In the current study, the authors elucidated 1) prevalence of EOL discussion and CPR/DNAR discussion, 2) participation rate of patients in those discussions, and 3) the factors associated with such participation. To the best of authors’ knowledge, this is the first study to elucidate the factors that relate to patients’ participation in the discussion on CPR/DNAR and EOL disclosure.
Approximately 5% of deceased patients had no chance of EOL discussion (including CPR/DNAR discussion), mostly due to unexpected in-hospital death. Considering that unexpected in-hospital cardiac arrest occurs in a nonnegligible proportion of hospitalized patients,[14] all the patients who admit to the hospital may be eligible for discussion on EOL care in preparation for deterioration. Indeed, the American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care recommends that physicians should initiate a discussion on CPR/DNAR with all the patients admitted for medical and surgical care.[15]
Only one-third of our participants had a chance to participate in their CPR/DNAR discussion, and only a half had a chance to participate in the EOL disclosure. These participation rates are higher than those in the previous studies in Japan, which ranged from 0 to 6%,[10, 12, 13] but lower than in those of the studies from USA, Switzerland, and Finland, which ranged from 37 to 80%.[7–9, 16–19] The definitions of EOL discussions vary among the studies and simple comparison is difficult, however, there are several possible reasons. First, the rate of patients’ participation increases as the time of the survey becomes more recent.[10–13] The rise in the participation rate probably reflects the growing interest in ACP. Second, the participation rate can be influenced by the legal background and health care system of the belonging society. Patient involvement ought to be higher in the societies where advance care planning and/or advance directive are mandatory in a certain situation (e.g. USA[20] and Taiwan[21, 22]). Third, cultural perspectives and patients' (and family’s) preferences that stem from them can be influential. Authentic Japanese culture lay value in ambiguity rather than explicitness compared to Western societies. For example, research on the concept of “good death” indicated that many Japanese do not want to know the seriousness of their condition. While 96% of USA patients emphasized that “knowing what to expect about one’s physical condition” helps them achieve a “good death”,[23] only 50% − 69% of Japanese patients agreed to such conception. A nationwide survey demonstrated that the majority of Japanese general population considered “dying without awareness that one is dying” as an important factor to achieve a good death, and approximately half of the general population considered that “not being informed of bad news” is an important issue during the last days of life.[24]
Decreased consciousness hampered approximately one third of the patients from participation in the CPR/DNAR discussion. The discussion should have been initiated earlier for these patients. In principle, ACP should be better started earlier than later, especially in patients who have life-threatening diseases[2, 4], however, actual timing of ACP is more complex since it is influenced by many factors such as patient's preferences, readiness, and medical conditions.[1, 2, 15] Using routine assessments, such as the Advance Care Planning Readiness Scale (ACPRS),[25] may help promote ACP. Paladino, et al demonstrated that a systematic intervention comprising training of clinicians based on a manual (the Serious Illness Conversation Guide: SICG), family materials, and system changes (patient identification using the “surprise question”, email reminders, and documentation templates on ACP) resulted in the improved implementation of serious illness conversations.[26]
Approximately 40% of the patients did not participate in their first discussion due to the judgment by their treating physicians. According to an international study that surveyed physicians’ attitudes toward patient autonomy, 82% of the Japanese palliative-care physicians agreed that patients should be informed first (rather than family first) of their serious medical condition,[27] however, the current study elucidated that, in reality, physicians tend to talk to the family first.
Medical professionals’ readiness, in addition to patients’ and families’ readiness, matters in initiating ACP. Medical professionals need to have the necessary skills and show an openness to talk about diagnosis, prognosis, death, and dying with individuals and their families.[1] Chu et al demonstrated that the implementation of palliative care training for healthcare providers and a program to routinely implement ACP for hospitalized patients with chronic life-limiting illness is feasible and successful in Taipei.[21] Providing physicians with appropriate educational opportunities and training, such as end-of-life care, psychological support for patients and their families, and communication skills training is essential.
The current study elucidated the factors that relate to patients’ participation in the discussion on CPR/DNAR and EOL disclosure. Whether a patient him/herself participates in the discussion on CPR/DNAR is mostly influenced by sociodemographic factors rather than medical factors.
Male patients are more likely to be involved in a discussion on CPR/DNAR. There have been only a few studies that examined the association between gender and patients’ involvement in CPR/DNAR discussion. In a multisite registry study in the United States, Perman et al. reported that women are more likely than men to establish DNAR.[28] Other studies, including a study that enrolled hospitalized older adults who required a surrogate decision-maker in the United States[16] and a Taiwanese study that enrolled cancer patients,[29] demonstrated no significant gender difference. We speculate that there are some paternalistic perspectives in Japan that autonomy is more emphasized among men than women, while women need to be “protected” from serious medical information. A further large-scale, multisite survey is needed to uncover potential mechanisms of gender difference.
The patients who live alone are more likely to participate in CPR/DNAR discussion. Probably this is due to a practical reason that patients living alone lack clear proxy decision-makers. Without a DNAR discussion, there is a risk of providing unnecessary life-sustaining treatment in case of sudden changes in their physical conditions and causing discomfort to the patient due to the treatments against their wishes.
The patients who had one year or more from the diagnosis were more likely to participate in CPR/DNAR discussion. This result was consistent with the previous Japanese study.[30] The longer the course of the illnesses is, the more prepared the patients become for their future (including death). Trust between the patient and the treating physician may be cultivated during the course of the illness. The treating physicians have more opportunities to understand the patient's background, personality, and sense of value, which makes sensitive discussion with the patient easier.
The patients who died of cancer were more likely to participate in EOL disclosure than patients who died of non-cancer illnesses. This is probably because the course of non-cancer diseases, such as heart failure and chronic respiratory diseases are generally less predictable than cancer.[31] However, clinical practice such as using of the “surprise question” - a simple question for clinicians to ask themselves “Would I be surprised if this patient died in the next 12 months?” - can help identify patients at high risk of death who might benefit from palliative care services.[32] This surprise question has been shown to have good predictability in patients with cancer,[33] decompensated heart failure,[34] and end-stage chronic kidney disease.[35]
The patients with a mental illness were less likely to participate in EOL disclosure. Various mental conditions were included in the current study, and they were roughly classified into the following three categories; psychological distress (depression and anxiety), serious mental illnesses (schizophrenia, bipolar disorder, and psychotic depression), and cognitive disorders (e.g. dementia). Several studies showed that accurate prognostic understanding is associated with elevated depression and anxiety,[36, 37] thus it is well-understandable that clinicians feel afraid that telling their patient that he or she is at his EOL stage may worsen his or her mental conditions. However, since EOL discussions are associated with less aggressive medical care near death and early hospice referrals,[9] and there is some evidence that most metastatic cancer patients want detailed prognostic information,[38] disclosing accurate prognostic information while minimizing psychological distress of patients is a challenging but important issue.[39] Clinicians may consider that patients with serious mental illnesses or cognitive disorders lack decision-making capacity and are not eligible for EOL discussions. The presence of serious mental illness and cognitive disorders does not directly mean that a patient lacks decision-making capacity, clinicians should try their best to let the patient involved in decision making while considering patients’ mental capacity.[40]
The current study has a few limitations. First, since this was a retrospective, chart review study, some potential factors that may influence patients’ participation in CPR/DNAR and EOL disclosure, such as the patient’s decision-making capacity, were not examined. Second, since this was a single-center study with moderate sample size, the generalizability of the results is limited. The study site was an urban acute-care hospital without a palliative care unit, and a substantial proportion of patients were transferred to another hospital or transited to a home-based hospice program at their EOL. The discussion on CPR/DNAR and EOL care may be held in such places. Third, the quality of the chart documentation may differ depending on the doctor who wrote it. Also, undocumented covert discussions between clinicians and patients were not detectable.