What ethical dilemmas are Japanese physicians faced with?

- Atsushi Asai, M.D., Wari Yamamoto, M.D., Tsuguya Fukui, M.D.

Department of General Medicine and Clinical Epidemiology
Kyoto University School of Medicine, Kyoto, Japan
Kyoto University Hospital
Sakyo-ku, Kyoto, 606-01 Japan
Tel/fax: 81-75-751-4246
Email: atsushi@kuhp.kyoto-u.ac.jp

Eubios Journal of Asian and International Bioethics 7 (1997), 162-5.


Each country may face some distinctive ethical problems. Little is known about what kind of ethical problems exist and how often physicians are faced with them in clinical settings in Japan. The authors conducted both retrospective and prospective studies to identify ethical dilemmas at a general medical ward of a university hospital in Japan. In the first phase of the study, retrospective chart reviews were conducted for 61 patients who had been admitted to our general medical ward. It revealed that ethical dilemmas were recorded in 17 (28%) cases. In the subsequent period, as the second phase of the study, 6-month prospective case findings were conducted, showing that 22 (47%) of 47 patients had ethical dilemmas that attending physicians were concerned about. Ethical dilemmas frequently identified included refusal of or unnecessary requests for diagnostic procedures or treatment by patients, issues concerning truth-telling to patients with a serious illness, and disagreement regarding plans for patient care between patients' family and physicians in charge. The prospective case findings identified a significantly higher frequency of ethical dilemmas than did retrospective chart reviews (47% versus 28%, chi-square = 4.1275, P<0.05). Demographics were not significantly different for the two patient groups. The data suggests that ethical dilemmas are not uncommon in general medical wards in Japan. The patients' family played a significant role in clinical decision making, leading on some occasions to complicated ethical dilemmas.


Medical practice is associated with ethical dilemmas and these may vary between countries. The role of family members in the decision making process may also be different from culture to culture. Ethical principles formulating physician's obligations in one country may not necessarily be regarded as appropriate in another. Many ethical problems in clinical settings have been identified in Western countries where patient autonomy is highly prized and the doctrine of informed consent tends to exclude family members from decision making (1-6). In Japan, on the other hand, it is not always clear whether physicians have an obligation to tell patients the truth and obtain informed consent from them. Family members are almost invariably deeply involved in the process of medical decision making, regardless of patient competency (7). Little is known about what kind of ethical problems physicians are faced with, and how often, in clinical settings in Japan. Furthermore, no studies have been done on ethical dilemmas resulting from physician-patient-family relationships in Japan.

We therefore conducted both retrospective and prospective studies to investigate what kinds of ethical dilemmas are commonly encountered and how often in clinical settings in Japan. The prospective study included interviews with physicians in charge of patients to get in-depth qualitative data about the nature of dilemmas. We then compared the results of the retrospective study with these of the prospective study to evaluate the former approach in terms of obtaining an overall view on ethical dilemmas in clinical settings.


The study was conducted in the general medical ward at Kyoto University Hospital. The general ward was opened in December, 1994, with five attending physicians and five residents. We classified ethical dilemmas to the following three categories: problems that are ethically controversial in Japan (e.g. disclosure of a diagnosis of cancer to a patient, appropriateness of informed consent in clinical settings, and medical decisions concerning the end of life.) (7-9); problems that put physicians in quandary in making decisions about patient care because of ethical reasons (e.g. Should a physician accept refusal of recommended treatment by a patient whose competency is doubtful? Should they follow the patient's wishes or those of family members?); problems that a member of a clinical team considers to be ethical dilemmas (e.g. Some physicians were entirely willing to comply with the request of family members not to tell a patient a diagnosis of cancer, but others were not.).

For this study, retrospective chart reviews (Phase 1) and prospective case findings (Phase 2) were conducted. In Phase 1, the medical charts of all patients who were admitted to the general medical ward between December, 1994 and October, 1995 were reviewed to identify and categorize ethical dilemmas recorded by attending physicians. One of the authors (A. A.) , who was not involved in patient care until October, 1995, reviewed all medical charts. In Phase 2, efforts of case findings were done for all patients admitted to the ward between November, 1995 and April, 1996, during morning reports held on a daily basis. When ethical dilemmas were identified, one of the authors interviewed the physician in charge to collect detailed information in order to clarify the qualitative nature of the ethical dilemmas. The phase 2 study was also used to compare the frequency of dilemmas with that of the phase 1 study to determine whether the retrospective chart reviews were as sensitive as prospective case findings for identifying ethical dilemmas. A chart review was not performed in the Phase 2 study.

Dichotomous variables were analyzed using chi-square test or Fisher's exact test. Continuous variables considered normally distributed were analyzed with Student's t-test. Correlations were calculated by using Person's r. A p value less than 0.05 was considered statistically significant.


Sixty-one patients were admitted between December, 1994 and October, 1995 to the general medicine ward and included in the Phase 1. Forty-seven patients were admitted to the ward between November, 1995 and April, 1996 and included in the Phase 2 study. Patient characteristics are shown in Table 1. Demographic (age, gender) and clinical (length of stay, number of patients who died, number of patients with cancer, mental illness, other illnesses) characteristics were not significantly different for the two groups of patients.

In Phase 1, 17 (28%) of the 61 cases involved at least one ethical dilemma. Forty such problems in total were recorded in the charts and eight patients had more than one predicament. In the Phase 2, 22 (47%) of the 47 cases involved at least one ethical dilemma. Fifty-nine dilemmas in total were identified, with 20 (91%) of these cases presenting with more than one. The frequency of ethical dilemmas in the Phase 2 thus differed significantly from that in Phase 1 (47% versus 28%, chi square = 4.13, P<0.05).

The characteristics and frequency of ethical dilemmas for the two phases are described in Table 2. Ethical dilemmas concerning patients' request for diagnostic procedures or regimens which were not medically indicated were identified more frequently in the Phase 2 than in the Phase 1 (8 versus 0, Fisher's exact probability = 0.013). On the other hand, ethical dilemmas regarding truth telling were identified more frequently in the Phase 1 than in the Phase 2 (12 versus 8, P=0.041). In both phases, patients with a malignant disease presented with more ethical dilemmas than those without (3.4 versus 0.4. P<0.05 in the Phase 1 and 4.0 versus 1.0. P<0.05 in the Phase 2 ) In the Phase 2, patients who died during the study period had more ethical dilemmas than those who survived (5.0 versus 1.4, P<0.05). In both phases, the incidence of ethical dilemmas was independent of the patients' age, sex, length of stay, and a diagnosis other than cancer.

The following section describes in detail the ethical dilemmas identified in the Phase 2.

Table 1 : Cases Retrospective Prospective
chart review (phase 1) case finding (phase 2)
Total number of patients 61 47
Patient characteristics
Mean age (Range) 53 years (14 - 90) 56 years (15 - 80)
Male 25 (41%) 20 (47%)
Female 36 (59%) 27 (53%)
Mean length of stay 27 days (1 - 125) 28 days (1 - 150)
Death 1 (2 %) 2 (4%)
Malignancy 5 (8%) 4 (8%)
Principal diagnosis
Gastrointestinal disorders 12 (20%) 10 (21%)
Infectious disorder 8 (13%) 7 (15%)
Psychiatric disorders 8 (13%) 6 (13%)
Neurologic disorder 8 (13%) 4 (9%)
Cardiovascular disorders 5 (8%) 6 (13%)
Respiratory disorders 2 (3%) 1 (2%)
Endocrine disorders 2 (3%) 3 (6%)
Hematologic disorder 2 (3%) 2 (4%)
Connective tissue disorder 2 (3%) 0 (0%)
Urinary tract disorders 1 (2%) 2 (4%)
Unknown 5 (8%) 5 (11%)
Others 7 (11%) 1 (2%)

Table 2: Identified ethical dilemmas and their frequency

Ethical dilemmas Number (frequency)
Retrospective / Prospective*
Refusal of diagnostic procedures or treatment by a patient 7 (18%) 10 (17%)
Request of diagnostic procedures or treatment by a patient# 0 (0%) 8 (14%)
Truth-telling# 12 (30%) 8 (14%)
Request or refusal of diagnostic procedures or treatment by family members 2 (5%) 8 (14%)
Decision about patient competency 3 (8%) 5 (8%)
Medical decisions concerning the end of life 3 (8%) 5 (8%)
Issues about informed consent 6 (15%) 3 (5%)
Allocation of resources 1 (3%) 2 (3%)
Distrust of medicine 3 (8%) 1 (2%)
Others 3 (8%) 9 (15%)
Total 40 59
#: There was a significant difference (P <0.05)
*chart review(phase 1) /case finding (phase 2)

Refusal of diagnostic procedures and treatment by patients

There were six cases of refusal of diagnostic procedures, two cases of attempted discharge against medical advice, and two cases of refusal of treatment. All patients but one were judged competent. One old man in his seventies with pulmonary tuberculosis and disturbed by isolation , and the other, in her eighties with iron deficiency anemia faced with unwanted gastrointestinal work-ups, both attempted to leave the hospital. A young patient who was admitted because of consciousness disturbance due to a conversion disorder refused all work-ups once she became fully conscious. An adolescent patient with long-standing epilepsy and conversion disorder refused to take anticonvulsants because of a strong prejudice against medication.

Ethical dilemmas regarding these refusals were intimately related with the issue of patient competency. It was left undecided whether a patient with a conversion disorder or with chronic epilepsy could be considered competent enough to make decisions. Such refusals were usually accepted when there was no convincing evidence that they were incompetent. The same course was followed in the case of a refusal of work-ups by a patient with an eating disorder.

Requests for diagnostic procedures and treatment by patients

There were five cases of requests for diagnostic procedures and three for treatment. A patient in her forties with abdominal pain with possible somatization disorders continually demanded work-ups including ultrasound and computerized tomography. Another patient in her fifties complaining of chronic pain in her upper and lower extremities demanded detailed muscular and dermatological work-ups. In both cases, the attending physicians were confident that these work-ups were not medically indicated. The physicians were concerned, however, that following their demands might not be ethically wrong, for those who believed that they had a serious disease might have a right to have their undue beliefs cleared up. The attending physicians could not decide whether to override their patients' wishes about work-ups or respect them.

A patient with pulmonary tuberculosis who had once tried to leave the hospital against medical advice asked his physicians to let him stay in the hospital beyond the medically indicated period. When such requests were made by the patients, they were usually met.

Truth telling

There were four patients with cancer, two of whom were not told of the true diagnosis. The issue of non-disclosure of the true diagnosis was found to usually entail a sequence of three stages. First, the physician told members of patient' family about the diagnosis of cancer. Second, the family of the patient requests that the physician not tell the patient the truth. Third, the physician complies with that request. In each case, physicians were distressed by the difficult choice of whether to first tell the diagnosis to the patient or the family members and whether it was ethically appropriate to comply with the request of the family. The physicians in charge felt, however, that they could not override the family's request not to tell the patient the truth. Clinical judgments in these situations were not always unanimous among the physicians involved.

The same dilemma also occurred in the case of a patient with neurological disorders of grave prognosis. Family members of that patient did not want the truth to be disclosed to the patient. The other two patients with cancer were informed of the actual diagnosis; one had no relative except for a demented spouse and the other explicitly wished to know the truth including the true diagnosis and its prognosis.

When the truth was not told, it was impossible for the physicians to obtain informed consent from the patient for subsequent diagnostic procedures and treatment. In such cases, the consent from family members was substituted. Insufficient administration of opioids to alleviate pain due to metastasized cancer was one of the unfavorable results of non-disclosure, since those who did not realize the terminal nature of their illnesses often tried to tolerate the pain out of the fear of becoming dependent on opioids and in the hope that the pain would soon subdue.

Requests for diagnostic procedures or treatment by family members

Eight cases involved requests for diagnostic procedures or treatment by family members. In three cases, including patients with persistent vegetative state (PVS), fever due to viral infection, and Parkinsonism, family members requested the physicians in charge to prolong hospitalization beyond the medically indicated period. In the PVS case, the spouse of the patient, having unreasonable expectations about recovery, asked the physicians to repeat brain function tests. The mother of a patient in his thirties with fever due to viral infection did not accept to treat the patient on an ambulatory basis, insisting on inpatient care. As for the case of Parkinsonism, the family requested prolonged hospitalization to avoid heavy burden on the family at home. The spouse of a patient with intra-abdominal infection, made strong requests for several diagnostic procedures and treatment and tried to intervene in the decision making process by the physicians in charge. The parents of a young patient with personality disorder refused the physician's recommendation to have their child undergo a psychiatric consultation, perhaps out of fear of being stigmatized.

Ethical dilemmas involving intervention from family members were thus encountered frequently regardless of patient competency or age.

Medical decisions concerning the end of life

Two patients died during the Phase 2 study. A patient in his sixties died of systematically metastasized cancer and another patient in her twenties died of sepsis and respiratory failure due to unknown etiology. Physicians who were taking care of these patients decided in advance not to perform cardiopulmonary resuscitation (CPR). The team of physicians in charge discussed the indication of CPR with the family members while the patient with cancer was still competent. The patient was not aware of the diagnosis of metastasized cancer or of the prognosis. A similar discussion regarding CPR for the young woman with sepsis was held when she was intubated and sedated.

The patient with a systematically metastasized cancer was given CPR even though the family and physicians had agreed in advance not to resuscitate at the time of cardiac arrest. When a physician who did not know the code status started CPR, the family asked them to continue and give more oxygen. An on-call physician took over the resuscitation, thinking it futile and not medically indicated. He went on pumping the chest of the dying patient to prolong his life, waiting for his frantic family to accept what was happening. He performed CPR for the sake of the family members, but he could not decide whether the withholding of CPR would have been appropriate or not in this case, because the patient might have been willing to undergo CPR for the sake of his family.

The decision on life-sustaining intervention was also perceived as an ethical dilemma. The cancer patient had also developed severe thrombocytopenia requiring platelet transfusion every three days. The patients would have died sooner of massive bleeding if platelet transfusion had not been continued. Although both the family and the attending physician decided not to administer vasopressors, intubation or CPR, they agreed to continue platelet transfusion until one of the family members came to the hospital. The patient with sepsis and respiratory failure died without administration of CPR, but the artificial ventilation continued after cardiac arrest, in anticipation of the rest of her relatives to come to her bedside.


First, we have to mention several methodological limitations. Our definition of ethical dilemmas might be problematic, because three criteria used here are not necessarily widely accepted as ethical dilemmas in Japan. We focused on problems that physicians are concerned about or agonized over in clinical judgment. Our department had been open for only one year when this study was conducted and had only ten beds at the time of the study. The total number of inpatients in the both phases was small although we included all patients admitted in the study period. Therefore the case-mix of inpatients may differ from that of departments of general medicine in other institutions. It is also possible that some of our patients were more articulate than those being seen in other departments or hospitals, because patients who visit our department sometimes have complicated medical and psychological problems and also have often had "bad" experiences with their previous physicians, failing to build a reliable physician-patient relationship . To obtain more representative data, more studies are needed at many institutions.

Many ethical issues have not yet been resolved in Japan and we do not have universally accepted ethical guidelines for physicians to follow. Attitudes of Japanese patients, family members, and physicians toward truth telling, informed consent, and medical decisions concerning the end of life are likely to vary substantially among the Japanese (10). For these reasons, we concentrated on the identification of ethical dilemmas that physicians are concerned with in the clinical setting. Any attempts to judge what is ethically appropriate or not were avoided.

Our findings revealed that ethical dilemmas were frequently related to family members involved in clinical decision making. Many family members requested diagnostic procedures and treatment for the patient regardless of patient competency or wishes. Dilemmas regarding truth telling have long been one of the most serious ethical issues in Japan (7), and family members were found to play a significant role in such cases. Circumstances concerning disclosure of a diagnosis of cancer would be different if the family members were willing to allow a physician to tell the truth to the patient, or if the physician tried to inform the patient of the true diagnosis right from the start. Disclosure to family members without permission of the patient is accepted as a standard medical practice and rarely regarded as a breach of confidentiality in Japan. Some physicians may be afraid of being blamed by the patient's family who believe that disclosure of the true diagnosis and prognosis has a psychologically devastating effect on a patient resulting in a miserable life at the end. Physicians' general belief that a patient with a serious disease should not know the truth also contributes to this "family first" disclosure in Japan.

Medical decisions concerning the end of life involved several dilemmas. Dilemmas about whether to perform CPR on a moribund patient, or which life-sustaining interventions should be continued and for how long, and the quantity of opioids to be administered were common. Our study also revealed ethical dilemmas related to the wishes of the patient's family. CPR was performed despite previous do-not-resuscitate orders and some life-sustaining procedures were administered to the patient for the sake of the family members. CPR and other life-sustaining interventions were sometimes started and continued based on the wishes of family members until they arrive at the bedside. One study indicated that it is extremely important for family members to be at the bedside at such a time and that the frequency of provision of CPR depended on whether family members had been present at the bedside (11). The patients may well have the same wishes at the end of their life. We therefore cannot judge whether these acts are ethically appropriate or not, because Japanese patients, if asked, might wish for CPR or aggressive intervention to make the family members feel comfortable, even though they know such an intervention would be of no benefit to themselves. Ethical guidelines need to be developed in this context to serve the patient's best interest at the end of life and to avoid unnecessary suffering, taking the patient's desire to meet the wishes of his or her family into consideration.

Disagreement between patients and physicians were not uncommon in our study. It has long been held that Japanese patients entrust physicians with full authority, emulating the authoritarian master-servant-like model (12). Case findings in the Phase 2 , however, showed that there were not a few patients who, not simply obeying a physician's recommendation, wanted their physicians to respect their self-determination. Some physicians expressed their discomfort about such patients in clinical discussion, but seldom recorded their feelings in the medical charts examined in the Phase 1. In many cases, the patients got what they wanted. A stereotypical model of a dependent patient and a paternalistic physician therefore does not fit the physician-patient relationship in our hospital and probably in much of current Japan.

Comparison of the study results of the two phases suggested that prospective case findings are more likely to identify ethical dilemmas than retrospective chart reviews. It is possible that case finding efforts at morning reports make participating physicians more sensitive to ethical dilemmas and encourage them to talk about such problems at the conference. It is also likely that physicians did not record ethical dilemmas even if they had recognized them. Although the studies for the two phases were conducted on different groups of patients, all of the patient characteristics were similar and the respective proportions of different types of dilemmas were similar except for the two issues, truth telling and requests for diagnostic procedures or treatment by family members as described. The illnesses, the length of stay, and the number of deaths were not significantly different. We thus think that the difference in frequency of ethical dilemmas identified between the two phases most likely reflects the differences of the sensitivity of the study method employed, i. e., the higher sensitivity of prospective case findings of the Phase 2.

In summary, our findings suggest that physicians in a general medical ward at a university hospital in Japan are frequently faced with various ethical dilemmas. Some are similar to those in other countries and others unique to Japan. The patients' family plays a significant role in clinical decision making in Japan, sometimes resulting in complicated ethical dilemmas. More studies are needed to clarify what ethical problems Japanese physicians encounter which will help resolve these dilemmas.

1. Wear AN, Brahams. To treat or not to treat: the legal, ethical and therapeutic implications of treatment refusal. JME 17 (1991), 131-135.
2. Lo B, Schroeder SA. Frequency of ethical dilemmas in a medical inpatient Service. Arch Intern Med 141 (1981), 1062-1064.
3. Conn DA, Pace N, Dougall JR, et al. Moral principles in withdrawing basic life support from patients in the persistent vegetative state. British Journal of Hospital Medicine 46 (1991), 391-392.
4. van der Mass PJ, van Delden JJM, Pijnenborg, et al. Euthanasia and other
medical decisions concerning the end of life. Lancet 338 (1991), 669-674.
5. Gordon MG, Singer PA. Decisions and care at the end of life. Lancet 346 (1995), 163-165.
6. Kuczewski MG. Receiving the family: The process of consent in medical decision making. HCR 26 (No.??? 1996) 30-37.
7. Asai A. Should physicians tell patients the truth? West J Med 163 (1995), 36-39.
8. Hattori H, Salzberg SM, Kiang WP et al. The patient's right to information Japan- Legal rules and doctor's opinions. SSM 32 (1991), 1007-1016.
9. Kai I, Gen Ohi, Yano E at al. Communication between patients and physicians about terminal care: A survey in Japan. SSM 36 (1993), 1151-1159.
10. Division of Public Policy, Ministry of Health and Welfare. Report from a survey on Japanese general public regarding their preferences toward terminal care. Tokyo: Division of Public Policy, Ministry of Health and Welfare, 1993.
11. Fukaura A, Tazawa H, Nakajima et al. Do-not-resuscitate orders at a teaching hospital in Japan. NEJM 333 (1995), 805-808.
12. Asai A. Barriers to informed consent in Japan. EJAIB 6 (1996), 91-93.
Go to commentary By Elwyn
Go back to EJAIB 7(6) November 1997
Go back to EJAIB
The Eubios Ethics Institute is on the world wide web of Internet: