Share this article on:

Negotiating End-of-Life Decision Making: A Comparison of Japanese and U.S. Residents' Approaches

Gabbay, Baback B. MD; Matsumura, Shinji MD, MSHS; Etzioni, Shiri MD; Asch, Steven M. MD, MPH; Rosenfeld, Kenneth E. MD; Shiojiri, Toshiaki MD; Balingit, Peter P. MD; Lorenz, Karl A. MD, MSHS

Featured Topic Research Report

Purpose To compare Japanese and U.S. resident physicians' attitudes, clinical experiences, and emotional responses regarding making disclosures to patients facing incurable illnesses.

Method From September 2003 to June 2004, the authors used a ten-item self-administered anonymous questionnaire in a cross-sectional survey of 103 internal medicine residents at two U.S. sites in Los Angeles, California, and 244 general medical practice residents at five Japanese sites in Central Honshu, Kyushu, Okinawa, Japan.

Results The Japanese residents were more likely to favor including the family in disclosing diagnosis (95% versus 45%, p < .001) and prognosis (95% versus 51%, p < .001) of metastatic gastric cancer. Of residents who favored diagnostic or prognostic disclosure to both the patient and family, Japanese residents were more likely to prefer discussion with the family first. Trainees in Japan expressed greater uncertainty about ethical practices related to disclosure of diagnosis or prognosis. Many Japanese and U.S. residents indicated that they had deceived a patient at the request of a family (76% versus 18 %, p < .001), or provided nonbeneficial care (56% versus 72%, p < .05), and many expressed guilt about these behaviors.

Conclusions The residents' approaches to end-of-life decision making reflect known cultural preferences related to the role of patients and their families. Although Japanese trainees were more likely to endorse the role of the family, they expressed greater uncertainty about their approach. Difficulty and uncertainty in end-of-life decision making were common among both the Japanese and U.S. residents. Both groups would benefit from ethical training to negotiate diverse, changing norms regarding end-of-life decision making.

At the time this was written, Dr. Gabbay was a fourth-year medical student at the David Geffen School of Medicine at UCLA. He began his psychiatry residency at Harbor/UCLA Medical Center, Torrance, California, in June 2005.

Dr. Matsumura is a visiting research fellow, the International Research Center for Medical Education at the University of Tokyo in Tokyo, Japan.

Dr. Etzioni is assistant program director, the Interdisciplinary Palliative Fellowship at VA Greater Los Angeles, Los Angeles, California.

Dr. Asch is the deputy associate chief of staff for Health Services Research and Development, VA Greater Los Angeles, and associate professor of medicine, David Geffen School of Medicine at UCLA, Los Angeles, California.

Dr. Rosenfeld is director, Veterans Integrated Palliative Program, VA Greater Los Angeles, and associate clinical professor, David Geffen School of Medicine at UCLA, Los Angeles, California.

Dr. Shiojiri is residency program director, Asahi General Hospital, Asahi, Japan.

Dr. Balingit is assistant clinical professor of medicine, David Geffen School of Medicine at UCLA, Olive View-UCLA Medical Center, Sylmar, California.

Dr. Lorenz is research associate, the VA Greater Los Angeles, and assistant professor of medicine, David Geffen School of Medicine at UCLA, Los Angeles, California. He is a VA Health Services Research and Development Career Development Awardee.

For information about the authors, see the end of this report.

Correspondence should be addressed to Dr. Gabbay, Veteran's Integrated Palliative Program, 11301 Wilshire Boulevard, Mail Code 111G, Los Angeles, CA 90073; telephone: (310) 441-0234; fax: (310) 268-4272; e-mail: 〈〉.

The ethical principle of autonomy can be defined as the ability of individuals to act intentionally in an informed manner, free from control or interference from others.1 As such, this principle guides physicians' practice of informing individuals about their condition–even if the disclosure may dishearten them–and investing them with responsibility for their own treatment decisions.2

However, the way that the principle of autonomy is applied is deeply enmeshed in the cultural values of physicians, patients and families, especially with regard to end-of-life decision making. For example, physicians in the United States strongly uphold the principle of universal disclosure with regard to cancer, and almost all American cancer patients are informed of their diagnosis in some manner when the disease is discovered.3 In contrast, Japanese physicians often view the family rather than the individual as the autonomous unit to be respected. Subsequently, they often do not inform patients of a cancer diagnosis, and physicians often defer to the family's rather than the patient's preferences for disclosure.4–7

Attitudes and practice regarding disclosure may be changing today in Japan as they changed in the United States during the 1960s and 1970s. In 1961, 88% of U.S. physicians did not inform cancer patients of their diagnosis, although approximately 20 years later a large majority upheld disclosure.3,8 A recent study described younger versus older Japanese physicians as more supportive of full disclosure and individual autonomy.9 Another study in Japan noted that the diagnosis of cancer shared was shared with 27% of adult patients in 1993 versus 71% of patients in 1998.10 These observations coincide with Japanese efforts to improve palliative care for its aging population. Although in 1990 there were only three specialized wards for palliative care in Japan, by 2002, there were 89 such wards.11

Physician trainees reflect the cutting edge of changing attitudes and thus may be the best source of information about future patterns of medical care delivery. With that in mind, we surveyed Japanese and U.S. medical residents regarding their attitudes, experiences, and emotional responses to diagnostic and prognostic disclosure to patients facing incurable illnesses. We hypothesized that changing Japanese cultural norms would be associated with higher levels of trainee uncertainty. We studied Japanese and U.S. medical residents' self-reports of diagnostic and prognostic disclosures to identify how conflicting paradigms of decision making influence trainees' attitudes, experiences, uncertainty, and guilt related to such disclosures.

Back to Top | Article Outline



We surveyed residents at five sites in central Honshu, Japan, and two sites in Los Angeles, California, between September 2003 and June 2004. The Japanese sites were Asahi General Hospital (AGH) in Asahi City; Aso Iizuka Hospital (AIH) in Iizuka; Ryukyu University Hospital (RUH) in Okinawa; St. Luke's International Hospital (SLIH) in Tokyo; and University of Tsukuba Hospital (UTH) in Tsukuba. All five hospitals care for a relatively homogenous elderly Japanese population. Of these five hospitals, only AGH and SLIH possess an inpatient palliative care unit with 20 and 25 palliative beds, respectively.

The U.S. sites were Olive View–UCLA Medical Center (OV) in Sylmar, California, and the VA Greater Los Angeles Health care System (VA GLA) in Los Angeles, CA. OV is a 337-bed county-owned public hospital with an approximately 60% Hispanic patient population. VA GLA is a 945-bed tertiary care hospital with a predominantly indigent veteran patient population that is 32% African American and 11% Hispanic. VA GLA has an inpatient palliative care consult service and fellowship training program. All sites have strong university affiliations.

Back to Top | Article Outline

Questionnaire development and study design

Items were adapted by two of the investigators (KAL, SM) from a previous patient questionnaire.12 Ten items assessed the residents' attitudes and experiences regarding end of life care, prognostic disclosure, and autonomy including case scenarios that addressed cancer diagnosis, prognosis, and treatment decisions. Scenario 1 described disclosing a metastatic gastric cancer diagnosis to a 50-year-old patient. Scenario 2 described disclosing a terminal prognosis to the same patient. Scenario 3 described a patient with postobstructive pneumonia and the participation of the patient, family, and physician in deciding whether or not to intubate.

The questionnaire assessed residents' attitudes about disclosure and decision making and their perceived certainty about their preferences. It also assessed residents' reports of providing treatment perceived as nonbeneficial, deceiving patients, and experiencing guilt about these actions. The questionnaire was translated into Japanese and back-translated into English to confirm accuracy. The study was reviewed and approved by the institutional review boards of the participating American and Japanese institutions. The questionnaire, which was administered anonymously, required fewer than five minutes to complete.

Back to Top | Article Outline

Participants, data collection, and analysis

We attempted to survey all the internal medicine residents who were training on site at the U.S. institutions and all the general medicine practice residents training on site at the Japanese institutions. Residents in Japan train for two years in general practice (internal medicine, obstetrics, pediatrics, primary care, surgery, and electives) before entering more specialized programs. First- and second-year residents at Japanese hospitals correspond to interns and second-year residents in the United States. Because some of the Japanese residents were surveyed at the beginning of their training year, and because they had received little clinical training in medical school, some of them had not yet exercised significant clinical responsibilities. The U.S. residents were all enrolled in three-year internal medicine residency programs. Since all of the first year residents were surveyed in the middle of their training year, the majority had begun to exercise significant clinical responsibilities, unlike some of the Japanese first-year residents.

A research assistant (BBG), fellow (SE), and attending physicians at each local institution approached residents as a group and up to three times individually during noon conferences, grand rounds, morning report, and before work rounds and asked them to complete the questionnaire. All questionnaire responses were entered in SPSS Version 10.0 Statistical Software. We used descriptive statistics to analyze and compare the Japanese and U.S. residents' responses using appropriate statistical tests. Tests of categorical responses used chi-square tests and proportional differences used t tests.

Back to Top | Article Outline


Description of respondents

A total of 347 questionnaires were collected from 244 Japanese and 103 U.S. residents. The response rate varied from 64% to 100% across the seven sites and was 74% for the Japanese sites versus 71% for the U.S. sites overall. The study was not adequately powered to evaluate site differences within the United States and Japan, although our findings did not suggest meaningful differences in site responses within each country. A total of 152 Japanese residents (62.3%) and 45 U.S. residents (43.7%) were in their first year of postgraduate training (PGY1). Ninety (37.7%) Japanese residents and 58 (56.3%) U.S. residents were in at least their second year of postgraduate medical training (group difference, p < .01).

Back to Top | Article Outline

Preferences for disclosure of diagnosis and prognosis

When asked whom should be told a cancer diagnosis, 229 (94%) of the Japanese residents indicated “both—the patient and the family” compared to the U.S. residents, 54 (53%) of whom indicated the “patient” only (group difference, p < .001; see Table 1). However, the Japanese residents were less certain than the U.S. residents of their approach. Only 20 (8%) of the Japanese residents were “completely certain,” compared with 64 (62%) of the U.S. residents (group difference, p < .001). Of those respondents who supported disclosing the diagnosis to both patients and family members, 99 (44%) of the Japanese residents reported they would inform the family first, while only a small minority (2%) of the U.S. residents said they would follow the same procedure (group difference, p < .001).

Table 1

Table 1

Regarding the disclosure of a poor prognosis, 175 (72%) of the Japanese residents again preferred to disclose the prognosis to “both—the patient and the family” compared to 46 (45%) of the U.S. residents, who indicated a preference to disclose to the “patient” only (group difference, p < .001). Furthermore, 56 (23%) of the Japanese residents chose to disclose the prognosis solely to the family, compared to only one (1%) U.S. resident (group difference, p < .001). The Japanese residents expressed more uncertainty than did the U.S. residents about their approach. Only 12 (5%) of the Japanese residents were “completely certain,” compared to 50 (49%) of the U.S. residents (group difference, p < .001). Of those who answered they would disclose to both patient and family, 67 (40%) of the Japanese residents favored disclosure to the family first, while 34 (85%) of the U.S. residents favored disclosure to the patient first (group difference, p < .001). The differences in attitudes and certainty we noted between Japanese and U.S. residents were maintained when we restricted these comparisons to housestaff with equivalent clinical experience.

Back to Top | Article Outline

Preferences for patient, family, and physician involvement in decision making

All the residents were asked to rate how involved the patient, family members, and the physician should each be in a decision about whether or not to intubate a patient with cancer and postobstructive pneumonia. The Japanese and U.S. residents assigned a similar relative hierarchy of importance in the involvement of the patient, family, and physician in the decision. On a five-point scale (1 = not at all involved, 5 = very involved) both groups of residents rated the importance of the patient's involvement highly, although the U.S. residents rated it as of slightly higher importance (4.67 for the Japanese residents versus 4.95 for U.S. residents, p < .001) Both Japanese and U.S. residents accorded family members a similar and only slightly lower importance (4.12 versus 4.14, p = .86). The Japanese residents rated the importance of physician involvement somewhat lower than did their U.S. counterparts (3.51 versus 4.39, p < .001).

Back to Top | Article Outline

Experiences deceiving patients, providing nonbeneficial care, and associated guilt

One hundred six (43%) Japanese and 101 (98%) U.S. residents had cared for at least one dying patient during their medical training (group difference, p < .001; see Table 2). Of these respondents, 78 (78%) Japanese, and 18 (18%) U.S. residents had hidden a cancer diagnosis at the family's request (group difference, p < .001). Of those who had concealed the diagnosis, 35 (45%) Japanese residents and 11 (61%) U.S. residents reported feeling guilty “all of the time” or “most of the time” as a result. Fifty-seven (56%) Japanese residents and 73 (72%) U.S. residents stated that they had provided nonbeneficial medical treatment to a cancer patient (group difference, p < .05). Twenty-eight (49%) of the Japanese residents who had provided such treatment felt guilty “all of the time” or “most of the time,” but only 21 (29%) of the corresponding U.S. residents reported feeling guilt “all of the time” or “most of the time” as a result.

Table 2

Table 2

Back to Top | Article Outline


Our study identified patterns of information disclosure and-end of-life care provision, including important cross-cultural differences, in the Japanese and U.S. residents we studied. We found that the Japanese residents were more likely than the U.S. residents to involve the family in disclosing both a cancer diagnosis and prognosis, and more likely to inform the family first. At the same time, the Japanese residents indicated more uncertainty than did their U.S. counterparts about their approach towards information disclosure.

The findings of our study regarding attitudes about disclosure and the relative importance of the family in decision making by Japanese and U.S. residents are consistent with those in previously published descriptions. Residents' cultural values have been associated with the approach to disclosure and with group versus individual decision making in prior surveys.12,13 Traditionally, in Japan, the individual is perceived as the socially embedded self and as part of several groups ranging from the family to the nation. As a result, groups, not individuals, tend to make decisions.14 The physician frequently accepts the family's final decision regarding whether or not to disclose.15 Families place importance on protecting members from the psychological stress of being diagnosed with cancer.13

In our study a substantial proportion of both Japanese and U.S. residents reported having concealed a cancer diagnosis from a patient, and over three-fourths of each group expressed at least some guilt about it. Among Japanese residents 76 percent reported concealing a cancer diagnosis, a figure consistent with Japan's traditional communication paradigm. The fact that nearly half expressed guilt most of the time and 82% expressed guilt at least some of the time regarding this behavior suggests that a shifting Japanese paradigm toward a more individual autonomy-driven model may be producing increased uncertainty and guilt among Japanese trainees.

A recent survey of Japanese physicians and patients demonstrated major differences in attitudes regarding disclosure, the relative roles of decisionmakers, and use of life-sustaining treatments.16 Other surveys indicate a greater desire for disclosure among Japanese cancer patients themselves, although families are still reluctant to disclose a cancer diagnosis to one another.17–19 Japanese trainees who are seeking to gain skill in how best to communicate diagnosis and prognosis are likely to be caught in the middle of this attitudinal shift, as our study shows.

We note the somewhat counterintuitive finding that the Japanese residents rated the importance of physician involvement in end-of-life decision making lower than did their U.S. counterparts. One might expect that in a culture where physicians are accorded more leeway about what to divulge to patients, the importance of the physician would be rated more highly than in the United States. On the other hand, although we cannot be certain, the responses of the Japanese residents in our study could also indicate that they regard their activities in managing information or making decisions as ones of service to the patient and family rather than as constituting a separate role.

Although hiding a diagnosis was less common among the U.S. residents, 18% of those residents in our sample reported having done so. Blackhall et al.'s20 previous study of African-American, European-American, Korean-American, and Mexican-American elders in Los Angeles demonstrated that many nonwhite respondents were uncomfortable with explicit disclosure and preferred a group decision making process. It may be that residents' reports of withholding information reflects the challenges residents face within Los Angeles' culturally diverse community.

Our findings suggest that both U.S. and Japanese residents could benefit from improved skills in negotiating potentially conflicting values. Given increasing cultural and racial diversity among the aging and seriously ill,21 U.S. medical training programs may be well-served to add curricular elements that address cross-cultural issues surrounding information disclosure. Medical trainees in Japan may benefit from acquiring skills to negotiate changing cultural norms. We have found that experiential approaches including role play, simulated family conferences,22 and film23,24 may offer effective methods to teach skills that require a high level of resident engagement and emotion.

Our study has several limitations. First, our survey respondents represent a convenience sample, although they were drawn from training programs in major Japanese and U.S. institutions. Second, some of the differences we noted (e.g., in certainty) could reflect cultural differences. Third, the survey evaluated the residents' perceptions only, and was unable to evaluate residents' actual behavior. This means we cannot examine the ethical issues involved in actual care being delivered in the two countries. Nevertheless, our findings do suggest a need to understand better why residents perceive conflict in information disclosure, and also to develop methods for helping trainees master skills in communication as part of their training.

In summary, we found that residents' approaches to end-of-life decision making in the United States and Japan reflect known cultural preferences related to the role of patient and family in disclosure of diagnosis and prognosis. We found that many trainees in both Japan and the United States experience uncertainty and report guilt about their approach towards communication as well as towards end-of-life care delivery. Our findings underscore the need for culturally sensitive ethics training in medical education, and a better understanding of how physicians' own cultural perspectives influence their practices.

The authors wish to thank the following local collaborators in Japan for their assistance with survey distribution: Dr. Hiroshi Imura (Aso Iizuka Hospital), Dr. Tetsuhiro Maeno (University of Tsukuba Hospital), Dr. Yuko Y. Takeda (Ryukyu University Hospital), and Dr. Mayuko Saito (St Luke' s International Hospital). An earlier version of this article was presented at the annual meeting of the American Academy of Hospice and Palliative Medicine in January 2005.

Back to Top | Article Outline


1Lo B. Resolving Ethical Dilemmas: A Guide for Clinicians. Philadelphia: Lippincott Williams & Wilkins, 2000, 11.
2Lo B, Quill T, Tulsky J. Discussing palliative care with patients. ACP-ASIM End-of- Life Care Consensus Panel. American College of Physicians/American Society of Internal Medicine. Ann Intern Med. 1999; 130:744–49.
3Novack DH, Plumer R, Smith RL, Ochitill H, Morrow GR, Bennett JM. Changes in physicians' attitudes toward telling the cancer patient. JAMA. 1979;241:897–900.
4Elwyn TS, Fetters MD, Sasaki H, Tsuda T. Responsibility and cancer disclosure in Japan. Soc Sci Med. 2002; 54:281–93.
5Mizushima Y, Kashii T, Hoshino K, et al. A survey regarding the disclosure of the diagnosis of cancer in Toyama Prefecture, Japan. Jpn J Med. 1990;29:146–55.
6Seo M, Tamura K, Shijo H, Morioka E, Ikegame C, Hirasako K. Telling the diagnosis to cancer patients in Japan: attitude and perception of patients, physicians and nurses. Palliat Med. 2000;14:105–10.
7Akabayashi A, Slingsby BT, Kai I. Perspectives on advance directives in Japanese society: a population-based questionnaire survey. BMC Med Ethics. 2003;4:E5.
8Oken D. What to tell cancer patients. A study of medical attitudes. JAMA. 1961;175:1120–28.
9Elwyn TS, Fetters MD, Gorenflo W, Tsuda T. Cancer disclosure in Japan: historical comparisons, current practices. Soc Sci Med. 1998;46:1151–63.
10Horikawa N, Yamazaki T, Sagawa M, Nagata T. Changes in disclosure of information to cancer patients in a general hospital in Japan. Gen Hosp Psychiatry. 2000;22:37–42.
11Takeda F. Status of Cancer Pain and Palliative Care. J Pain Symptom Manage. 2002;24:197–99.
12Matsumura S, Bito S, Liu H, Kahn K, Fukuhara S, Kagawa-Singer M, Wenger N. Acculturation of attitudes toward end of life care: a cross-cultural survey of Japanese Americans and Japanese. J Gen Intern Med. 2002;17:531–39.
13Mizuno M, Onishi C, Ouishi F. Truth disclosure of cancer diagnoses and its influence on bereaved Japanese families. Cancer Nurs. 2002;25:396–403.
14Davis AJ, Konishi E. End of life ethical issues in Japan. Geriatr Nurs. 2000;21:89–91.
15Fukui S. Information needs and the related characteristics of Japanese family caregivers of newly diagnosed patients with cancer. Cancer Nursing. 2002;25:181–86.
16Ruhnke GW, Wilson SR, Akamatsu T, et al. Ethical decision making and patient autonomy: a comparison of physicians and patients in Japan and the United States. Chest. 2000;118:1172–82.
17Long SO. Family surrogacy and cancer disclosure: physician–family negotiation of an ethical dilemma in Japan. J Palliat Care. 1999;15:31–42.
18Matsumura S, Fukuhara S, Kurokawa K, Bito S, Ohki M. Nihonjin no gankokuchi ni kansuru kibou to soreni eikyou wo ataeru shoyouin no kento. [The preferences of disclosure about cancer diagnosis among Japanese and the factors associated with these preferences.] Jpn Med J. 1997;3830:37–43.
19Kakai H. A double standard in bioethical reasoning for disclosure of advanced cancer diagnosis in Japan. Health Commun. 2002;14:361–76.
20Blackhall LJ, Murphy ST, Grank G, Michei V, Azen S. Ethnicity and attitudes toward patient autonomy. JAMA. 1995;274:820–25.
21Addressing diversity. U.S. Department of Health and Human Services, Administration on Aging ( Accessed 11 April 2004.
22Jubelirer SJ, Welch C, Babar Z, Emmett M. Competencies and concerns in end of life care for medical students and residents. W V Med J. 2001;97:118–21.
23Lorenz KA, Steckart MJ, Rosenfeld KE. End of life education using the dramatic arts: the Wit Educational Initiative. Acad Med. 2004;79:481–86.
24Wit Film Project: Innovative medical education for end-of-life care. Inter-Institutional Collaborating Network On End-Of-Life Care (IICN) ( Accessed 11 April 2005.
© 2005 Association of American Medical Colleges