Deathbed Observations by Physicians and Nurses: A Cross-Cultural Survey

THE JOURNAL OF THE AMERICAN SOCIETY FOR PSYCHICAL RESEARCH VOLUME 71 JULY «> 1977 NUMBER 3 Deathbed Observations by Physicians and Nurses: A Cross-Cultural Survey KARLIS OSIS AND ERLENDUR HARALDSSON |! ABSTRACT: Surveys of deathbed observations were conducted in the United States and in India to replicate the findings of a pilot survey carried out in 1959-60 and to gather more detailed data relevant to the question of post-mortem survival. Physicians and nurses filled in questionnaires and subsequently were interviewed concerning 442 cases in the United States and 435 in India. The most frequently reported phenomenon was that of terminal patients having hallucinations of human figures. The main findings of the pilot survey were confirmed in the present survey in both cultures. Again, four-fifths of the apparitions were ‘‘survival related'’; that is, they portrayed deceased persons and religious figures. This is in sharp contrast to the hallucinations of a normal population. Three out of four apparitions were experienced as having come to take the patients away to a post-mortem modus of existence, to which 72% of them consented. More. patients responded with serenity, peace, and elation (41%) than with negative emotions (29%) to this ostensible invitation to die. The data were analyzed for interaction with various medical, psychological, and cultural factors which could cause or shape hallucinations. In conformity with the survival hypothesis, the deathbed visions were found to be relatively independent of these factors as they were assessed in the population surveyed. INTRODUCTION At times dying patients ‘‘see’’ persons and visionary landscapes which others present do not see. Usually such deathbed visions are 'The American survey was conducted with the support of the late Chester F. Carlson. The greater part of the Indian survey was funded by the James Kidd bequest. We are grateful to our Indian collaborators, particularly Dr. J. Prasad and Mr. P. Dayal, and to our American consultants, especially Dr. Gardner Murphy.

238 Journal of the American Society for Psychical Research interpreted as mere hallucinations which have no basis in external reality. In the early years of psychical research, Myers (1903) and Hyslop (1908) recognized some possibly paranormal elements in a few selected cases of visions of dying patients. Sir William Barrett, a physicist of the Royal College of Science, Dublin, was interested in such cases and presented a number of them in a small book entitled Death-Bed Visions (1926). While some of his cases were carefully observed by physicians and nurses, others are of lesser evidential quality. Barrett was particularly impressed by visions that seemed to mirror some form of contact between patients who were fully rational and cognizant of their physical surroundings and their deceased relatives who, presumably, had passed on to the ‘‘other world.” Often in these cases the ostensible purpose of the deceased was to take the patient away to a post-mortem plane of existence. Barrett emphasized cases in which apparitions ran contrary to the patient's expectations, for example, apparitions of persons the patient thought were still living, but who in fact were dead. In several of Barrett's cases the apparitions were experienced either with exalted feelings or with emotions of serenity and peace. Deathbed visions, Barrett pointed out, often did not conform to cultural stereotypes, e.g., dying children were surprised to see ‘‘angels’’ without wings. Some 30 years later, Barrett's work inspired one of us (K.O.) to systematically study deathbed experiences using modern survey methods and statistical evaluations. In 1959-60, under the auspices of the Parapsychology Foundation, K.O. conducted such a survey. It was the first of its kind and will be referred to in this paper as the **pilot survey.’* The report on the pilot survey was published by the Parapsychology Foundation as a monograph, Deathbed Observations by Physicians and Nurses (Osis, 1961). K.O. found that deathbed visions which appear to be suggestive of post-mortem survival tended to be independent of factors known to cause hallucinations, enhance their occurrence, or influence their content. Medical factors—such as illness predisposing to halluci- natory experiences, high fevers, medication with morphine, etc.—did not seem to generate an increase in the frequency of after-life related experiences. In some instances, such medical factors even appeared to suppress survival-related phenomena. Moreover, personal vari- ables such as the patient's sex, age, education, socio-economic a and religious affiliation also appeared to be of little m ce. e contents of the dying patients’ hallucinations were analyzed and found to be different from those of hallucinations in the general population and the mentally ill. For example, hallucinations of the dying are usually visual, as is the case in most ESP experiences, and rarely auditory, the predominant mode in mentally disturbed d Observations by Physicians and Nurses 239 ns. Terminal patients were reported to have seen apparitions of eased, rather than of the living, two or three times more often do people in the general population. Of all the apparitions of fied persons, 90% were of relatives of the patient; of these, were close relatives: mother, father, spouse, sibling, and off- . This occurs infrequently in the hallucinations of the general . also performed extensive interaction analyses on the data of pilot study. This led to the discovery of many patterns which ed the post-mortem survival hypothesis. He believed, , since most of the findings were post hoc, that without cation in later surveys the weight of the findings would be rather | out to replicate the pilot study. The range of questions was y expanded in order to obtain more information which might pe pport the evidence for post-mortem survival, or provide facts dicting it and thus lend weight to what we have termed the tion hypothesis."’ Information from the pilot survey and i sources was used to formulate a model, or series of hypotheses ing deathbed visions. The model is a bi-polar one which y contrasts two mutually exclusive concepts: the survival hesis and the destruction hypothesis. A i IODEL OF THE TWO BASIC HYPOTHESES OF DEATHBED VISIONS Survival Destruction \is the transition to another Death is the ultimate destruction de of existence. of the personality. A. Sources of Deathbed Visions ___ Extrasensory Perception Sick Brain or Delusions |. Extrasensory awareness of 1. Malfunction of the nervous mate entities, ¢.g., de- system and the dying brain. relatives and religious 2. Schizoid reactions to severe stress. 8 Influence of Hallucinogenic Factors on Deathbed Visions Ind ependent of Medical Factors — Dependent on Medical Factors presence of halluci- 1. The presence of halluci- medical factors will not nogenic medical factors will in-

240 Journal of the American Society for Psychical Research increase the frequency of visions related to post-mortem ex- istence. 2. Conditions detrimental to ESP will decrease the frequency of after-life related phenomena. crease the frequency of halluci- nations related to post-mortem existence—i.e., the more dis- turbed the brain processes, the more numerous the ‘“‘other- world’’ fantasies. 2. ESP not involved. C. Content of Deathbed Visions Perceptions 1. After-life related visions will be relatively coherent, and orien- ted to the situation of dying and the transition to another mode of existence, including ‘‘other- worldly"’ messengers and en- vironments. Hallucinations 1. Hallucinations will portray only memories already stored in the brain and express desires, expectations, and fears of the in- dividual, as well as beliefs char- acteristic of his culture. D. Influence of Psychological Factors on Deathbed Visions Conditions Related to Awareness of an “Other World” 1. Clarity of consciousness and an intact sense of reality will facilitate awareness of an ‘‘other world’’ and its messengers, while states in which contact with external reality is absent will impair such awareness. 2. Patients’ expectation of re- covery or dying will not influence the occurrence of after-life re- lated visions. 3. Presence of stress will not increase the frequency of visions related to an after-life. Conditions Related to Halluci- nations of This World or “Other-World’’ Fantasies 1. Clarity of consciousness and an intact sense of reality will be less conducive to all kinds of hallucinations than states in which contact with reality is absent. 2. Patients’ expectation of re- covery will facilitate this-life hallucinations, while expec- tation of dying will facilitate hallucinations of an after-life. 3. Presence of stress will in- crease the frequency of halluci- nations related to an after-life. by Physicians and Nurses 241 Variability of Content Across Individuals and Cultures Hallucinations 1. Much variability. 2. Hallucinations are purely subjective, They will vary widely young Peal a ics, a04) galearsd and ynamics, background of the individual. METHOD sached physicians and nurses in both the United States ‘two steps: (a) a two-page initial questionnaire was serning the extent and kinds of observations they had patients, and of those who were close to death but individual interviews were held with respondents on of cases reported in the questionnaires and which fell ppe of the survey. re Tait Procedure in the U.S. was conducted between 1961 and 1964 in rk, New Jersey, Connecticut, Rhode Island, and Pennsyl- : second survey was carried out in Northern India during A gh the same basic questionnaire was used in both surveys, ight adjustments were made in the questions asked of the Indian 1 .g., tropical diseases and the Hindu and Moslem ns were covered—in order to accommodate them to the cultural al questionnaire we asked the medical personnel about ations of the following: lucinations of human figures experienced by (a) terminal those not recovering), and (b) by non-terminal patients ho were close to death but recovered). allucinations of surroundings (landscapes, etc.) experienced minal patients, and (b) by non-terminal patients. elevation (sudden rise of mood to elation or serenity) in

242 Journal of the American Society for Psychical Research In the U.S. the questionnaire, with a covering letter, was mailed to a stratified random sample of 2500 physicians and 2500 nurses. Those not responding received another letter asking for a reply. A total of 1004 responses was received. Our Indian consultants advised us not to use the mails to distribute the questionnaire. We therefore worked mainly in large university hospitals. Usually the professor of medicine or professor of surgery arranged meetings with the hospital staff during which we gave a short talk and distributed the questionnaires to be filled out. Practically all the physicians and nurses we approached returned the completed questionnaires (a total of 704). Interviews American respondents who reported pertinent cases were inter- viewed by telephone. In India, telephone contacts had to be replaced by personal interviews, mainly in hospitals but sometimes in the homes of the respondents. We developed three separate follow-up questionnaires for the following types of experience: (a) hallucinations of human figures, (b) hallucinations of surroundings, and (c) mood elevation. Each of these questionnaires consisted of 69 questions used to guide the interview. Open-ended questions were used, e.g., “What was the patient's behavior indicating that he/she was experiencing hallucinations?” Questions proposing a set of alternative answers were also used, e.g., **Was the patient calmed by the hallucination, did he/she become excited, or was there no apparent effect?"’ Questions covered (a) characteristics of the patient such as sex, age, education, religious belief and degree of involvement in it, and belief in an after-life; and (b) medical factors such as diagnosis, medical history, medication, temperature, and clarity of conscious- ness. Additional questions elicited information from the respondents concerning their date of graduation from professional school, degree received, religious beliefs, belief in life after death, and attitudes toward hallucinations. The main part of the questionnaire was devoted to obtaining as many details as possible about the experience reported, ¢.g., how the patient described the hallucination. A total of 877 cases, about evenly divided between the U.S. and India, comprise the main part of the data. Evaluation of the Data The interview data were coded and recorded on computer cards. Each item in the various categories was analyzed for frequency of occurrence. For example, responses to the question, ‘‘What was the primary diagnosis of the illness?’’ were grouped into basic categories eathbed Observations by Physicians and Nurses 243 5 malignancies, cardiovascular disease, respiratory disease, : se. brain disease or injury, etc. Then the data were sub- sree tions. Items describing factors a hat each other were considered jointly: for example, compari- | en the number of patients suffering from diseases known ‘hallucinations (brain diseases, uremia) and the number of uffering from other kinds of diseases were made in terms of quency which the patients hallucinated living persons, ns, etc. Differences were assessed by chi-square statistics. ‘below only probabilities associated with the chi-square We used the significance level of P = .05 (two-tailed). RESULTS We completed a similar number of interviews in the U.S. (442) and in India (435). The vast majority of patients involved were terminally 714). We also had 163 cases of patients who recovered from 1 conditions. Hallucinations of human figures, or seeing . was the type of phenomenon most frequently reported 91 patients). A total of 112 vision cases were primarily of ly abodes, landscapes, gardens, buildings. In 174 cases, did not report seeing anything unusual, but their moods elevated to serenity, peace, elation, or religious emotions. t covers only cases of apparitions of human figures seen by patients (471 cases). wistics of Apparitions Seen by Terminal Patients ‘note .re of terminal patients ‘‘seeing’’ persons not oe cher en t soursetie ty far the largest and most sting part of our data. The sample derives from 216 interviews A respondents and 255 from Indian respondents. ration of apparition. As in most instances of spontaneous ESP, iences were usually of quite brief duration: lasted for five minutes or less, 17% from six to 15 minutes, and 17% for more than an hour (Table 1, Row a). a apparition. The closer in time the apparition was to "s pins (Table 1, Row b), the more frequently it had teristics suggestive of an after-life. The time between seeing pparition and losing consciousness was generally shorter than the between losing consciousness and clinical death. mtity of apparition. The apparitions in our sample portrayed g persons, dead persons, and mythological or historical religious According to the findings of the pilot study, apparitions of ng have nothing to do with post-mortem survival. On the other

244 Journal of the American Society for Psychical Research Table I CHARACTERISTICS OF THE APPARITIONAL EXPERIENCE IN TERMINAL PATIENTS —— —_ereeeeee—eeeeeeeeeeeeeeeeeeeeeeee Number of Cases _— Percentage* Variables Characteristics U.S. India Total U.S. India Total 1 sec.-5 min. 8 83 168 65 38 48 6-15 min, 7 4 © 13 2 17 a. Duration of 16-S9 min. 11 6SO) 61 9 23 18 apparition I hr.-1 day 13 3 44 «(0 Oo Longer + 4 3 os 8 No information %6& 3% 14 -—- — — 0-10 min. 7 636 CSB 9: 4. (2 b. Interval between 11-59 min. 7 3 6 4 2 45 apparition and 1-6 hrs. 2 6 %W 13 2 2 death 7-24 hrs. 2% 41 #6 14 #17) «55 Longer 17) 0 S2 169 OC 8 No information 21 34 —- — Living ” = so 16 20 18 Dead 1 1 6 2 47 e en hey Religious figure 22 93 115 12 48 30 “rer Combination of above 11 7 18 6 4 = § No information 9 8 2 —- — — Male 9 13 102 3 7 57 So hex Female 1% 30 12 6f 23 43 apparition No information 6. 122.3% (= == Taken for visitor 144 2 42 4 4 14 To comfort patient 13 4.17 2 6 To take patient away, with consent 40 102 142 44 SO 47 ce. Purpose of To take patient away, apparition without consent | a < 1 6 8 To send patient back 0 2 2 0 1 1 Threatening 4,300 7 4 6 6 Reliving memories 26 Reh y, Rakes 1 9 No information 18 #S2#10—- — — No effect or relaxation 60 65 125 31 28 3 f. Emotional Serenity 4% 4 & 2 18 2 reactions, Elation % 32 8 2 14 «21 Ist group Negative 33. (91 12417 sD No information 12 2 82- — — No effect or relaxation 60 65 125 31 28 W g- Emotional Negative 33.0 (91 «124 C17 40s reactions, Positive, nonreligious 77 36 113 39 16 27 by Physicians and Nurses 245 Table 1 (Continued) Number of Cases Percentage* Characteristics U.S. India Total U.S. India Total Positive, religious 2 3% 61 #13 16 14 ‘No information a :27 ek do not include cases about which no information was available, Saietetsee cottmae do sot ad up to 100 des to rounding Ott ons of the dead and of religious figures may have ristics sugge of life after death. We termed apparitions gory ‘‘survival-related apparitions,’’ and they comprised cases in the pilot survey. The proportion of survival- ons in the present survey was remarkably similar to t survey: 83% in the U.S, and 79% in India (Table 1, » proportions be characteristic of hallucinations in z persons who are not near death? Fortunately, two of hallucinations experienced by the general ded data for comparison. In the ‘Census of Halluci- dgwick and Committee, 1894) it was reported that only mple had hallucinations similar to those in our survival- ip; D. J. West (1948, p. 190) reported 22% in a ‘‘mass ” survey. (Categories that do not fit our classification ‘cluded in the calculation of these percentages.) Thus at terminal patients in both the pilot and in the present paritions of the dead and of religious figures about ore frequently than the general population sampled in ‘ish studies sroportion of survival-related apparitions in the U.S. and bly stable, the identity of the apparitions experienced ‘persons while indian patients predominantly saw fes (Table 1, Row c; Table 2, Rows a and b). In a s to elsewhere we were able to trace

246 Journal of the American Society for Psychical Research Table 2 IDENTITY OF APPARITIONAL FIGURES* Number of Figures = Percentage** Variables Identity U.S. India Total U.S. India Total Mother 60 16 76 «628 14 23 Father 1S 1 7 14 9 Spouse 49 | a & | 8 18 Sibling 27 iS 42 12 13 413 Offspring Other relatives, previous generation 5 (fan eb! 2 6 4 a. Secular Other relatives, same generation 2 10 12 1 8 4 Other relatives, next tion 0 4 + 0 3 I Unidentified relatives 9 14 23 4 12 7 Friends, acquaintances 21 & 2 10 7 9 Unidentified persons 25 1€62€%&- i-—- i= Totals; 2006=«i78 OE ll ll God or Jesus 13 Pin 30" 3425 £22 28 Shiva, Rama, Krishna 0 13; 6B Oo) 37, 122 Mary, Kali, Durga 5 a 9 16 $ 8 God of death & messengers 0 18 18 Oo 24 17 Angels, Devi, etc. 9 |) ae. ee. same” See | Demons & devils 1 2 3 3 3 3 Other religious figures, unidentified 2 SHS VB - - = Totals: 33 1077 400 0- — — *Totals include cases in which several figures were seen by the same patient. **Percentages do not include cases about which no information was available. Figures in some percentage columns do not add up to 100 due to rounding off. Ninety-one percent of all identified apparitions of persons were relatives of the patient. (In 20%, the identity of the apparition was not ascertained.) Of these, 90% were close relatives, i.e., mother, spouse, offspring, sibling, and father—in that order of frequency (Table 2, Row a), The religious figures were usually described merely as an angel or god, or were unidentified. When identified they were named according to the patient's religion, e.g., no Hindu reported seeing Jesus; no Christian a Hindu deity. Purpose of apparition. Quite often patients told respondents why the apparition had visited them. In 50% of the cases in the pilot survey, the stated purpose of the apparition was to aid patients in their transition to another world: ‘‘to take them away."’ In the present athbed Observations by Physicians and Nurses 247 udy this purpose was reported in 65% of the cases (Table 1, Row e). ther analyses we excluded two somewhat ambiguous cate- : that the figure came ‘‘to comfort’’ the patient, which could either a ‘‘this-life’’ purpose or a “‘take-away"’ purpose, and . where patients were said to be ‘‘reliving memories’’ which, of indicates no contemporary purposes. Thus, after adjusting ata in this way, the take-away purpose is clearly dominant in all . surveys: pilot, 76%; U.S., 69%; India, 79%. ients' response to apparition. In the pilot survey it was found large majority (89%) of the patients who saw apparitions with a away purpose eagerly consented ‘‘to go’’ with them. Although mt was expressed in 72% of the present survey's take-away ;, 28% did not consent, and some patients reacted with fright screams for help. Practically all these negative responses came n patients who refused to consent (Table 1, Row e). In our hed interaction analyses we found that this difference in t between the U.S. and Indian samples may be partly due to fients’ religion and partly to their nationality. nts were said to have reacted to the apparition with noticeable ns in 70% of the cases.? Many of them reacted with positive ns (41%), while a considerable number (29%) had negative ns ticularly in cases where the patient did not consent ‘‘to with the apparition. Of those with positive emotional reactions, were serene and peaceful and half were elated (Table 1, Row f). also asked our respondents to evaluate patients’ positive Is as religious or nonreligious feelings (Table 1, Row g). They that 35% of those positive emotions were of religious erminal patients usually suffer from pain and other kinds of ; consequently their moods are rather depressed. The and serenity that the survival-related apparitions aroused in the patients contrasted sharply against the gloom of dying. gs. Various medical factors are known to increase the likelihood icinatory behavior. Medication consisting of certain analgesics Sedatives, such as morphine and Demarol, might have caused tinations in some of our cases. However, the majority (61%) of 425 patients about whom we have such information had not -d drugs which could cause hallucinations. Half of those who under sedation had received such small doses or such weak ents’ emotional reactions to apparitional experiences were ascertained on an question scale. For our analyses items registering positive emotions were ded together as (a) cither serenity or clation (Table 1. Row f), or (b) either positive igious emotions or positive religious emotions (Table 1, Row g).

248 Journal of the American Society for Psychical Research drugs that the respondents did not consider them to have been psychologically affected. Of the 20% who were influenced, more than half (11%) were said to be only mildly affected. Eight percent were moderately affected and only 1% strongly affected (Table 3, Row d). Thus the evidence indicates that in most cases the apparitional experiences were not drug-induced. Temperature. High body temperature sometimes leads to halluci- nations. Only 8% of the patients ran fevers of over 103 degrees (measured orally) which might have facilitated hallucinatory behavior (Table 3, Row c). Diagnoses. Hallucinations may be associated with injury and diseases of the brain, and with uremic poisoning caused by kidney malfunction, although many brain-injured patients, especially those with strokes, do not hallucinate, Only 12% of the patients in our sample, including stroke cases, had such diagnoses (Table 3, Row a). In addition to primary diagnoses, we also considered secondary illnesses, previous illnesses, and any other factors in the patient's history which might have been hallucinogenic, e.g., alcoholism or mental illness. This measure is rather inflated since we included in it diagnoses only suspected by the physician and diseases which were not active at the time of the terminal illness. We also included the primary diagnoses involving the three hallucinogenic categories discussed above. Only 25% of the patients had secondary diagnoses which could have been hallucinogenic (Table 3, Row b). “‘Hallucinogenic index."’ It was also important to know how many patients might have had at least one of the following possible hallucinogenic factors: drugs, high fever, and primary and/or secondary diagnoses of a hallucinogenic nature. We therefore established a ‘‘hallucinogenic index'’ which includes every patient who had one or more of the above-mentioned indices. It should be noted that we included in this index cases which do not strongly suggest that the patients’ hallucinations were of an abnormal origin, ¢.g., stroke cases, cases in which medication only slightly affected clarity of consciousness, etc. Nevertheless, such indices are present in only 38% of the cases; the majority (62%) are free of them. In the pilot survey, it was found that deathbed visions are relatively unaffected by medical factors. The data from the present survey give the same impression. Clarity of consciousness. We inquired into the clarity of consciousness of the patient at the time of the apparitional experience, a condition which is closely related to medical factors. We had this information for 457 cases. Almost half (43%) the patients were in a normal state of consciousness; they were fully aware of and responsive to their environment, In 29% awareness was mildly impaired, but the respondents could still communicate with their ‘Deathbed Observations by Physicians and Nurses 249 patients. Only 17% were in such a severely impaired state of cor ess that little or no communication was possible. In 11% clarity fluctuated and could not be accurately determined for the times the hallucinations were experienced (Table 3, Row e). Hallucinogenic medical factors are clearly absent in two-thirds of ‘our data. Could these factors nevertheless have affected the ‘remaining third of the patients who were included in the “‘halluci- Table 3 MEDICAL STATUS OF TERMINAL PATIENTS SEEING APPARITIONS eee ———OOOOOoOeOOewe"eneweyeyes=q?=—e—_e_e_eaeaaSSSsSsaa—aaaa—e— ooo Number of Patients Percentage* Variables Medical Status U.S. India Total U.S. India Total Cancer 9). 2 40T 2 eB ss Heart & circulat disease = 6l 39 100 29 16 2 Injury & post- operative 10 62 2 $s 8S 1% a. Primary Respiratory Gagnosis disease o a Aes sera 4a injury /dis- = uremia 28 26 S4 13 il 12 Miscellaneous a6) 2 6§ JEW 2” ode tg No information 4 10 tA re b. Secondary Present 68 «40 = 108 18 «25 diagnosis, Absent 137 187 324 67 8&2 75 hal- No information o. 82D = = = Less than 100° 128 #129 #1257 & S3 S88 s. Body 100°-103° ss 449834 —_— Above 103° 161: ap $36 a” Sets (oral) No information 72 3Sb— | = None o4 6165 sia2S89—sis4Ds—iCTCSC ‘sé Médication, no effect 3990Ctéi«‘«2 79: 201 37" Ig 4. Medication Mildly affected 31 Si) ees! 28 ae affecting con ly sciousness affected 22 10 32 ‘- ‘ : Strongly affected 5 1 6 No information 25 21 4% = — _ Clear 9 100 a8 ” > 2 Mildly impaired Mom) 4 ©. Clarity of con- Severely impaired 36 39 S Te 1S 38 2 SD. 1%. “S$ No information 13 1 “4 —_ _ — | SSeS a oe EES *Percentages do not include cases about which no information was available

250 Journal of the American Society for Psychical Research nogenic index’? Could they have spuriously enhanced the frequency of those characteristics which were found in the pilot study to support the post-mortem survival hypothesis? These characteristics are (a) of survival-related apparitions of dead persons and religious figures, (b) their *‘take-away’’ purpose, and (c) the patients’ appropriate emotional reactions. Cross-tabulations between the “hallucinogenic index’’ and the nature of the apparition (living, dead, or religious figure) show that there was no significant inter- action. The presence of hallucinogenic factors did not increase the frequency of survival-related trends such as apparitions of the deceased, religious figures, or expression of the “‘take-away"’ purpose. Hallucinogenic factors did, however, significantly affect the expected emotional reactions of the patients in the American sample (P = .03). They seemed to su serenity, peace, and religious emotions, and to increase the incidence of negative reactions. This trend is not significant in the Indian sample. We conclude that the medical variables ascertained in the survey seem to be relatively unrelated to the apparitional experiences in terminal patients. Demographic Factors Demographic factors such as age, sex, educational level, and occupation (Table 4, Rows a-d) did not interact significantly with any aspects of the patients’ apparitional experiences. Psychological Factors We analyzed several psychological factors to determine whether they tended to shape the phenomenological aspects of the main phenomena: apparitions of the living, the dead, and religious figures; purpose of the apparitions; and the patients’ emotional reactions to them. Stress. Hallucinations tend to occur in situations of severe stress and social deprivation (Siegel and L. J. West, 1976; L. J. West, 1962). Not only are visits to terminal patients by relatives and friends often restricted, but most such patients are going through very stressful situations, compounded by having to cope with severe pain. Therefore, could their hallucinatory experiences be due to stress rather than to extrasensory awareness of ‘‘visitors’’ from another mode of existence? We attempted to answer this question by evalu- ating an indirect indication of stress found in the data of the patients in our sample—their mood on the day before the hallucination occurred, We assumed that negative moods such as anxiety, anger, or depression would indicate more stress than would positive moods. The least stress, we believed, would be indicated by moods designated by our respondents as ‘‘normal’’ or ‘‘average."* Peathbed Observations by Physicians and Nurses 251 Table 4 CHARACTERISTICS OF TERMINAL PATIENTS SEEING APPARITIONS Oooo Number of Patients Variables Characteristics U.S. India Total U.S. India Total C—O " 1-30 9 68 8 9 2 19 31-50 22 97 119 10 8638 2S a. Age Over 50 174 9 24 81 «3S S No information 1 0 1 — = = Male 99 175 274 46 09 58 b. Sex Female 17-80) 197, SHH sR None, school 13 7 90 ~ f 32 21 pies 57 S89 16 30 2 27 c. Education High school 73 65 138 x” 27 R College Ss 3% 8 M4 6 2 No information 28 16 “4 6S —_- — espe f % 20 8 4 19 3 aun beac mer 9 0 4 +7 #6 17 4 Farmer, gga . se- —— 70 83 153 §2 58 53 No information 8! 103 184 os _ a = 214 214 — 85 48 po ae = 26 26 _ 10 6 Moslem — 12 12 = 5 3 Protestant 97 — 97 Si — 22 © Religion Catholic S. Se wns SB Jewish 12 a 12 6 = 3 Other or none 14 _ 14 7 _ 3 No information 25 3 28 = = No involvement 12 3 1S 8 2 5 ine Sti 27 12 # 18 9 14 4 cena in rel te “4. 2 2 © BM B religion 6406S iid Asi No information 69 127 196 = a Belief 70 139 92 92 92 Poe ‘ ‘a ee ie No belicf 6 6 No information 141 179 no — = — _. oo *Percentages do not include cases about which no information was available. Figures in some percentage columns do not add up to 100 due to rounding off.

252 Journal of the American Society for Psychical Research There were no significant interactions between the patients’ moods on the day prior to the apparitional experiences and what the apparition represented. We also failed to find any appreciable interaction between mood and patients’ emotional reactions to the apparitional experience. The purpose of the apparition was not significantly related to mood in cither the American or the Indian sample taken separately. However, this relationship is significant in the pooled data from both populations (P = .001), and it is in the direction opposite to what would be expected on the hypothesis that stress is a causative factor in apparitional experiences. Patients with normal moods experienced apparitions with a peaceful ‘‘take-away”’ purpose more frequently (54%) than did those who had positive (31%) or negative (27%) moods. From these data we may infer that while the stress experienced by the terminal patients might have caused other kinds of hallucinations, it is unlikely that it affected the incidence of apparitions which expressed purposes related to post- mortem survival, Desires and expectations. A patient's desires, expectations, or ‘wishful thinking’’ might be possible causes of hallucinations. For example, a thirsty traveler in the desert might have the illusion of seeing water when none was there. We ascertained from our respondents the number of patients who had expressed a desire to have a visit from a living person, such as a spouse or a child, and then checked on how many of these persons were later hallucinated. We found only 13 such cases, an insignificant fraction of the total sample. Furthermore, there were no indications in the data to suggest that persons who had recently visited the patient appeared frequently in his hallucinations. Of those visitors, only nine were hallucinated. Fear of dying. \n order to cope with their fear of dying, patients who expect to die might be motivated to hallucinate ‘‘messengers"’ from the after-life. (This would not be the case for patients who expect to recover.) In neither the American nor the Indian sample, however, were the intentions or identity of the survival-related apparitions significantly correlated with the patients’ expectations of living or dying. This is particularly apparent in cases where the patient did not consent to the ‘‘take-away’’ purpose of the apparition and screamed for help. Patients’ emotional reactions to the apparitional experiences also failed to relate significantly to the motivational variables ascertained in the surveys. Cultural Factors We hoped that cross-cultural comparisons would throw light on the hypothesis that some deathbed visions may portray certain aspects of a reality external to the patient. Deathbed Observations by Physicians and Nurses 253 ing to our model we assume that some apparitions may in ee apes independent of the observer. As cultural factors have a more powerful effect on subjective hallucinatory experiences than on observations of external reality, the degree of influence of such factors as religion and belief in post-mortem survival might give a clue as to the true nature of deathbed fone Pen : eligion. Our sample lation consisted mainly of Christians ia) and Hindus (48%) in the U.S. the stratifications were 51% Protestant, 36% Catholic, 6% Jewish, and 7% unaffiliated or belonging to other religions. Eighty-five percent of the Indian patients were Hindu, 10% were Christian, and 5% were Moslem (Table 4, Row e). This elaborate hea pag the scanner proportions among the general popu n the areas , except for the small unaffiliated group in the U.S. This discrepancy disappears if we assume that most of the patients whose affiliations were not reported actually were not affiliated with any religious denomination. Apparently religious affiliation was not a factor in determining the phenomena. The question remains, however, whether the patients’ religion could have determined the important core characteristics of the apparitions. Religion did not significantly influence the purpose or the kind of apparition seen (living, dead, or religious figures). And both the occurrences of survival-related apparitions (of the dead and/or religious figures) and their after-life purposes appear to have transcended the widely divergent religious ideologies of Hindus, Catholics, Protestants, Jews, and Moslems. The patients’ emotional reactions of serenity, peace and ape feelings engendered by apparitional experiences were imilar Catholics and morales the U.S. Unfortunately, we had too few patients of other religions for effective comparison in the American sample. Therefore Catholics and Protestants were compared only with the rest of the patients as a whole—including those who were unaffiliated, those of other religions, and those whose affiliations were not reported. This mixed group showed different emotional reactions such as less serenity (P = .02) and less religions feelings. (P = .06). India min of Christian patients was reported to fe shir bay a oiails and religious emotion than the Hindus did. Part of this difference was traced to the respondent bias of Christian nurses and therefore it cannot be interpreted with reasonable certainty. However, there were more similarities than differences: like Americans, many Hindu patients responded with serenity, ce, and religious emotions. The veal difference between the American and Indian reactions to the apparitional experience lies in the patients readiness to consent

254 Journal of the American Society for Psychical Research to the ‘‘take-away’’ purpose of the apparition: wi exception, all the American patients were fey **to el while 34% of the Indian patients were not. Can this be explained in terms of their differences in religion? There was indeed a difference, though not a significant one: only 16% of the Indian Christians did sok consent ‘‘to go,"’ as compared to 37% of the Hindu patients. It seems ro that this no-consent attitude is due to both national and religious factors. Patients’ involvement in religion did not signifi- cantly affect the nature of the apparition experienced, its sd - Be apy ss if Rp ons to it. ety oe lief in life after death. A patient's belief in life after d Pia for understanding his ways of coping with oie eath, Yet surprisingly few (one third) of our respondents were aware of their patients’ beliefs, or lack thereof, in an after-life. The majority of them reported that they either did not discuss the matter with the — or did not pay enough attention to remember it (68% of the Ce t is remarkable that 12 patients who did not believe in life er death saw apparitions (Table 4, Row g). This, of course, is too small a sample for detailed interaction analysis. We assumed that patients whose beliefs were weak or non-existent might have been more likely to neglect mentioning the matter to their physicians or nurses than would those who had strong convictions. We therefore contrasted this ‘‘no information’’ group with the believers , oe analysis revealed that belief in an after-life has no significant Pept on the frequency of the kind of apparition seen, though it - seem to influence the patient's ostensible communication with it ed patients in the ‘‘believers’’ group than those in the “no ormation’’ group experienced apparitions with a ‘‘take away” purpose and consented “‘to go’’ with them. This difference is significant in both the U.S. (P = .0S) and the Indian (P = .004) sample. There was no such difference in Indian patients who did not consent ‘‘to go."* (Since there was only one ‘‘non-consenter’’ in the American sample, this comparison could not be made.) Belief in life after death did not significantly affect serenity and oe in the American patients, but it did increase such feelings at oe Sacred of —. reactions in the Indian sample (P = .005). In ws. : = “ah “Sar ae positive religious responses number of other variables were ascertained and ana smog them the possibility of respondent bias, as sd above—but they are not discussed here due to lack of space. Reports on phenomena other than hallucinations of human figures were also Pascage and evaluated: visions of scenery, etc., mood elevation ¢ stad before death, and experiences of patients who were near eath but recovered, We hope to report on these elsewhere. Deathbed Observations by Physicians and Nurses 255 DISCUSSION The American and Indian surveys were designed to replicate the findings of the pilot survey and to provide more detailed data bearing on the hypothesis of post-mortem survival. While the pilot survey unearthed many findings which were interpreted as being consistent with the after-life hypothesis, it had severe limitations concerning the statistical certainty of these findings. Because the pilot survey was the first of its kind, previous information was inadequate in helping to predict many of the trends which emerged in that study. Therefore the possibility that such unexpected trends were due to chance variations could not be ruled out. However, most trends in the present survey are reasonably consistent among themselves and with those of the pilot survey. This diminishes the probability of chance as an acceptable explanation. Our model of deathbed experiences related to post-mortem survival assumes (a) that survival-oriented apparitions may to some extent be due to ESP of or from “‘another world’’ (e.g., deceased relatives or religious figures) or (b), if this is not the case, are entirely subjective. Therefore we hypothesized that medical factors which often cause hallucinations but which are not known to affect ESP will not increase the frequency of seeing after-life related apparitions. In all three surveys, the data conformed to this hypothesis. Further- more, we postulated that medical conditions which impede sensory (and, we presume, extrasensory) contact with the external world also reduce the incidence of seeing after-life oriented hallucinations. This was confirmed. We carefully considered psychological factors which might have caused hallucinations. Severe stress, especially in situations of drastically reduced social contact, can find release in hallucinations. Psychiatrists suggest that deathbed visions actually are schizoid episodes through which patients cope with very stressful situations by hallucinating pleasing fantasies of another world. A careful analysis of the data revealed no support for this counterhypothesis. Stress was not significantly related to the core phenomena of deathbed, visions. In both the American and Indian samples, the trend went in the opposite direction: stress tended to reduce the survival-related aspects of these experiences. Patients’ desires, wishes, and expectations also had no significant influence. In a large number of cases, patients experienced apparitions which appeared to be in opposition to their own motivations, though consistent with our hypothesis of post-mortem survival. Some psychiatrists have de- veloped the concept of ‘‘latent motivation’’—motivation which is not expressed verbally or exhibited in behavior. However, this concept has been severely criticized and generally rejected in scientific

256 Journal of the American Society for Psychical Research research. We did not find any definite indices in our da $ one ye" — to the main phenomena of Seated vaieon, a counterhypothesis is cultural childhood and youth, cultural beliefs are Restart sapere epee ways. Could they re-emerge in the visions of the dying—a kind of playing back of old records? The cross-cultural survey in India was Siro vA done with this question in view. Our model assumes that pre and cultural factors will completely shape deathbed » provided they are caused by these factors. However, if they are based on perception of some form of external reality, or ESP glimpses of ‘another world,’ we hypothesized that only modest nectar between cultures would emerge, with the main features smte 4 the same. An analogy could be found by contrasting a ‘yp y American and a typically Indian painting of a mountain: the etails would be quite different while the basic characteristics of gore Nase be clearly recognizable. 4 ¢ found a very close agreement in all three surveys w the frequencies of survival-related apparitions: dead ae eiicos figures versus those of the living. The ostensible ““*take-away”’ pee of the apparition was also equally present. Absence of uences by medical and psychological variables was indicated in all three samples. The core phenomena are the same. Mb isoreivt coloring, however, was present. The sex of the ucinatory figure was largely determined by culturally conditioned eee which, in turn, seem to influence the proportion of allucinations of dead and religious figures. Religion had a comparably slight influence on the main phenomena, though it did, of oo determine the naming of the religious figures. We interpret ese modest cultural differences according to our model: they seem ix renjanithngil perce Ap 9 ay visions are, in part, based on n of some i having sed eabsattie recW Pg of external reality rather than ch culture develops dominant attitudes or values co what is desirable and meritorious to say or do, and one undesirable and degrading. In Western culture, talking about oe contact with the dead is often felt to be undesirable. In spite e a that 27% of an American sample studied by Greeley (1975) pe 4a yes’" to the question ‘‘Have you ever felt that you were y touch with someone who had died?”’ only rarely were such arty cea told to professional people. In a British survey, Rees an contacted 277 widows and 66 widowers in selected localities. Po is number, 94% were suitable for interview. Forty-seven percent that ap reported hallucinations of the presence of a dead Spee one of them discussed their experiences with their doctors, and only one out of 137 did so with a clergyman. The main reason for Deathbed Observations by Physicians and Nurses 257 not discussing the experience was fear of ridicule. It is likely that patients in our survey also had a negative response bias; that is, they avoided telling medical personnel about ‘‘secing’’ apparitions of the dead. If this is so, our sample represents fewer survival-related apparitions than the number actually experienced, Another possible way in which cultural conditioning could shape the data is through respondent bias. The medical observers might have reported what they believed they were supposed to, according to cultural norms, and left out what went against the grain of their particular culture. We found no serious distortion in favor of the after-life hypothesis. On the contrary, we detected some under- reporting of those phenomena which we hypothesize as being related to post-mortem survival. Our data came from interviews with physicians and nurses rather than with the patients themselves. This could introduce a source of bias in reporting and sampling. However, some studies by Moody and Kiibler-Ross are based upon interviews with patients. In Moody's (1975) account of the experiences of resuscitated patients, he states that quite a few of these patients, while in a near-death state, became aware of the presence of deceased relatives as well as what we have called religious figures ‘‘who apparently were there to ease them through their transition to death’’ (p. 43). In a personal communication Ktibler-Ross (1976) has on the basis of her experience with terminal patients confirmed the main characteristics of our own findings: a predominance of survival-related apparitions, their “take-away” purpose, and patients’ reactions of serenity, peace, and religious emotion. The issue of survival after death obviously cannot be assessed solely on the basis of experiences of dying patients. The entire range of other phenomena suggestive of an after-life—such as out-of-body experiences, reincarnation memories, apparitions collectively per- ceived, and certain kinds of mediumistic communications—have to be considered together with the various explanatory hypotheses (other than survival) that have been advanced (see, ¢.g., Hart, 1956, 1959; ~ Murphy, 1961; Roll, 1974; E. M. Sidgwick, 1923; Stevenson, 1974a, 1974b, 1975; Tyrrell, 1953). Noyes (1972), Noyes and Kletti (1972, 1976), and Garfield (1975) have published surveys of cases which involve deathbed experiences characterized by altered states, panoramic memories, and also some phenomena similar to those covered in the present report, but without having ascribed to them an after-life interpretation. Discussion of the full range of data and theories related to the survival question does not fall into the scope of this paper. We conclude our report on the cross-cultural survey of the experiences of dying patients by stating that the main findings are

258 Journal of the American Society for Psychical Research consistent among the three surveys that have been conducted in the United States and in India over a 15-year period. The central tendencies of the data support the after-life hypothesis as it is formulated in the model we outlined briefly earlier in this paper. REFERENCES BARRETT, W. F. Death-Bed Visions. London: Methuen, 1926. GARFIELD, C. Consciousness alteration and fear of death. Journal of Transpersonal Psychology, 1975, 7, 147-175. GREELEY, A.M. Sociology of the Paranormal: A Reconnaissance. Beverly Hills, Calif.: Sage Publications, 1975. Hart, H. Six theories about apparitions. Proceedings of the Society for Psychical Research, 1956, 50, 153-239. HART, H. The Enigma of Survival. Springfield, Ill.: Charles C Thomas, 1959. Hys.op, J. H. Psychical Research and the Resurrection. Boston: Small, Maynard, 1908. KUBLER-ROSS, E. Personal communication, 1976. Moopy. R. A., JR. Life After Life. Atlanta: Mockingbird Books, 1975. Murpny, G. Challenge of Psychical Research. New York: Harper & Row, 1961. Myers, F. W. H. Human Personality and its Survival of Bodily Death. London: Longmans, Green, 1903. 2 vols. NOYES, R., JR. The experience of dying. Psychiatry, 1972, 35, 174-183. NOYES, R.,JR., AND KLETTI, R. The experience of dying from falls. Omega, 1972, 3, 45-52. NOYES, R.,JR., AND KLETTI, R. Depersonalization in the face of life- threatening danger: A description. Psychiatry, 1976, 39, 19-27. Osis, K. Deathbed Observations by Physicians and Nurses. New York: Parapsychology Foundation, 1961. REES, W. D. The hallucinations of widows. British Medical Journal, 1971, 4, 37-41, ROLL, W. G. Survival research: Problems and possibilities. Theta, 1974, 39-40, 1-13. SIDGWICK, E. M. (MRS. H.). Phantasms of the living. . . Proceedings of the Society for Psychical Research, 1923, 33, 23-429. SIDGWICK, H., AND COMMITTEE. Report on the census of hallucina- tions. Proceedings of the Society for Psychical Research, 1894, 10, 25-422. SIEGEL, R. K., AND WeEsT, L. J. (Eps.). Hallucinations: Behavior, Ex- perience and Theory. New York: Wiley, 1975. STEVENSON, |. Twenty Cases Suggestive of Reincarnation. (2nd ed. rev.) Charlottesville: University Press of Virginia, 1974. (a) Deathbed Observations by Physicians and Nurses 259 STEVENSON, I. Xenoglossy: A Review and Report of a Case. Charlottes- ville: University Press of Virginia, 1974. (b) STEVENSON, I. Cases of the Reincarnation Type. Volume I. Ten Cases in India. Charlottesville: University Press of Virginia, 1975. TYRRELL, G. N. M. Apparitions. London: Duckworth, 1953. ein West, D.J. A mass observation questionnaire on hallucinations. Journal of the Society for Psychical Research, 1948, 34, 187-196. West, L. J. (Ep.). Hallucinations. New York: Grune and Stratton, 1962. A.S.P.R. Department of Psychology 5 West 73rd Street University of Iceland New York, N.Y, 10023 Reykjavik, Iceland