Advertisement

Nationwide Japanese Survey About Deathbed Visions: “My Deceased Mother Took Me to Heaven”

Open AccessPublished:September 19, 2016DOI:https://doi.org/10.1016/j.jpainsymman.2016.04.013

      Abstract

      Objectives

      Primary aim was to clarify the prevalence and factors associated with the occurrence of deathbed visions, explore associations among deathbed visions, a good death, and family depression. Additional aim was to explore the emotional reaction, perception, and preferred clinical practice regarding deathbed visions from the view of bereaved family members.

      Methods

      A nationwide questionnaire survey was conducted involving 3964 family members of cancer patients who died at hospitals, palliative care units, and home.

      Results

      A total of 2827 responses (71%) were obtained, and finally 2221 responses were analyzed. Deathbed visions were reported in 21% (95% CIs, 19–23; n = 463). Deathbed visions were significantly more likely to be observed in older patients, female patients, female family members, family members other than spouses, more religious families, and families who believed that the soul survives the body after death. Good death scores for the patients were not significantly different between the families who reported that the patients had experienced deathbed visions and those who did not, whereas depression was more frequently observed in the former than latter, with marginal significance (20 vs. 16%, respectively, adjusted P = 0.068). Although 35% of the respondents agreed that deathbed visions were hallucinations, 38% agreed that such visions were a natural and transpersonal phenomenon in the dying process; 81% regarded it as necessary or very necessary for clinicians to share the phenomenon neutrally, not automatically labeling them as medically abnormal.

      Conclusions

      Deathbed vision is not an uncommon phenomenon. Clinicians should not automatically regard such visions as an abnormal phenomenon to be medically treated and rather provide an individualized approach.

      Key Words

      Introduction

      Deathbed vision is an end-of-life experience commonly observed throughout the world,
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      End of life experiences and their implications for palliative care.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      but very few systematic studies have been reported in peer-reviewed journals.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      The Grotto painting of the 15th century Assisi tells the story of a dying monk who experienced a vision of the deceased Saint Francis.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      From 1926 to 1977, Osis K. published pioneering studies of deathbed visions in dying patients across many cultures.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      Through several empirical observations, Fenwick P. proposed a concept of transpersonal end-of-life experience and the final meaning of end-of-life experience.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      The former includes deathbed visions (e.g., visions of deceased persons) and deathbed coincidence (e.g., clocks stopping, animal behaviors) and the latter includes a brief recovery of lucidity to help the patients and families settle unresolved matters.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      They stress that, although medical professionals often regard such a phenomenon as meaningless or harmful symptoms, the phenomenon itself has therapeutic value, and thus, it is of importance for clinicians caring for dying patients to understand the nature of deathbed visions.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      End of life experiences and their implications for palliative care.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      In preliminary small-sample studies, deathbed visions were observed in 50%–60%, and were typically associated with feelings of comfort of patients and families.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      Some researchers stress that deathbed visions are different from hallucinations due to delirium because in hallucinations, patients usually see animals or insects and feel anxious or agitated; in deathbed visions, patients see the deceased with comfortable feelings and are often mentally lucid. In some cases, not only patients themselves but also health care professionals or family caregivers see the visions. On the other hand, some studies on families of terminally ill patients with delirium reported that patients sometimes see the deceased during episodes of delirium.
      • Morita T.
      • Akechi T.
      • Ikenaga M.
      • et al.
      Terminal delirium: recommendations from bereaved families' experiences.
      • Namba M.
      • Morita T.
      • Imura C.
      • Kiyohara E.
      • Ishikawa S.
      • Hirai K.
      Terminal delirium: families' experience.
      Existing studies are, however, mainly based on views of health care professionals;
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      and only one study was performed on patients,
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      and three small studies were performed on families.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      A recent Japanese single-center study on 575 bereaved families of patients receiving a home hospice service revealed that 39% reported their loved patients experiencing deathbed visions, termed Omukae in Japanese (literally, someone visiting a dying patient to accompany them on death's journey).
      • Morooka R.
      Deathbed visions in the context of the end-of-life care.
      Understanding what the families experienced through deathbed visions could be of importance, given that pioneering literature suggested that this phenomenon is not uncommon and is a subjectively meaningful event.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      A large study is required to explore the following research questions: how often do deathbed visions occur? What factors contribute to the occurrence of deathbed visions? Is the prevalence of deathbed visions different among locations patients are dying (e.g., hospital vs. home)? How do families and patients feel to see the deathbed visions, comfortable or scared? Are deathbed visions associated with a patient's good death? How do families regard deathbed visions, such as hallucinations or meaningful episodes to prepare for death? Is the experience of deathbed visions associated with family grief outcomes? What is the preferred clinical practice from the view of family members to face deathbed visions?
      The primary aims of this study were 1) to clarify the prevalence and factors associated with the occurrence of deathbed visions on a large bereaved family sample, and 2) to explore the potential association between deathbed visions and a patient's good death and family depression. Additional aim was to explore the emotional reaction, perception, and preferred clinical practice regarding deathbed visions from the view of bereaved family members.

      Subjects and Methods

      This was a nationwide survey of bereaved family members of cancer patients to evaluate quality of end-of-life care across Japan: the Japan Hospice and Palliative care Evaluation study.
      • Miyashita M.
      • Morita T.
      • Hirai K.
      Evaluation of end-of-life cancer care from the perspective of bereaved family members: the Japanese experience.
      A multicenter questionnaire survey was conducted involving bereaved family members of cancer patients who died at hospital, palliative care units, or home. We mailed questionnaires to bereaved families in May 2014, and again in June 2014 to nonresponding families. The completion and return of the questionnaire was regarded as consent to participate in this study, and families who did not want to participate were asked to return the questionnaire with “no reply.” Ethical and scientific validity was confirmed by the institutional review board of all participating institutions.

      Settings and Subjects

      Participating institutions were recruited from those belonging to the Japan Hospice Association. From all 321 certified palliative care units in Japan, 296 belonged to the association; of those, 133 agreed to participate in this survey. There were no national registries of hospitals or home hospice services in Japan; among all 49 hospitals and 51 home hospice services belonging to the association, 20 and 22, respectively, agreed to participate in this survey.
      Primary physicians identified potential participants following the inclusion criteria: 1) bereaved family members of an adult cancer patient (one family member was selected for each patient), 2) aged 20 or older, 3) capable of replying to a self-reported questionnaire, and 4) aware of the diagnosis of malignancy. Exclusion criteria included the following: 1) inability to complete the questionnaire (dementia, cognitive failure, psychiatric illness, language difficulty, or visual loss), 2) treatment-associated death or death in intensive care units, 3) family member unavailable, 4) receiving palliative care services less than 3 days, and 5) no serious psychological distress recognized by the primary physician. The final criterion was, as in our previous studies,
      • Miyashita M.
      • Morita T.
      • Hirai K.
      Evaluation of end-of-life cancer care from the perspective of bereaved family members: the Japanese experience.
      • Morita T.
      • Miyashita M.
      • Yamagishi A.
      • et al.
      Effects of a programme of interventions on regional comprehensive palliative care for patients with cancer: a mixed-methods study.
      • Shinjo T.
      • Morita T.
      • Hirai K.
      • et al.
      Care for imminently dying cancer patients: family members' experiences and recommendations.
      adopted on the assumption that primary physicians could identify families who may suffer a serious psychological impact due to the present study, and no formal criteria or psychiatric screening was applied. Families were surveyed 6–12 months after the patients' deaths.

      Measurements

      The questionnaire was developed by the authors on the basis of a literature review and local preliminary surveys.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      • Morita T.
      • Akechi T.
      • Ikenaga M.
      • et al.
      Terminal delirium: recommendations from bereaved families' experiences.
      • Namba M.
      • Morita T.
      • Imura C.
      • Kiyohara E.
      • Ishikawa S.
      • Hirai K.
      Terminal delirium: families' experience.
      • Morooka R.
      Deathbed visions in the context of the end-of-life care.
      To avoid a lack of clarity for responding families, we focused our questions on deathbed visions, not expanding to broader end-of-life experiences such as deathbed coincidence. We defined deathbed visions as visions of deceased persons or afterlife scenes.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      Afterlife scenes were defined as an afterlife world (celestial landscape, heaven, fields of flowers), the presence of a border (river, tunnel, bridge), God/Buddha, or light.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • van Lommel P.
      • van Wees R.
      • Meyers V.
      • Elfferich I.
      Near-death experience in survivors of cardiac arrest: a prospective study in the Netherlands.
      • Greyson B.
      Dissociation in people who have near-death experiences: out of their bodies or out of their minds?.
      Sensitivity analyses were a priori scheduled for different definitions.

      Deathbed Visions

      We asked the respondents whether the patients did or did not clearly say they saw deceased persons or scenes of the afterlife during the last two weeks. Possible choices included yes (patients themselves clearly stated so), yes (patients did not clearly state so but the family witnessed it), no, and unsure. We regarded deathbed visions as present if families reported either of the former two answers. We specified the periods for family recall as two weeks for clarification. For the respondents who agreed that the patients saw visions of deceased persons, we asked the families to identify them from the mother, father, siblings, a child, spouse, grandmother/grandfather, other family members, friend, pets, or others; for those who agreed that the patients saw afterlife scenes, we asked the families to identify them from scenes of the afterlife, the presence of borders, God/Buddha, the light, or others. Others items were divided into existing items, and responses inconsistent with prepared options were excluded. Additionally, families were asked whether they did or did not talk about their experience to other family members, physicians, nurses, or professional care workers.

      Good Death

      Quality of death and dying was evaluated using the Good Death Inventory.
      • Miyashita M.
      • Morita T.
      • Sato K.
      • et al.
      Good death inventory: a measure for evaluating good death from the bereaved family member's perspective.
      • Hirai K.
      • Miyashita M.
      • Morita T.
      • et al.
      Good death in Japanese cancer care: a qualitative study.
      This was developed to represent important concepts relating to a good death and has 10 subscales: physical and psychological comfort, living in a favorite place, maintaining hope and pleasure, a good relationship with medical staff, not feeling a burden to others, a good relationship with the family, independence, environmental comfort, being respected as an individual, and a feeling of fulfillment at life completion. Reliability and validity were confirmed.
      • Miyashita M.
      • Morita T.
      • Hirai K.
      Evaluation of end-of-life cancer care from the perspective of bereaved family members: the Japanese experience.
      • Morita T.
      • Miyashita M.
      • Yamagishi A.
      • et al.
      Effects of a programme of interventions on regional comprehensive palliative care for patients with cancer: a mixed-methods study.
      • Miyashita M.
      • Morita T.
      • Sato K.
      • et al.
      Good death inventory: a measure for evaluating good death from the bereaved family member's perspective.
      The comfort domain includes three items: free from pain, free from physical distress, and psychological distress. Bereaved family members rate the patient's quality of death and dying in their final place of care using a seven-point Likert-type scale, with higher values indicating a higher quality of death and dying.

      Mental Health of the Bereaved Families

      We measured depression of the bereaved family members using Patient Health Questionnaire (PHQ-9).
      • Kroenke K.
      • Spitzer R.L.
      • Williams J.B.
      The PHQ-9: validity of a brief depression severity measure.
      • Ito M.
      • Nakajima S.
      • Fujisawa D.
      • et al.
      Brief measure for screening complicated grief: reliability and discriminant validity.
      The reliability and validity of the Japanese version of both scales were confirmed, and we adapted a score of 10 to indicate clinical depression.
      • Kroenke K.
      • Spitzer R.L.
      • Williams J.B.
      The PHQ-9: validity of a brief depression severity measure.
      • Ito M.
      • Nakajima S.
      • Fujisawa D.
      • et al.
      Brief measure for screening complicated grief: reliability and discriminant validity.

      Emotional Reaction, Perceptions, and Preferred Clinical Practice

      We also asked the respondents to rate the level of agreement with four statements about emotional reactions of families and patients on a five-point Likert-type scale from 1: strongly disagree to 5: strongly agree (i.e., scared or anxious, comfortable, or reassured). Moreover, we asked the respondents to rate the level of agreement with three statements about family perception of deathbed visions on a five-point Likert-type scale from 1: strongly disagree to 5: strongly agree; and three statements about preferred care from medical professionals on a three-point Likert-type scale from 1: unnecessary to 3: very necessary.
      As covariates, we asked family members to report the following demographic data: patient age; sex; tumor sites; income (low, <8 million yen/year vs. high, ≥8 million yen/year); living area (urban, ≥300,000 population vs. rural, <300,000 population); family age; sex; relationship to patients; the periods family members stayed with the patients during the last week; education (high, university or graduate school vs. low, others); religion; frequency of visiting a temple, church, or other religious places (regularly, often vs. rarely, none); and agreement with the statement that the soul survives the body after death (on a four-point Likert-type scale, disagree to agree).

      Statistical Analyses

      We initially calculated the frequency with 95% CIs for the prevalence of deathbed visions for the total sample. To identify the factors associated with the family-reported occurrence of deathbed visions, we compared demographic data using the univariate logistic regression analysis. Multivariate analyses were not performed.
      The potential association between deathbed visions and patients' good death was assessed by comparing the total score and comfort subscale score of the Good Death Scale between the respondents with and without deathbed visions. Comparisons were performed using Student t-test and adjusted for all background demographic data by regression analyses. The potential association between deathbed visions and family depression was assessed by comparing the prevalence of depression (i.e., PHQ-9 score of 10 or more) and adjusted for all background demographic data by regression analyses.
      To explore emotional reactions, perceptions, and preferred clinical practice, we divided the responses into several categories and calculated the frequency of each item. For additional analyses, we explored sex-related differences regarding the contents of deathbed visions (e.g., whether male patients saw their mothers more frequently than female patients) and demographic factors significantly related to the belief that deathbeds visions were one of the natural and transpersonal phenomena.
      As sensitivity analyses, all analyses were performed on the basis of two definitions: deathbed visions were defined as visions of either deceased persons or afterlife scenes; and defined as visions of only deceased persons. Both analyses obtained essentially the same results, and we presented the results on the basis of the first definition (results of the latter definition are available as Web Materials). In another sensitivity analysis where deathbed visions were regarded as present only when patients themselves clearly stated, the results were essentially the same.
      A P-value of 0.050 was regarded as significant. All analyses were performed using the Statistical Package for the Social Sciences (version 11.0; IBM, Tokyo, Japan).

      Results

      A total of 4440 family members met the inclusion criteria, but 478 were excluded (Fig. 1). We sent 3964 questionnaires, and 2827 (71%) were returned. Of them, 348 refused to reply. For the present study, 147 responses were excluded because of missing data on the primary end points, and 111 responses were excluded because they reported “patients had deathbed visions” but contents reported in the “others” category were inconsistent with the definitions (e.g., insects, funeral/grave, unknown persons, shoulders/war, sea/mountain, and out-of-body experience). We thus analyzed a total of 2221 responses (79% of the obtained data). Background characteristics are summarized in Table 1. There were significant differences in patient age, living area, family sex, relationship to patients, the period family members stayed with the patients, education, religion, and the belief about the soul after death.
      Fig. 1
      Fig. 1Recruitment of participants (number in preintervention survey, number in postintervention survey). ICU = intensive care unit.
      Table 1Respondents' Characteristics (N = 2221)
      Total, n (%)HospitalPalliative Care UnitsHomeP
      Patients
      Age (mean, standard deviation)72.9 (12)70.4 (12)74.3 (11)74.0 (12)<0.001
       Sex
      Male1276 (58)431 (60)315 (55)530 (57)0.15
      Female936 (42)284 (40)259 (45)393 (43)
       Primary tumor sites
      Lung473 (21)145 (20)134 (23)194 (21)0.095
      Liver, bile duct, pancreas454 (20)135 (19)121 (21)198 (22)
      Stomach, esophagus325 (15)98 (14)87 (15)140 (15)
      Colon, rectum267 (12)89 (12)58 (10)120 (13)
      Prostate, kidney, bladder163 (7.3)47 (6.4)43 (7.3)73 (8.0)
      Breast128 (5.8)59 (8.3)24 (4.1)45 (4.9)
      Uterus, ovary109 (4.9)39 (5.5)33 (5.7)37 (4.0)
      Head and neck, brain99 (4.5)31 (4.3)30 (5.2)38 (4.1)
      Blood84 (3.8)34 (4.8)18 (3.1)32 (3.5)
      Others112 (5.0)38 (5.3)33 (5.7)41 (4.5)
       Income
      High1947 (93)632 (93)516 (95)799 (92)0.11
      Low153 (7.3)49 (7.2)30 (5.5)74 (8.5)
       Living area
      Urban951 (48)245 (38)235 (44)471 (57)<0.001
      Rural1045 (52)402 (62)294 (56)349 (43)
      Families
       Age (mean, standard deviation)61.9 (12)62.1 (12)61.7 (12)62.0 (12)0.83
       Sex
      Male681 (31)258 (37)186 (32)237 (26)<0.001
      Female1510 (69)445 (63)391 (68)674 (74)
       Relationship with patients
      Spouse1151 (52)408 (58)267 (46)476 (52)<0.001
      Child724 (33)199 (28)206 (36)319 (35)
      Siblings90 (4.1)37 (5.2)37 (6.4)16 (1.7)
      Parents52 (2.4)21 (3.0)8 (1.4)23 (2.5)
      Others188 (8.5)45 (6.3)62 (11)81 (8.9)
       Stay with patients during the final week
      Four days or more1916 (87)593 (84)466 (81)857 (94)<0.001
      Less than four days282 (13)117 (17)112 (19)53 (5.8)
       Education
      High898 (41)242 (35)234 (41)422 (47)<0.001
      Low1276 (59)459 (66)332 (59)485 (54)
       Religion
      None778 (36)245 (35)226 (40)307 (34)0.013
      Buddhism1224 (56)408 (58)306 (54)510 (56)
      Christianity43 (2.0)5 (0.7)12 (2.1)26 (2.9)
      Others131 (6.0)41 (5.9)28 (4.9)62 (6.9)
       Frequency of visiting religious places
      Often or regularly1514 (69)494 (70)384 (67)636 (70)0.34
      Rarely or never670 (31)212 (30)190 (33)268 (30)
       Belief that the soul survives the body after death
      Rated on a four-point Likert scale from agree to disagree.
      Agree1509 (70)490 (70)372 (66)647 (72)0.041
      Disagree657 (30)207 (30)195 (34)255 (28)
      Some data do not add up to 100% due to missing values. Percentages (numbers) are presented.
      a Rated on a four-point Likert scale from agree to disagree.

      Prevalence of Deathbed Visions

      Deathbed visions were reported in 21% (95% CIs, 19–23; n = 463) in total. Of those, 351 families stated that patients themselves clearly described them, 113 families stated that patients did not clearly state so but the family witnessed them, 1392 families replied no, and the remaining 365 families replied that they were unsure.
      Contents of deathbed visions are summarized in Table 2. Each patient had a median of two contents (range 1–7). Of patients with deathbed visions, 87% had visions of deceased persons, and 54% had visions of afterlife scenes. Among the deceased persons, parents were most frequently listed, followed by siblings and friends. There were no significant sex-related differences in the contents of deathbed visions (data not shown). Families reported that they talked about their experience to other family member (83%, n = 384), while they less frequently talked to physicians (16%, n = 76), nurses (20%, n = 92), and professional care workers (5.0%, n = 23). A total of 12% of the families (n = 54) reported they did not talk about their experience to anyone.
      Table 2Contents of Death Visions (N = 464)
      Contents%n
      Deceased persons87403
       Parents67310
      Mother38177
      Father29133
       Siblings24113
       Child1152
       Spouse8.640
       Grandmother/grandfather1.36
       Other family members4.320
       Friends1676
       Pets1.78
      Afterlife scene54250
       Afterlife world1988
       Presence of border1359
       God/Buddha9.745
       The light7.133

      Factors Associated With Family-Reported Occurrence of Deathbed Visions

      Deathbed visions were significantly more likely to be observed in older patients, female patients, female family members, family members other than spouses, families with more religious activities, and families who believed the soul survive the body after death (Table 3).
      Table 3Factors Associated With Occurrence of Deathbed Visions
      VariablesOdds Ratio95% CIsP
      Death locations
       Hospital1.0 (reference)
       Palliative care units1.00.80–1.40.74
       Home1.10.89–1.40.32
      Patient characteristics
       Patient age
      ≤691.0 (reference)
      70–891.51.2–1.9<0.001
      ≥902.51.6–3.8<0.001
       Patient sex
      Male1.0 (reference)
      Female1.31.0–1.60.020
       Primary tumor sites
      Lung1.0 (reference)
      Stomach, esophagus1.00.73–1.50.85
      Colon, rectum1.10.76–1.60.62
      Liver, bile duct, pancreas1.30.94–1.80.11
      Breast1.20.76–1.90.42
      Prostate, kidney, bladder1.00.65–1.60.96
      Uterus, ovary0.70.40–1.30.25
      Head and neck, brain1.00.61–1.80.86
      Blood0.970.54–1.80.93
      Others1.40.85–2.20.19
       Income
      Low1.0 (reference)
      High1.00.70–1.60.86
       Living area
      Rural1.0 (reference)
      Urban0.90.73–1.10.36
      Family characteristics
       Family age
      ≤491.0 (reference)
      50–600.900.68–1.20.47
      ≥600.760.55–1.10.091
       Family sex
      Male1.0 (reference)1.1–1.70.007
      Female1.4
       Relationship with patients
      Husband/wife1.0 (reference)
      Child1.41.1–1.80.004
      Siblings, parents, others1.51.1–2.00.006
       Stay with patients during the final week
      Less than four days1.0 (reference)0.98–1.90.066
      Four days or more1.4
       Education
      Low1.0 (reference)0.79–1.20.86
      High0.98
       Religion
      None1.0 (reference)
      Buddhism0.880.71–1.10.27
      Christianity0.810.37–1.80.60
      Others1.10.71–1.70.67
       Frequency of visiting religious places
      Rarely or never1.0 (reference)1.2–2.0<0.001
      Often or regularly1.6
       Belief that the soul survives the body after death
      Disagree1.0 (reference)1.2–1.90.001
      Agree1.5

      Association Between Deathbed Visions and Patients' Good Death and Family Depression

      Good death scores were not significantly different between the families who reported the patients experienced deathbed visions and those who did not (Table 3). Depression was more frequently observed in the families who reported the patients experienced deathbed visions compared those who did not, with marginal significance (20%, 95% CIs = 16–25 vs. 16%, 95% CI = 16–19; Table 4).
      Table 4Association Between Deathbed Visions and Patient's Good Death and Family's Depression
      With Deathbed Visions (n = 464)Without Deathbed Visions (n = 1757)PAdjusted P
      Good death
       Total score (mean, standard deviation)4.8 (0.92)4.7 (0.9)0.460.80
       Comfort subscale (mean, standard deviation)5.0 (1.4)5.0 (1.4)0.280.94
      Depression (%, n)20% (n = 92)16% (n = 281)0.0580.068
      Good death was measured using the Good Death Inventory, with a higher score indicating a higher quality of death and dying, ranging from 1 to 7. Depression was measured by PHQ-9. P-values were adjusted by patient age, sex, primary tumor sites, patient income, living area, family age, sex, relationship to patients, periods family members stayed with the patients during the last week, education, religion, frequency of visiting a religious place, and belief that soul survive the body after death.

      Family Emotional Reaction, Perception, and Preferred Care From Medical Professionals

      The proportions of the respondents who reported deathbed visions as causing fear were 19% for patients and 22% for families, whereas the proportions of the respondents who reported deathbed visions as comfortable were 24% for patients and 13% for families (Table 5). Although 35% of the respondents agreed that deathbeds visions were hallucinations due to impaired general conditions, 38% agreed that deathbed visions were one of the natural and transpersonal phenomena (Table 5). Female family members and those with a belief that the soul survives the body after death were significantly more likely to agree that deathbed visions were natural and transpersonal phenomena (male, 26% vs. female, 45%, P < 0.001; belief, 44% vs. nonbelief, 26%, P < 0.001). About 80% of the respondents regarded it as necessary or very necessary for clinicians to share the phenomena neutrally, and use psychotropics if a patient was distressed because of the deathbed visions. Less than 30% of the respondents regarded pastoral care as necessary (Table 5).
      Table 5Family Emotional Reaction, Perception, and Preferred Practical Practice
      Emotional Reaction and PerceptionStrongly Agree or AgreeUnsureDisagree or Strongly Disagree
      Emotional reaction
      Responses of agree or strongly agree on a five-point Likert-type scale.
      % (n)
       Patients
      Seemed to be scared, or anxious19 (88)27 (123)47 (220)
      Seemed to be comfortable or reassured24 (111)40 (185)30 (138)
       Families
      Scared or anxious22 (103)19 (89)52 (240)
      Comfortable or reassured13 (58)34 (159)45 (208)
      Perceptions
      Responses of agree or strongly agree on a five-point Likert-type scale.
       Deathbed visions were hallucinations due to medications25 (114)24 (112)45 (207)
       Deathbed visions were hallucinations due to impaired general conditions35 (162)23 (105)36 (165)
       Deathbed visions were one of natural and transpersonal phenomena in the dying process38 (174)27 (124)30 (140)
      Preferred Practical Practice
      Responses of necessary or strongly necessary on a three-point Likert-type scale.
      UnnecessaryNecessaryVery Necessary
      Share the phenomenon neutrally, not automatically labeling as medically abnormal18 (81)54 (251)24 (111)
      Use psychotropics, if patient was distressed because of deathbed visions12 (57)60 (280)23 (105)
      Coordinate to receive pastoral care66 (304)21 (99)5 (23)
      Percentages with numbers in parentheses are shown.
      a Responses of agree or strongly agree on a five-point Likert-type scale.
      b Responses of necessary or strongly necessary on a three-point Likert-type scale.

      Discussion

      This is, to our knowledge, the first large-scale study on bereaved family members to systemically investigate deathbed visions. The strengths of this study were its large sample size of a nationwide population at home, inpatient hospices, and acute care hospitals; relatively high response rates; and the use of validated measurement tool of good death and depression.
      The first important finding of the present study was clarification of the estimated prevalence of deathbed visions. In this study, deathbed visions were observed in about 20% of the patients. Previous studies demonstrated a variety of prevalence: 43% of clinicians experienced deathbed phenomena per year,
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      a hospice nurse encountered a median of 4.8 patients with deathbed visions per month,
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      88% of terminally ill patients with a median survival of 15 days reported that they had at least one vision or dream related to deceased persons,
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      and 36% of bereaved families reported that dying patients had deathbed visions different from hallucinations.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      In all studies, the core concept of deathbed visions is the same, but target subjects (clinicians, patients, and families), study methods (interview and questionnaire survey), periods of investigation (a day, whole trajectory, and specific periods: two weeks in this study), and operational definition are inconsistent. A direct comparison of existing studies, therefore, is of less value; however, these studies indicate that deathbed visions are not uncommon in the world.
      Whether patients and families are comfortable with deathbed visions is a focus of research.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      End of life experiences and their implications for palliative care.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      Previous studies emphasized that deathbed visions are generally comfortable experiences for patients and families,
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      End of life experiences and their implications for palliative care.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      and some researchers regarded this as a distinct difference between deathbed visions and hallucinations.
      • Mazzarino-Willett A.
      Deathbed phenomena: its role in peaceful death and terminal restlessness.
      The literature suggests that 50%–80% of clinicians agreed with the opinion that deathbed visions offered comfort for patients and that patients who experienced deathbed visions achieved a peaceful death;
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      51% reported being happy to see “visitors”;
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      and visions of deceased were more comforting than visions of living persons for patients themselves.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      In this study, there were no significant differences in good death scores including the comfort subscale, but rather there was a slightly higher tendency of depression in families with experience of deathbed visions. The responses to unvalidated questions directly asking if deathbed visions were feared or comfortable were divisive: feared in about 20% vs. comfortable in about 20% of patients and families. This inconsistency may come from the possibility that families might recall a variety of episodes and not isolate “pure” deathbed visions from other similar episodes, especially delirium.
      • Morita T.
      • Akechi T.
      • Ikenaga M.
      • et al.
      Terminal delirium: recommendations from bereaved families' experiences.
      • Namba M.
      • Morita T.
      • Imura C.
      • Kiyohara E.
      • Ishikawa S.
      • Hirai K.
      Terminal delirium: families' experience.
      • Bruera E.
      • Bush S.H.
      • Willey J.
      • et al.
      Impact of delirium and recall on the level of distress in patients with advanced cancer and their family caregivers.
      • Breitbart W.
      • Gibson C.
      • Tremblay A.
      The delirium experience: delirium recall and delirium-related distress in hospitalized patients with cancer, their spouses/caregivers, and their nurses.
      • Brajtman S.
      The impact on the family of terminal restlessness and its management.
      Depression or grief is a complex process, and a single factor such as a deathbed vision could not explain the development of postbereavement depression.
      • Schulz R.
      • Hebert R.
      • Boerner K.
      Bereavement after caregiving.
      This study nonetheless highlights that deathbed visions are not distressing phenomena for all patients and families, and some regard them as transpersonal phenomena in the dying process, not hallucinations, consistent with previous preliminary studies.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      • Lawrence M.
      • Repede E.
      The incidence of deathbed communications and their impact on the dying process.
      • Brayne S.
      • Farnham C.
      • Fenwick P.
      Deathbed phenomena and their effect on a palliative care team: a pilot study.
      • Brayne S.
      • Lovelace H.
      • Fenwick P.
      End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
      • Kerr C.W.
      • Donnelly J.P.
      • Wright S.
      • et al.
      End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
      • BarBato M.
      • Blunden C.
      • Reid K.
      • Irwin H.
      • Rodriguez P.
      Parapsychological phenomena near the time of death.
      • Kellehear A.
      • Pogonet V.
      • Mindruta-Stratan R.
      • Gorelco V.
      Deathbed visions from the Republic of Moldova: a content analysis of family observations.
      • Muthumana S.P.1
      • Kumari M.
      • Kellehear A.
      • Kumar S.
      • Moosa F.
      Deathbed visions from India: a study of family observations in northern Kerala.
      Clinicians should not automatically regard deathbed visions as abnormal phenomena to be medically treated, and an individualized approach is strongly needed.
      The findings that the contents of deathbed visions were mostly related to deceased persons, not religious figures, and that patients and families were reluctant to talk about this to health care professionals confirmed earlier observations.
      • Fenwick P.
      • Lovelace H.
      • Brayne S.
      Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
      • Fenwick P.
      • Brayne S.
      End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
      Factors associated with the development of deathbed visions identified in this study, such as an older age and stronger religious beliefs, are reasonable, and future studies should confirm these findings.
      This study has several large limitations. First, although we had made maximum efforts to define deathbed visions as clearly as possible, interpretation might be different among respondents. This is, we believe, an acceptable limitation because there is no universally accepted operational definition of deathbed visions. As using different definitions might lead to different results, consensus regarding the operational definition of deathbed visions is strongly needed in future studies. Second, this study involved a bereaved family survey substantially long periods after patient death (six months after), and so recall and proxy bias cannot be avoided. This study design, however, has a unique merit in obtaining family outcomes, such as depression or a validated measure of a patient's quality of death and dying very close to death. This study applied only quantitative study design, and deathbed visions should be further understood through a variety of research methods, including patient interview or ethnography studies. Thirdly, primary physicians identified potential participants, and there might be a selection of family members of deceased patients. This study was performed in a Japanese population, and so generalizability of the findings to other cultures needs caution.
      In conclusion, deathbed visions are not uncommon phenomena. Clinicians should not automatically regard such visions as abnormal, and an individualized approach is needed.

      Disclosures and Acknowledgments

      This study was funded by the Japan Hospice Palliative Care Foundation in Japan. The authors have no conflicts of interest.

      Appendix

      The same analyses were performed on the respondents on the basis of definitions of deathbed visions as visions of only deceased persons. Analyses obtained essentially the same results.
      eTable 1Respondents' Characteristics (N = 2160)
      Total, N (%)HospitalPalliative Care UnitsHomeP
      Patients
      Age (mean, standard deviation)72.9 (12)70.4 (12)74.2 (11)74.1 (12)<0.001
       Sex
      Male1236 (58)418 (60)306 (55)512 (57)0.20
      Female915 (42)278 (40)249 (45)388 (43)
       Primary tumor sites
      Lung461 (21)141 (20)131 (23)189 (21)0.20
      Liver, bile duct, pancreas440 (20)132 (19)117 (21)191 (21)
      Stomach, esophagus310 (14)96 (14)82 (15)132 (15)
      Colon, rectum264 (12)88 (13)58 (10)118 (13)
      Prostate, kidney, bladder157 (7.3)44 (6.3)41 (7.3)72 (8.0)
      Breast127 (5.9)58 (8.3)24 (4.3)45 (5.0)
      Uterus, ovary107 (5.0)37 (5.3)33 (5.9)37 (4.1)
      Head and neck, brain94 (4.4)29 (4.2)27 (4.8)38 (4.2)
      Blood83 (3.9)33 (4.7)18 (3.2)32 (3.6)
      Others110 (5.1)38 (5.5)31 (5.5)41 (4.6)
       Income
      High1891 (93)617 (93)498 (95)776 (91)0.072
      Low148 (7.3)45 (6.8)29 (5.5)74 (8.7)
       Living area
      Urban925 (48)240 (38)228 (44)457 (57)<0.001
      Rural1015 (52)389 (62)285 (56)341 (43)
      Families
       Age (mean, standard deviation)62.0 (12)62.2 (12)61.7 (12)62.1 (12)0.74
       Sex
      Male668 (31)252 (37)182 (33)234 (26)<0.001
      Female1462 (69)432 (63)376 (67)654 (74)
       Relationship with patients
      Husband/wife1117 (52)397 (58)258 (46)462 (52)<0.001
      Child708 (33)194 (28)199 (36)315 (35)
      Siblings89 (4.2)37 (5.4)36 (6.4)16 (1.8)
      Parents49 (2.3)20 (2.9)7 (1.2)22 (2.5)
      Others181 (8.4)43 (6.2)61 (11)77 (8.6)
       Stay with patients during the final week
      Four days or more1858 (87)576 (84)447 (80)835 (94)<0.001
      Less than four days279 (13)115 (17)112 (20)52 (5.9)
       Education
      High877 (42)236 (35)227 (41)414 (47)<0.001
      Low1236 (58)446 (66)320 (59)470 (54)
       Religion
      None756 (36)235 (35)220 (40)301 (34)0.014
      Buddhism1189 (56)399 (59)294 (53)496 (56)
      Christianity42 (2.0)5 (0.7)11 (2.0)26 (2.9)
      Others128 (6.0)41 (6.0)28 (5.1)62 (6.7)
       Frequency of visiting religious places
      Often or regularly1467 (69)478 (70)371 (67)618 (70)0.40
      Rarely or never656 (31)209 (30)184 (33)263 (30)
       Belief that the soul survives the body after death
      Rated on a four-point Likert scale from agree to disagree.
      Agree1462 (70)475 (70)359 (66)628 (72)0.056
      Disagree643 (30)203 (30)189 (34)251 (28)
      Some data do not add up to 100% due to missing values. Percentages (numbers) are presented.
      a Rated on a four-point Likert scale from agree to disagree.
      eTable 2Contents of Death Visions (N = 403)
      Contents%n
      Deceased persons10043
       Parents77310
      Mother44177
      Father33133
       Siblings28113
       Friends1976
       Child1352
       Spouse9.940
       Grandmother/grandfather1.56
       Other family members5.020
       Pets2.08
      Afterlife scene38153
       Afterlife world1666
       Presence of border8.936
       God/Buddha8.434
       The light5.723
      eTable 3Factors Associated With Occurrence of Deathbed Visions
      VariablesOdds Ratio95% CIsP
      Death locations
       Hospital1.0 (reference)
       Palliative care units1.10.76–1.40.90
       Home1.20.90–1.50.27
      Patient characteristics
       Patient age
      ≤691.0 (reference)
      70–891.71.3–2.2<0.001
      ≥902.81.8–4.5<0.001
       Patient sex (female; male as reference)1.41.1–1.70.004
       Primary tumor sites
      Lung1.0 (reference)
      Stomach, esophagus0.920.62–1.30.66
      Colon, rectum1.20.81–1.80.36
      Liver, bile duct, pancreas1.30.93–1.80.13
      Breast1.30.83–2.20.23
      Prostate, kidney, bladder0.950.58–1.50.82
      Uterus, ovary0.720.39–1.30.29
      Head and neck, brain0.900.50–1.70.74
      Blood1.10.57–1.90.87
      Others1.50.89–2.40.13
       Income
      Low1.0 (reference)0.67–1.60.93
      High1.0
       Living area
      Rural1.0 (reference)0.72–1.10.37
      Urban0.9
      Family characteristics
       Family age
      ≤491.0 (reference)
      50–600.920.68–1.20.57
      ≥600.750.53–1.10.10
       Family sex
      Male1.0 (reference)1.0–1.70.022
      Female1.3
       Relationship with patients
      Husband/wife1.0 (reference)
      Child1.51.2–1.90.001
      Others1.51.1–2.10.005
       Stay with patients during the final week
      Less than four days1.0 (reference)
      Four days or more1.30.89–1.80.20
       Education
      Low1.0 (reference)
      High1.00.82–1.30.85
       Religion
      None1.0 (reference)
      Buddhism0.870.69–1.10.24
      Christianity0.820.36–1.90.63
      Others1.10.72–1.80.57
       Frequency of visiting religious places
      Rarely or never1.0 (reference)
      Often or regularly1.61.2–2.0<0.001
       Belief that the soul survives the body after death
      Disagree1.0 (reference)
      Agree1.51.1–1.90.003
      eTable 4Association Between Deathbed Visions and Patient's Good Death and Family's Depression
      With Deathbed Visions (N = 377)Without Deathbed Visions (N = 1626)PAdjusted P
      Good death
       Total score (mean, standard deviation)4.8 (0.9)4.7 (0.9)0.170.68
       Comfort subscale (mean, standard deviation)5.1 (1.4)5.0 (1.4)0.210.84
      Depression (%, n)20% (n = 76)17% (n = 281)0.190.19
      Good death was measured using the Good Death Inventory, with a higher score indicating a higher quality of death and dying, ranging from 1 to 7. Depression was measured by PHQ-9. P-values were adjusted by patient age, sex, primary tumor sites, patient income, living area, family age, sex, relationship to patients, periods family members stayed with the patients during the last week, education, religion, frequency of visiting a religious place, and belief that soul survive the body after death.
      eTable 5Family Emotional Reaction, Perception, and Preferred Practice
      Emotional Reaction and PerceptionStrongly Agree or AgreeUnsureDisagree or Strongly Disagree
      Emotional reaction
      Responses of agree or strongly agree on a five-point Likert-type scale.
      % (n)
       Patients
      Seemed to be scared, or anxious18 (74)26 (105)49 (197)
      Seemed to be comfortable or reassured24 (98)41 (165)29 (116)
       Families
      Scared or anxious21 (85)20 (79)53 (212)
      Comfortable or reassured12 (50)35 (140)45 (182)
      Perceptions
      Responses of agree or strongly agree on a five-point Likert-type scale.
       Deathbed visions were hallucinations due to medications25 (100)24 (98)45 (181)
       Deathbed visions were hallucinations due to impaired general conditions36 (144)23 (91)35 (142)
       Deathbed visions were one of natural and transpersonal phenomena in the dying process38 (153)26 (104)31 (126)
      Preferred Practical Practice
      Responses of necessary or strongly necessary on a three-point Likert-type scale.
      UnnecessaryNecessaryVery Necessary
      Share the phenomenon neutrally, not deciding medically18 (73)53 (215)24 (98)
      Use psychotropics, if a patient was distressed because of deathbed visions12 (49)61 (244)22 (90)
      Coordinate to receive pastoral care66 (266)21 (85)4.5 (18)
      Percentages with numbers in parentheses are shown.
      a Responses of agree or strongly agree on a five-point Likert-type scale.
      b Responses of necessary or strongly necessary on a three-point Likert-type scale.

      References

        • Fenwick P.
        • Lovelace H.
        • Brayne S.
        End of life experiences and their implications for palliative care.
        Int J Environ Stud. 2007; 64: 315-323
        • Mazzarino-Willett A.
        Deathbed phenomena: its role in peaceful death and terminal restlessness.
        Am J Hosp Palliat Care. 2010; 27: 127-133
        • Fenwick P.
        • Lovelace H.
        • Brayne S.
        Comfort for the dying: five year retrospective and one year prospective studies of end of life experiences.
        Arch Gerontol Geriatr. 2010; 51: 173-179
        • Fenwick P.
        • Brayne S.
        End-of-life experiences: reaching out for compassion, communication, and connection-meaning of deathbed visions and coincidences.
        Am J Hosp Palliat Care. 2011; 28: 7-15
        • Lawrence M.
        • Repede E.
        The incidence of deathbed communications and their impact on the dying process.
        Am J Hosp Palliat Care. 2013; 30: 632-639
        • Brayne S.
        • Farnham C.
        • Fenwick P.
        Deathbed phenomena and their effect on a palliative care team: a pilot study.
        Am J Hosp Palliat Care. 2006; 23: 17-24
        • Brayne S.
        • Lovelace H.
        • Fenwick P.
        End-of-life experiences and the dying process in a Gloucestershire nursing home as reported by nurses and care assistants.
        Am J Hosp Palliat Care. 2008; 25: 195-206
        • Kerr C.W.
        • Donnelly J.P.
        • Wright S.
        • et al.
        End-of-life dreams and visions: a longitudinal study of hospice patients' experiences.
        J Palliat Med. 2014; 17: 296-303
        • BarBato M.
        • Blunden C.
        • Reid K.
        • Irwin H.
        • Rodriguez P.
        Parapsychological phenomena near the time of death.
        J Palliat Care. 1999; 15: 30-37
        • Kellehear A.
        • Pogonet V.
        • Mindruta-Stratan R.
        • Gorelco V.
        Deathbed visions from the Republic of Moldova: a content analysis of family observations.
        Omega (Westport). 2011-2012; 64: 303-317
        • Muthumana S.P.1
        • Kumari M.
        • Kellehear A.
        • Kumar S.
        • Moosa F.
        Deathbed visions from India: a study of family observations in northern Kerala.
        Omega (Westport). 2010-2011; 62: 97-109
        • Morita T.
        • Akechi T.
        • Ikenaga M.
        • et al.
        Terminal delirium: recommendations from bereaved families' experiences.
        J Pain Symptom Manage. 2007; 34: 579-589
        • Namba M.
        • Morita T.
        • Imura C.
        • Kiyohara E.
        • Ishikawa S.
        • Hirai K.
        Terminal delirium: families' experience.
        Palliat Med. 2007; 21: 587-594
        • Morooka R.
        Deathbed visions in the context of the end-of-life care.
        Jpn J Palliat Med. 2014; 24 ([in Japanese]): 108-111
        • Miyashita M.
        • Morita T.
        • Hirai K.
        Evaluation of end-of-life cancer care from the perspective of bereaved family members: the Japanese experience.
        J Clin Oncol. 2008; 26: 3845-3852
        • Morita T.
        • Miyashita M.
        • Yamagishi A.
        • et al.
        Effects of a programme of interventions on regional comprehensive palliative care for patients with cancer: a mixed-methods study.
        Lancet Oncol. 2013; 14: 638-646
        • Shinjo T.
        • Morita T.
        • Hirai K.
        • et al.
        Care for imminently dying cancer patients: family members' experiences and recommendations.
        J Clin Oncol. 2010; 28: 142-148
        • van Lommel P.
        • van Wees R.
        • Meyers V.
        • Elfferich I.
        Near-death experience in survivors of cardiac arrest: a prospective study in the Netherlands.
        Lancet. 2001; 358: 2039-2045
        • Greyson B.
        Dissociation in people who have near-death experiences: out of their bodies or out of their minds?.
        Lancet. 2000; 355: 460-463
        • Miyashita M.
        • Morita T.
        • Sato K.
        • et al.
        Good death inventory: a measure for evaluating good death from the bereaved family member's perspective.
        J Pain Symptom Manage. 2008; 35: 486-498
        • Hirai K.
        • Miyashita M.
        • Morita T.
        • et al.
        Good death in Japanese cancer care: a qualitative study.
        J Pain Symptom Manage. 2006; 31: 140-147
        • Kroenke K.
        • Spitzer R.L.
        • Williams J.B.
        The PHQ-9: validity of a brief depression severity measure.
        J Gen Intern Med. 2001; 16: 606-613
        • Ito M.
        • Nakajima S.
        • Fujisawa D.
        • et al.
        Brief measure for screening complicated grief: reliability and discriminant validity.
        PLoS One. 2012; 7: e31209
        • Bruera E.
        • Bush S.H.
        • Willey J.
        • et al.
        Impact of delirium and recall on the level of distress in patients with advanced cancer and their family caregivers.
        Cancer. 2009; 115: 2004-2012
        • Breitbart W.
        • Gibson C.
        • Tremblay A.
        The delirium experience: delirium recall and delirium-related distress in hospitalized patients with cancer, their spouses/caregivers, and their nurses.
        Psychosomatics. 2002; 43: 183-194
        • Brajtman S.
        The impact on the family of terminal restlessness and its management.
        Palliat Med. 2003; 17: 454-460
        • Schulz R.
        • Hebert R.
        • Boerner K.
        Bereavement after caregiving.
        Geriatrics. 2008; 63: 20-22