Published ahead of print on January 24, 2003, doi:10.1164/rccm.200207-752OC
© 2003 American Thoracic Society Discrepancies between Perceptions by Physicians and Nursing Staff of Intensive Care Unit End-of-Life DecisionsRéanimation Chirurgicale, Hôpital Henri Mondor; Réanimation Médicale, Hôpital Universitaire Henri Mondor, AP-HP, Créteil; Laboratoire d'Ethique Médicale et de Santé Publique, Hôpital Necker Enfants-Malades; Réanimation Médicale, Hôpital Universitaire Bichat-Claude Bernard; Réanimation Médicale, Hôpital Universitaire Cochin, AP-HP, Paris; Réanimation Médicale, Hôpital La Source, Orléans; Réanimation Médicale, Hôpital Universitaire La Croix Rousse, Lyon; Réanimation Médicale, Centre Hospitalier Universitaire, Angers; Réanimation Chirurgicale, DAR B, Hôpital Universitaire Saint-Eloi, Montpellier; Réanimation Polyvalente, Centre Hospitalier Général, Valenciennes; Réanimation Médicale, Hôpital Universitaire de la Cavale Blanche, Brest; and Réanimation Médicale, Hôpital Universitaire Jean Bernard, Poitiers, France Correspondence and requests for reprints should be addressed to Edouard Ferrand, Service d'Anesthésie-Réanimation, Unité de Réanimation Chirurgicale et Traumatologique, Hôpital Henri-Mondor, AP-HP, 51 rue du Mal de Lattre de Tassigny, 94010 Créteil cedex, France. E-mail: edouard.ferrand{at}hmn.ap-hop-paris.fr
Several studies have pointed out ethical shortcomings in the decision-making process for withholding or withdrawing life-supporting treatments. We conducted a study to evaluate the perceptions of all caregivers involved in this process in the intensive care unit. A closed-ended questionnaire was completed by 3,156 nursing staff members and 521 physicians from 133 French intensive care units (participation rate, 42%). Decision-making processes were perceived as satisfactory by 73% of physicians and by only 33% of the nursing staff. More than 90% of caregivers believed that decision-making should be collaborative, but 50% of physicians and only 27% of nursing staff members believed that the nursing staff was actually involved (p < 0.001). Fear of litigation was a reason given by physicians for modifying information given to competent patients, families, and nursing staff. Perceptions by nursing staff may be a reliable indicator of the quality of medical decision-making processes and may serve as a simple and effective tool for evaluating everyday practice. Recommendations and legislation may help to build consensus and avoid conflicts among caregivers at each step of the decision-making process.
Key Words: end-of-life decisions • ethics • decision-making • critical care • intensive care Over the last few decades, decisions to forego life-sustaining treatments (DFLSTs) have become common in intensive care units (ICUs) (1–4). The guidelines or legal precedents that legitimate these decisions in some countries (5–8) are probably used as a point of reference elsewhere (9, 10). Several studies have pointed out ethical shortcomings in the decision-making process, including failure to consider nurses' opinions (4, 11). Whereas physicians are primarily concerned with curing their patients, nurses focus on the impact of care on their patients (12–14). A major issue is whether consent of the patient or surrogate is sufficient to make DFLSTs ethically legitimate when there is disagreement among the caregivers (15). The considerable moral responsibility conferred on nurses by their unique proximity to the patient and his/her relatives and their interaction with the physician team are strong arguments in favor of including nurses in the specific DFLST process (14, 16–19). Recently, two lawsuits in France and one in Belgium have been filed against ICU physicians who had withdrawn mechanical ventilation from hopelessly ill patients (20–22). In these three instances, the lawsuits were filed by nurses, who charged the physicians with euthanasia. This indicates a major dissent among caregivers and great dissatisfaction of nurses about the handling of DFLSTs in the ICU. In one of the French cases, the physician was found guilty in 1995 of homicide, a ruling that gave rise to considerable debate (20). The Belgian case is awaiting trial, but the Belgian Society for Intensive Care has issued a position paper supporting the physician, and a change to existing legislation on DFLSTs is being considered by the Belgian parliament (22). The absence in most European countries of recommendations from scientific bodies and of legislation on DFLSTs in critical care medicine probably contributes to the occurrence of conflicts among ICU caregivers. In May 2002, the French Language Society of Critical Care Medicine (Société de Réanimation de Langue Française) issued its first recommendations on DFLSTs (23). We conducted a survey to evaluate the perceptions of all caregivers in DFLSTs for ICU patients who are mentally incompetent. In France, competent patients decide for themselves, but decisions for incompetent patients are left to the physicians, not the family members. Conflicts about values and ambiguity in relationships among caregivers have been suggested (24) and may be amplified by the need to make DFLSTs. Few studies have evaluated the hypothesis that good collaboration, a major determinant of nurse satisfaction, may improve the experience of dying patients. We specifically sought to assess associations linking ICU policies, decision-making processes, co-operation among nurses and physicians, and caregiver satisfaction, as reported by the nursing staff and physicians. This study confirms that perceptions of nurses and physicians differ widely.
Study Participants We sent a study project and reply form to the medical director and senior head nurse of each ICU in the 320 university or general hospitals on a list published by the French Language Society of Critical Care Medicine. The reply form asked who would be the local investigator, how many physicians and other professionals worked in the ICU, their job title, and their work shift. In each ICU, all caregivers who had been working in the ICU for at least three months were invited to participate in the study. Physicians included residents, fellows, attending physicians, and the ICU director. The nursing staff included nurses, nursing assistants, physiotherapists, and head nurses from both day and night shifts. Physiotherapists approach patients in a manner similar to nurses, in terms of the nature of the care they provide and the attention they give to what patients express; consequently, we included the few physiotherapist participants into the nursing staff category.
Survey Instrument For the pretest validation of the questionnaires, we conducted semistructured interviews with attending physicians, head nurses, nurses, and nurse assistants on the staffs of the medical ICU in Poitiers, France and the surgical ICU in Créteil, France. These interviews showed that the questionnaires were easily understood and that the full range of response options was used. Because the terms "withdrawing," "withholding," "ethical standards," and "high-quality decision-making" can be unclear or can give rise to a variety of interpretations, participants were provided with the following definitions:
Questionnaire Administration Data collected from the questionnaires were double keyboarded.
Ethics Committee
Statistical Analysis Multiple logistic regression analysis was performed to examine the relation between the characteristics of the caregivers and their perceptions of DFLSTs.
Of the 320 ICUs canvassed for the study, 157 (49%) agreed to participate. Of these 157, only the 133 units with more than 10% of the personnel returning completed questionnaires were included in the study. Of these 133 ICUs, 90 (67.7%) were mixed medical–surgical, 22 (16.5%) were surgical, and 21 (15.8%) were medical. Ninety-eight (73.6%) ICUs were in university hospitals and 35 (26.4%) in general hospitals. Questionnaires with answers to more than 90% of the items were returned by 3,156 of the 6,341 (49.8%) nursing staff members (Table 1) and by 521 of the 915 (56.9%) physicians (Table 2) working in the 133 ICUs.
Ninety-one percent (n = 2,875) of the 3,156 nursing staff members and 99% (n = 517) of the 521 physicians had personal experience with DFLSTs as part of their work in the ICU. Tables 3 and 4 show how caregivers perceived DFLSTs and the place of these decisions in the ICU.
ICU Commitment to High Ethical Standards Sixty-five percent of nursing staff members (n = 2,036) and 78% of physicians (n = 415) believed that their ICU was committed to high ethical standards. Physicians were more likely than nursing staff members to believe that the nursing staff was involved in this commitment (75% of physicians [n = 396] vs. 43% of nursing staff members [n = 1,360]; p < 0.001) with no differences between ICUs of university and general hospitals (data not shown). Nursing staff members in surgical ICUs were more likely to believe that they were not sufficiently involved by physicians than were their counterparts in medical or medical–surgical ICUs (16.7, 20.1, and 31.1%, respectively; p < 0.0001).
Satisfaction with the Process for Making Decisions to Forego Life-Sustaining Treatment
Decision-Making Involvement in the decision-making process of a professional who had no role in patient care was viewed favorably by 58% of nursing staff members (n = 1,830) and 42% of physicians (n = 221). Most nursing staff members favored a psychologist, whereas physicians' responses were equally distributed among the various possibilities suggested to them (Table 5) .
Among physicians, 79% (n = 418) believed that, before making a DFLST, they considered the opinion of the nursing staff regarding the course of the patient's treatment in the ICU, as compared with only 31% of nursing staff members (n = 953) (p < 0.001). Furthermore, 32.2% of physicians (n = 170) and 8.8% of nursing staff members (n = 277; p < 0.001) believed that DFLSTs were followed by adequate discussion of these decisions. Also, 16% of physicians (n = 85) and 21% of nursing staff members (n = 647) reported that they felt isolated most of the time.
Perceptions by Nursing Staff Members According to Work Shift and Time in the ICU
Communication with the Family Presence of the nurses at meetings to discuss DFLSTs with the family was considered necessary by 56% of nursing staff members (n = 1,758) and 36% of physicians (n = 189) (p < 0.05). Seventy-five percent of nursing staff members (n = 2,362) and 75% of physicians (n = 500) believed that the family should always be informed of DFLSTs. However, only 42% (n = 1,339) and 66% (n = 348), respectively, believed that families were always informed in actual clinical practice (p < 0.05). Only 69% of nursing staff members (n = 2196) and 61% of physicians (n = 323) (not significant) believed that families should be informed fully; the main reason for not providing full information was that this might add to the family's distress (35% of nursing staff [n = 1,100] and 59% [n = 311] of physicians).
Criteria for Decisions to Forego Life-Sustaining Treatments
Liability A total of 42% of nursing staff members (n = 1,312) and 30% of physicians (n = 159) believed that the nursing staff in charge of the patient should share with the physicians the responsibility for DFLSTs, including legal responsibility. Twelve percent of nursing staff members (n = 391) believed that their role during the discussion was only to make their opinion heard clearly, without sharing in the responsibility for the decision. Seventy-eight percent of physicians (n = 411) but only 48% of nursing staff members (n = 1,526) believed that nurses (in the presence of the physician) could implement a DFLST made by the physician and consisting in increased sedation (p < 0.05); corresponding figures were 76% (n = 402) and 58% (n = 1,846) for discontinuing vasoactive therapy (p < 0.05), 63% (n = 334) and 51% (n = 1,604) for decreasing the FIO2 (p < 0.05), and 30% (n = 160) and 28% (n = 898) for extubating the patient. Most physicians (76.7%, n = 405) did not believe they were breaking the law when they made DFLSTs. However, some physicians reported that they worried about malpractice suits (Table 8) , and concern about litigation was one of the reasons given by physicians for modifying the information they provided to competent patients.
Written DFLST procedures were available in only five ICUs. Thirty-three percent of physicians (n = 175) believed that the recent increase in litigation made written procedures desirable and 58% reported that their reports of DFLSTs in medical records did not faithfully describe reality (n = 92, 17% of all medical respondents). Fifty-seven percent of physicians (n = 301) were favorable to a change in current legislation about DFLSTs in the ICU.
These findings carry several messages. First, they indicate that nursing staff members are often dissatisfied with the DFLST process in French ICUs. Second, we found marked differences between perceptions of physicians and nursing staff members, with most physicians being satisfied with these procedures. Third, fear of litigation clearly had an unfavorable influence on the quality of DFLST procedures. In this study, 75% of nursing staff members reported dissatisfaction with DFLSTs. In this area of heated controversy on both sides of the Atlantic, the negative opinion of the caregivers who are closest to dying patients is very disturbing, if not surprising. In a study conducted in five hospitals in the United States, 75% of 759 nurses felt dissatisfied with management strategies and with their ICU's commitment to ethical standards and 50% said that, when caring for dying patients, they performed acts that contradicted their moral beliefs (25). Nurses who feel dissatisfied may perform acts that are not consonant with professional values. In a questionnaire study conducted by Asch, 17% of 1,139 nurses reported that they had engaged in euthanasia or assisted suicide, including 8% without an order from a physician (24). Some nurses reported injecting saline instead of vasopressors ordered by physicians. Nurse satisfaction is closely dependent on the amount of collaboration within the caregiver staff (12, 26, 27). In our study, nearly 75% of nursing staff members believed that collaboration was inadequate during decision-making, although the overwhelming majority of caregivers in both groups believed that collaboration was mandatory, as previously reported (28) or recommended (29, 30). This finding is in keeping with the lack of involvement of the nursing staff in half the DFLSTs recorded in the French national LATAREA study (4). Close interdisciplinary collaboration in the ICU is ethically desirable and improves clinical outcomes (11, 12, 31–36). Differences between predicted and observed mortality in ICU patients were significantly associated with the degree of interaction among ICU staff members (31), and staff satisfaction with the decision-making process was significantly related to patient outcomes (32). In another study, the degree of physician–nurse collaboration as perceived by the nurses was associated with patient outcomes in a medical ICU but not in a surgical ICU or a medical–surgical ICU; collaboration as perceived by the physicians was not associated with outcomes (35). One of the limitations of these studies is that only ICU death and a need for readmission to the ICU were evaluated: other outcomes such as patient/family satisfaction, cost, and longer-term mortality were not considered (35). Furthermore, these studies excluded patients for whom DFLSTs were made. Few studies have evaluated the hypothesis that good collaboration may improve the experience of dying patients. The observational phase of the SUPPORT study published in 1995 showed a high rate of deficient physician–patient communication and inappropriate treatment, with inadequate pain management in dying patients and absence of knowledge of patient wishes regarding cardiopulmonary resuscitation in over 50% of cases (11). In the interventional phase of the SUPPORT study, a specially trained nurse interviewed patients and families about their preferences regarding end-of-life care and encouraged caregivers to direct sufficient attention to pain control (11). This intervention failed to improve outcomes reflecting the experience of dying patients. The authors suggested that the intervention may have occurred too late in the decision-making process or that the physician–patient relationship might have been better had the patient spoken with the physician rather than with the research nurse (37). Furthermore, the nurse, although specially trained, was not part of the ICU staff, raising the possibility that a nurse from the ICU would perhaps have been more successful in improving communication among the caregiver staff (38). We found a significant association between the degree of nursing staff involvement in the ICU's general commitment to ethical standards and nursing staff satisfaction with DFLSTs, in keeping with several earlier studies (34, 35, 39). In a 1996 single-center study among nurses who were involved in decisions to withdraw mechanical ventilation and who believed the decision was morally correct, 84% were very satisfied with withdrawal procedures (33). Another interesting finding from our study is that nursing staff satisfaction with DFLST procedures was significantly better in medical ICUs than in surgical ICUs, although a majority of nursing staff members were very dissatisfied with these procedures in medical, surgical, and medical–surgical ICUs. Baggs and coworkers made a similar observation and suggested that the need for close collaboration may be greater in ICUs with more complex patients, such as those admitted to medical ICUs (35). Finally, in our study, fear of litigation was probably an obstacle to communication during DFLST process. About 50 and 20% of physicians reported that they gave inaccurate information to families and to nursing staff members, respectively. At the time of our study, changes in legislation about end-of-life care in ICUs and other related issues were under discussion in France; we found that 58% of physicians were favorable to a change in legislation (40). Thus, the setting was very different from that encountered in the U.S., where family members generally make DFLSTs for ICU patients, with the guidance and advice of physicians (10, 41, 42). Conceivably, the lack of recommendations from official bodies like scientific societies and failure to acknowledge the right of patients to full autonomy—from informed consent to refusal of care—may lead to covert and consequently illegal practices, to inadequate support of the patient and family, and to insufficient trust among ICU caregivers, a situation that may increase the likelihood of malpractice suits (20–22). Decisions made openly and discussed in depth with all those involved may be less likely to lead to litigation, rather than the opposite. The recommendations on DFLSTs in ICUs published in May 2002 by the French Society of Critical Care Medicine (SRLF) strongly emphasize that physicians are under a legal obligation to document these decisions in the patient's medical records (23). Some limitations of this study should be pointed out. First, the questionnaire dealt with the physician and nurses' perceptions as to end-of-life care and was not intended to address the issue of patients or family members opinions. Second, although we pretested our questionnaire, we acknowledge that a closed-ended questionnaire offers an assessment that is driven by those who write the response options, precluding new input from the respondents. Third, although the response rate was similar to that in many previous studies, the opinions of 75% of potential participants escaped evaluation by our study. Finally, the perceived poor quality of decision-making procedures, together with the lack of an official statement from scientific bodies, suggests that nurses' perceptions may play a role in preventing inappropriate decisions about patients whose consent is not obtained (34, 43, 44). Physicians should initiate interdisciplinary collaboration by allowing all involved staff members to communicate their own opinions. DFLSTs generate painful conflicts between competing ethical values. Neither recommendations issued by learned societies nor changes in legislation can lighten the weight of the decision nor shift the responsibility away from the physician. However, recommendations, laws, and good practices can help to build consensus and avoid disagreement among caregivers at each step of the decision-making process. Compatible with our study results, we suggest that operating procedures should be developed to detect reservations, passive opposition, or resistance to decision-making processes, particularly regarding DFLST, which require a high degree of collaboration and serenity. Measuring satisfaction of the various members of the healthcare team with these decisions may be a simple and effective tool for evaluating everyday practice.
The authors are indebted to Dr. A. Wolfe for helpful advice and thoughtful reading of this manuscript and to the hospitals and their staffs for participating in the study.
Supported by a grant from the Direction des Hôpitaux (Programme Hospitalier de Recherche Clinique AOM 98 301). This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form July 26, 2002; accepted in final form January 6, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2 statistic was used to compare categoric variables. Categoric variables are expressed as percentage of the group from the group from which they were derived, with their 95% confidence intervals. 
