that identify ways to improve teamwork between these professionals, so vital to patient and family experiences at the end of life. ACKNOWLEDGMENTS We thank Stephen Petterson, PhD, for his valuable help with the statistical analyses in this study, and Ms. Sarah Delgado, MSN, RN, for her assistance with literature review. REFERENCES 1. Pendergast TJ, Luce JM: Increasing incidence of withholding and withdrawal of life support from the critically ill. Am J Respir Crit Care Med 1997; 155: Tsai MH, Chen YC, Ho YP, et al: Organ system failure scoring system can predict hospital mortality in critically ill cirrhotic patients. J Clin Gastroenterol 2003; 37: Groeger JS, White P, Nierman DM, et al: Outcome for cancer patients requiring mechanical ventilation. J Clin Oncol 1999; 17: Price KJ, Thall PF, Kish SK, et al: Prognostic indicators for blood and bone marrow transplant patients admitted to an intensive care unit. Am J Respir Crit Care Med 1998; 158: Depuydt PO, Benoit DD, Vandewoude KH, et al: Outcome in noninvasively and invasively ventilated hematologic patients with acute respiratory failure. Chest 2004; 126: Angus DC, Barnato AE, Linde-Zwirble WT, et al: Use of intensive care at the end of life in the United States: An epidemiologic study. Crit Care Med 2004; 32: Lynn J, Harrell F Jr, Cohn F, et al: Prognoses of seriously ill hospitalized patients on the days before death: Implications for patient care and public policy. New Horiz 1997; 5: Truog R, Cist AF, Brackett SE, et al: Recommendation for end-of-life care in the intensive care unit: The Ethics Committee of the Society of Critical Care Medicine. Crit Care Med 2001; 29: Nelson JE, Danis M: End-of-life care in the intensive care unit: Where are we now? Crit Care Med 2001; 29(Suppl):N2 N9 10. Curtis JR, Rubenfeld GD (Eds). Managing Death in the ICU: The Transition From Cure to Comfort. New York, Oxford University Press, Rubenfeld GD, Curtis JR: Improving care for patients dying in the intensive care unit. Clin Chest Med 2003; 21: Pendergast TJ, Claessens MT, Luce JM: A national survey of end-of-life care for critically ill patients. Am J Respir Crit Care Med 1998; 158: End-of-life practices in European intensive care units: The Ethicus study. JAMA 2003; 290: Foley KM: Pain and symptom control in the dying ICU patient. In: Managing Death in the Intensive Care Unit: The Transition From Cure to Comfort. Curtis JR, Rubenfeld GD (Eds). New York, Oxford University Press, Payen JF, Bru O, Bosson JL, et al: Assessing pain in critically ill sedated patients by using a behavioral pain scale. Crit Care Med 2001; 29: Nelson JE, Meier DE, Litke A, et al: The symptom burden of chronic critical illness. Crit Care Med 2004; 32: Nelson JE, Meier DE, Oei EJ, et al: Selfreported symptom experience of critically ill cancer patients receiving intensive care. Crit Care Med 2001; 29: SUPPORT Principal Investigators: A controlled trial to improve care for seriously ill hospitalized patients. JAMA 1995; 274: Bergbom-Engberg I, Haljame H: Assessment of patients experience of discomforts during respirator therapy. Crit Care Med 1989; 17: Azoulay E, Sprung CL: Family-physician interactions in the intensive care unit. Crit Care Med 2004; 32: Azouley E, Chevret S, Leleu G, et al: Half the families of intensive care unit patients experience inadequate communication with physicians. Crit Care Med 2000; 28: Azouley E, Pouchard F, Chevret S, et al: Meeting the needs of intensive care unit families: A multicenter study. Am J Respir Crit Care Med 2001; 163: Pochard F, Azoulay E, Chevret S, et al: Symptoms of anxiety and depression in family members of intensive care unit patients: Ethical hypothesis regarding decision-making capacity. Crit Care Med 2001; 29: Mission statement, Society of Critical Care Medicine. Available at: Accessed December 26, Baggs JG, Mick DJ: Collaboration: A tool addressing ethical issues for elderly patients near the end of life in intensive care units. J Gerontological Nurs 2000; 26: Gardner DB: Ten lessons in collaboration. Online J Issues Nurs [serial online]. July 428 Crit Care Med 2007 Vol. 35, No. 2

8 2005; 10: manuscript 1. Available at: htm. Accessed December 26, Lindeke LL, Sieckert AM: Nurse-physician workplace collaboration. Online J Issues Nurs [serial online] July 2005; 10: Manuscript 4. Available from: org/ojin/topic26/tpc26_4.htm. Accessed December 26, Baggs JG, Schmitt MH, Mushlin AI, et al: Association between nurse-physician collaboration and patient outcomes in three intensive care units. Crit Care Med 1999; 27: Baggs JG, Ryan SA, Phelps CE, et al: The association between inter-disciplinary collaboration and patient outcomes in medical intensive care. Heart-Lung 1992; 21: Zimmerman JE, Shortell SM, Rousseau DM, et al: Improving intensive care: Observations based on organizational case studies in nine intensive care units: A prospective, multicenter study. Crit Care Med 1993; 21: Breen CM, Abernathy AP, Abbott KH, et al: Conflict associated with decisions to limit life-sustaining treatment in intensive care units. J Gen Intern Med 2001; 16: Ferrand E, Lemarie F, Regnair B, et al: Discrepancies between perceptions by physicians and nursing staff of intensive care unit end-of-life decisions. Am J Respir Crit Care 2003; 167: Bucknall T, Thomas S: Nurses reflections on problems associated with decision-making in critical care settings. J Adv Nurs 1997; 25: Solomon MZ, O Donnell L, Jennings B, et al: Decisions near the end of life: Professional views regarding life sustaining treatments. Am J Public Health 1993; 83: Kirchhoff KT, Spuhler V, Walker L, et al: Intensive care nurses experiences with endof-life care. Am J Crit Care 2000; 9: Reckling JB: Who plays what role in decisions about withholding and withdrawing life sustaining treatment? J Clin Ethics 1997; 8: Thomas EJ, Sexton JB, Helmreich RL: Discrepant attitudes about teamwork among critical care nurses and physicians. Crit Care Med 2003; 31: Meltzer LS, Huckabay LM: Critical care nurses perceptions of futile care and its effect on burnout. Am J Crit Care 2004; 13: Corley MC: Moral distress of critical care nurses. Am J Crit Care 1995; 4: Jameton A: Dilemmas of moral distress: Moral responsibility and nursing practice. AWHONNS Clin Issues Perinat Womens Health Nurs 1993; 4: Hamric AB: Moral distress in everyday ethics. Nurs Outlook 2000; 48: Corley MC: Nurse moral distress: A proposed theory and research agenda. Nurs Ethics 2002; 9: Kalvemark S, Hoglund AT, Hansson MG, et al: Living with conflicts Ethical dilemmas and moral distress in the health care system. Soc Sci Med 2004; 58: Hamric AB, Davis W, Childress MD: Moral distress in health-care providers: What is it and what can we do about it? Pharos of Alpha Omega Alpha Honor Med Soc 2006; 69: McDaniel C: Ethical environment: Reports of practicing nurses. Nurs Clin North Am 1998; 33: Olson L: Ethical climate in health care organizations. Int Nurs Rev 1995; 42: Rushton CH: Creating an ethical practice environment: A focus on advocacy. Crit Care Clin North Am 1995; 7: Olson LL: Hospital nurses perceptions of the ethical climate of their work setting. Image J Nurs Sch 1998; 30: Corley MC, Elswick RK, Gorman M, et al: Development and evaluation of a moral distress scale. J Adv Nurs 2001; 33: McDaniel C: Development and psychometric properties of the ethics environment questionnaire. Med Care 1997; 35: Hojat M, Fields SK, Veloski JJ, et al: Psychometric properties of an attitude scale measuring physician-nurse collaboration. Eval Health Prof 1999; 22: Shortell SM, Rosseau DM, Gillies RR, et al: Organizational assessment in ICUs: Construct development, reliability and validity of the ICU nurse-physician questionnaire. Med Care 1991; 29: Cardoso T, Fonseca T, Pereira S, et al: Life sustaining treatment decisions in Portuguese intensive care units: A national survey of intensive care physicians. Crit Care (London) 2003; 7:R167 R Oberle K, Hughes D: Doctors and nurses perceptions of ethical problems in end-of-life decisions. J Adv Nurs 2001; 33: Frick S, Uehlinger DE, Zenklusen RM: Medical futility: predicting outcome of intensive care unit patients by nurses and doctors A prospective comparative study. Crit Care Med 2003; 31: Hart SE: Hospital ethical climates and registered nurses turnover intentions. J Nurs Schol 2005; 37: Lilly CM, DeVeo DL, Sonna LA, et al: An intensive communication intervention for the critically ill. Am J Med 2000; 109: Shannon SE, Mitchell PH, Cain KC: Patients, nurses, and physicians have differing views of quality of critical care. J Nurs Schol 2002; 34: Lilly CM, Sonna LA, Haley KJ, et al: Intensive communication: Four year follow-up from a clinical practice study. Crit Care Med 2003; 31: Campbell ML, Guzman JA: Impact of a proactive approach to improve end-of-life care in a medical ICU. Chest 2003; 123: Crit Care Med 2007 Vol. 35, No