Journal of Cancer Education
ISSN: 0885-8195 (Print) 1543-0154 (Online) Journal homepage: http://www.tandfonline.com/loi/hjce20
A study of the dying process in elderly hospitalized males R. Peter Mogielnicki MD , William A. Nelson PhD & Jennie Dulac BSN, RN To cite this article: R. Peter Mogielnicki MD , William A. Nelson PhD & Jennie Dulac BSN, RN (1990) A study of the dying process in elderly hospitalized males, Journal of Cancer Education, 5:2, 135-145 To link to this article: http://dx.doi.org/10.1080/08858199009528051
Published online: 01 Oct 2009.
Submit your article to this journal
Article views: 1
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hjce20 Download by: [Australian Catholic University]
Date: 10 August 2017, At: 18:03
0885-8195/90 $3.00 + .00 © 1990 American Association for Cancer Education
J. Cancer Education. Vol. 5, No. 2, pp. 135-145, 1990 Printed in the U.S.A. Pergamon Press plc
A STUDY OF THE DYING PROCESS IN ELDERLY HOSPITALIZED MALES R. PETER MOGIELNICKI, MD*; WILLIAM A. NELSON, PhD†;
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
JENNIE DULAC, BSN, RN‡ Abstract—The dying process was studied by questioning nurses and next of kin of 40 consecutive patients who died in an acute care Veterans Hospital. Information regarding problems commonly thought important in the dying process was elicited and attempts were made to relate this information to global assessments of quality of life during the preterminal week and quality of the moments surrounding death. Despite long-standing awareness of many of these problems, important pain, respiratory difficulty, mood problems, blunted alertness, stooling problems, urination problems, and oral intake problems each was present in at least 50% of patients. Multivariate analysis did not define a convincing relationship between these problems and global assessments of quality. Responses of nurses and next of kin were similar most of the time, but nurses systematically rated pain as less severe than next of kin and next of kin systematically demonstrated less awareness of urinary and stooling problems than nurses.
Optimal management of a hospitalized patient in the terminal phase of a fatal illness is one of the most difficult tasks faced by clinicians. The challenge is not a new one, but the demographically-driven bulge in our increasingly elderly population and the AIDS epidemic have both worked to make death more visible to health professionals as well as to the lay public. The situation is further complicated by the increasing ability of medical care providers to sustain life and the escalating costs of such care. In the midst of these issues, the care provider in the hospital struggles with what is the best manner to manage
This study was supported by Veterans Administration Northeast Regional HSR and D Field Program-Project #048. *Professor of Clinical Medicine, Dartmouth Medical School; Chief, Medical Service, White River Junction Veterans Administration Hospital. †Associate Professor of Clinical Psychiatry, Dartmouth Medical School; Chief, Chaplain Service, White River Junction Veterans Administration Hospital. ‡Project Coordinator, Quality of Death Study, Nurse Researcher, Dartmouth Medical School. This work was performed at the White River Junction Veterans Administration Hospital, White River Junction, Vermont. Reprint requests to: Peter Mogielnicki, MD, Medical Service, White River Junction Veterans Administration Hospital, White River Junction, Vermont, 05001.
the medical, emotional, spiritual, and social needs of the terminal patient.1'2 Prompted by several patients whose deaths were accompanied by particularly difficult circumstances, an informal review group began to meet to examine the quality of the deaths of patients at our hospital. In a manner analogous to a conventional "Death and Complications Conference" this "Quality of Death Review Group" reconstructed each case from the medical record and from personal accounts. This was done from the perspective that terminal care focused on the broad needs of a dying person rather than on the technical aspects of diagnosis and management of disease. The group then attempted to understand what could be expected from the hospital staff in facilitating a "good" quality of death. A recurring question during these discussions was what, after all, is a good quality death? The death and dying literature, since Elisabeth Kubler-Ross's landmark writings in the late 1960s can be divided into five main categories.3 The first category contains first-person accounts that describe personal experiences with an illness and care process as the author faced the reality of imminent death.4-5-6 Works in the second category focus on how patients are thought to experience the dying pro-
135
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
136
R. P. MOGIELNICKI et al.
cess. The third category addresses techniques of patient management. For example, Weisman describes what should be done for the terminally ill (especially the cancer patient) to promote "safe conduct" for a better death.12 This is a method of guiding the patient through the emotional distress and discomfort of dying. Cassem, in his instructive chapter in the MGH Handbook on Psychiatry, presents nine features that he believes to be essential for a good death.13 In comparing the management literature to our own experiences, the suggested styles of care seemed challenging for the busy clinician yet intuitively appropriate for a good death. Few articles gave quantitative insight into how often these problems can be expected to occur or, more importantly, how often they can be successfully managed without unsatisfactory trade-offs (ie, successful control of pain at the expense of clouded mental status).14"29 The fourth category of literature contains review articles that describe special circumstances such as children and dying.30-31'32 The last category focuses on the ethical dilemmas that often accompany the process of death.33'40 Except for one excellent study which examined trends in quality of life during the twomonth period prior to death from cancer, there was little to be found in the recent literature that dealt with a global assessment of the dying experience, or that integrated in some way the many aspects and potential trade-offs into some unified expression of the quality of the process.41 Also lacking were systematic descriptions of how frequently the specific problems actually occurred or how frequently the management programs resulted in successful outcomes.42 For these reasons we developed a project to study systematically a series of patients dying at our hospital in order to measure the frequency with which specific problems thought to be important in the dying process were present, and to determine how successfully they were managed. A secondary goal was to attempt to develop a summary technique that might be used to quantitatively define the quality of an individual's death.
METHODS For 40 consecutive patients dying on the acute care services of the White River Junction, Vermont, Veterans Administration Medical Center (WRJ, VAMC) an initial draft questionnaire was used to collect data from two nurses and the next of kin on the circumstances surrounding each patient's death. The information collected was analyzed with particular attention to conflicting responses and to questions that had been interpreted in more than one way. Based on these observations and comments, the questionnaire was extensively rewritten and is available from the authors upon request. The general areas covered by the questions are shown in Table 1. A second series of 40 consecutive patients who died in the hospital at least 72 hours after admission were then studied using this revised instrument. Hospital census sheets were screened each weekday morning, and the head nurse of each ward on which a death had occurred was contacted and asked to identify the two nurses—at either the RN or LPN level — who best knew the patient. Based upon the shift worked, amount of contact with the patient, and nursing degree status, one of these two was designated as the principal nurse for that patient. Within 72 hours of each patient's death, both of these identified nurses were interviewed and the questionnaire was administered by one of the authors (JD), preferably in person
Table 1. Factors examined as indicators of quality of death Oral intake Problems with urine Problems with stool Nausea Breathing difficulty Pain Mental alertness Mood Fear
Psychosocial/spiritual issues Hope Invasive procedures performed Explicitness of code status Communication with caregivers Communication with family/ friends Aloneness Control Global perceived quality of life during week before death Global perceived quality of dying moments
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
Dying process in elderly males
or by telephone. At least two and no more than three weeks thereafter, the next of kin was contacted, and (provided that consent to participate was granted) the questionnaire was administered by the same interviewer by telephone. After the interviewing process was completed, each patient's medical record was audited for demographic data, resuscitation status documentation in doctor's order sheets, and progress note and anesthesia dosing orders. Multiple regression analysis was performed with the assistance of the Dartmouth Medical School Biostatistics Shared Service. RESULTS Descriptive statistics All 40 patients were males. Their ages ranged from 41 to 91 years (average 67.2 years). Forty-five percent were married, 34% were divorced or separated, 10% were widowed and 10% were single. Sixty-five percent were retired, 7.5% were employed, and the remainder were either disabled or unemployed. Sixty percent were Protestant and 35% were Roman Catholic. Duration of hospitalization prior to death was more than 3 but less than 7 days for 20%, between 7 and 14 days for 32.5%, between 15 and 30 days for 30% and over 30 days for 17.5%. The primary diagnosis was cancer in 60% of the patients, disease of the circulatory system in 20%, nonmalignant disease of the respiratory system in 5% and other disease in 15%. Table 2 displays the frequency with which our series of patients experienced the elements postulated to affect the quality of the dying experience. Certain elements such as oral intake problems (vomiting, dysphagia, anorexia, aspiration, other problems sufficient to necessitate restricted oral intake or nasogastric tube placement) or invasive procedures (defined as including intravenous therapy and phlebotomy) were experienced by almost every patient. Others, such as problems with stooling, problems with urination, problems with alertness, problems with mood, fearfulness, not
137
Table 2. Prevalence of selected elements among forty consecutive patients during week before death
Element Oral intake problem Invasive procedures Urination problems (incontinence or catheterized) Lack of alertness Respiratory distress Important pain (score 2-5) Stooling problems (incontinence, constipation, or diarrhea) Mood problems Fearfulness No goodbye said Died alone Nausea Nonexplicit code status
Number (%) in which present 38 (94) 33 (83) 27 (68) 27 (68) 25 (62) 24 (60) 20 (50) 20 (50) 19 (48) 19 (48) 17 (43) 8(20) 2 (5)
having said goodbye to significant others, respiratory difficulty, and important pain were experienced by about half of our patients. Nausea, noted by some authors to be an important symptom in dying patients, was present only 20% of the time in our series. Finally, a nonexplicit code status was present in only 5% of our patients at the time of their deaths. Figure 1 demonstrates the distribution of pain scores assigned by the principal nurse to the 40 patients during the week before death. The question eliciting this information defined pain as that occurring with the patient immobile irrespective of analgesic dose. Thirty percent experienced no pain or unimportant pain (scores 0-1); 60% experienced important pain (scores 2-5); 27.5% experienced moderate to severe pain (scores 3-5); and 20% experienced severe pain (scores 4-5). The distribution of global assessments of the quality of life during the week prior to death as assigned by the principal nurse is shown in Figure 2 and that of global assessment of the quality of dying moments assigned by the same rater is shown in Figure 3. The high frequency of unknown responses in the latter reflects the fact that the principal nurse often was not present at the time of death and, therefore, could not assess the
138
R. P . MOGIELNICKI et al.
40-i c\i CO
30-
in
c\i CM
u c
20-
3
£ ul
in
10-
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
in
Unknown
0 None
Unbearable
Figure 1. Frequency distribution of pain scores assigned by principal nurse for 40 dying patients. Score represents degree of pain experienced by immobile patients during week prior to death. Scores reflect pain experienced, not amount of analgesia needed or used.
40-i
in
cvi CO
o CO
30in
u c
20-
3
s u.
10 in
in
0 1 (Excellent)
2
3 SCORES
4
5 Unknown (As poor as possible)
Figure 2. Frequency distribution of global quality of life during week before death scores assigned by principal nurse.
quality of the moments of death. Assignments of global assessment of quality of life during the week prior to death were unimodal and strongly skewed toward poor quality (mode = score of 4). Assignments of global assessments of the quality of dying moments were bimodal with one peak in the quite good
range (score =1) and another in the as poor as possible range (score = 5). Multiple regression analysis We employed multiple regression analysis to examine the relationships between the spectrum of individual variables explored in the
Dying process in elderly males
139
30CM
a*
o 20-
in
CM
o c « in
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
10 -
0 (Excellent)
1
2
3 SCORES
5 Unknown (As poor as possible)
Figure 3. Frequency distribution of global quality of dying moments scores assigned by principal nurse.
questionnaire and the responses to the two global score questions. This process yielded only three variables in our study with an identifiable relationship to the perceived quality of life during the week before death. Problems with oral intake score, pain score, and respiratory difficulty score, based on the modeling process, appeared to account for approximately 25% of the variation in global quality of life scores. Problems with oral intake dropped the global quality of life score an average of two points; pain and respiratory problems each dropped the score an average of one point. However, since this relationship depended on only two or three patients in the data set, it was felt to be insufficiently strong to represent a firm conclusion of the analysis. Similarly, three variables on initial runs appeared to have some effect on reported quality of the moments surrounding death. Depression, dying alone, and believing that one was going to leave the hospital alive all were associated with lowered scores in this area. However, these associations, too, depended upon only two or three patients in the data set and hence seem tenuous.
Resuscitation status Ninety-five percent of our patients had, at the time of their deaths, an explicit resuscitation status recorded in the doctor's order sheet. For 33 patients this status was "Do not resuscitate should cardiac or respiratory arrest occur (DNR)." For 5 patients, a specific order defining that the patient was to be resuscitated had been written (Not DNR). The average age for the DNR group was 65.4 years and for the Not DNR group was 73.8 years. Interrater concordance As expected, given the nature of many of the questions asked, there was a moderate amount of discordance in the responses of the two nurses and of the next of kin interviewed about each patient. For the most part, this discordance was fairly similar in degree between the two nurses and between the principal nurse and the next of kin. In addition, for the most part, there appeared to be no systematic differences between the two types of respondents. These generalizations were not true, however, in three specific areas. In questions about incontinence of both stool and urine, next of kin proved to be dis-
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
140
R. P. MOGIELNICKI et al.
tinctly more likely to deny the existence of these conditions or, at least, to deny knowledge of their existence. Figure 4 displays the responses to several questions in a series of three-by-three charts which attempt to show the range of agreement and disagreement between the two nurses and between the principal nurse and the next of kin for all patients. Visual inspection reveals similar and symmetric scatter on both sides of concurrence— represented by the left to right downward diagonal in each chart—for responses to questions regarding oral intake trouble and the meeting of religious needs. This was the type of concurrence characteristic of all of the other questions asked as well, except for those regarding incontinence of stool and urine. Most of the time, when a nurse stated that incontinence was present, the next of kin either stated that incontinence was not present or did not know whether or not incontinence was present. The reverse situation—next of kin stating that incontinence was present and nurses denying knowledge of its presence or asserting its absence—was not
YES
NPO
YES RN
In the last decade or two it has become clear that other paradigms in addition to the biomedical one can productively be brought to bear on the care of the sick. Psychosocial, spiritual, behavioral, connexional, and func-
YES
NO
RN
NO
UNK
YES
NO
RN
NO
m
RN
NO
NPO
YES
NO
NO
\ \
UNK N.S.
N.S.
RN
RN
RN
RN YES
• : •
• • .^^
N.S.
N.S.
YES
UNK
UNK
YES
UNK
•
NO
YES
UNK
.-"
YES
.*
UNK
RN
RN
. . .^^ NPO
RELIGIOUS NEEDS MET
STOOL INCONTINENCE
RN
RN NO
DISCUSSION
URINARY INCONTINENCE
ORAL INTAKE PROBLEMS
YES
observed. Chi squared analysis confirmed a statistically significant difference between the responses of nurse and next of kin in comparison to the responses of the nurses at the
ISSN: 0885-8195 (Print) 1543-0154 (Online) Journal homepage: http://www.tandfonline.com/loi/hjce20
A study of the dying process in elderly hospitalized males R. Peter Mogielnicki MD , William A. Nelson PhD & Jennie Dulac BSN, RN To cite this article: R. Peter Mogielnicki MD , William A. Nelson PhD & Jennie Dulac BSN, RN (1990) A study of the dying process in elderly hospitalized males, Journal of Cancer Education, 5:2, 135-145 To link to this article: http://dx.doi.org/10.1080/08858199009528051
Published online: 01 Oct 2009.
Submit your article to this journal
Article views: 1
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hjce20 Download by: [Australian Catholic University]
Date: 10 August 2017, At: 18:03
0885-8195/90 $3.00 + .00 © 1990 American Association for Cancer Education
J. Cancer Education. Vol. 5, No. 2, pp. 135-145, 1990 Printed in the U.S.A. Pergamon Press plc
A STUDY OF THE DYING PROCESS IN ELDERLY HOSPITALIZED MALES R. PETER MOGIELNICKI, MD*; WILLIAM A. NELSON, PhD†;
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
JENNIE DULAC, BSN, RN‡ Abstract—The dying process was studied by questioning nurses and next of kin of 40 consecutive patients who died in an acute care Veterans Hospital. Information regarding problems commonly thought important in the dying process was elicited and attempts were made to relate this information to global assessments of quality of life during the preterminal week and quality of the moments surrounding death. Despite long-standing awareness of many of these problems, important pain, respiratory difficulty, mood problems, blunted alertness, stooling problems, urination problems, and oral intake problems each was present in at least 50% of patients. Multivariate analysis did not define a convincing relationship between these problems and global assessments of quality. Responses of nurses and next of kin were similar most of the time, but nurses systematically rated pain as less severe than next of kin and next of kin systematically demonstrated less awareness of urinary and stooling problems than nurses.
Optimal management of a hospitalized patient in the terminal phase of a fatal illness is one of the most difficult tasks faced by clinicians. The challenge is not a new one, but the demographically-driven bulge in our increasingly elderly population and the AIDS epidemic have both worked to make death more visible to health professionals as well as to the lay public. The situation is further complicated by the increasing ability of medical care providers to sustain life and the escalating costs of such care. In the midst of these issues, the care provider in the hospital struggles with what is the best manner to manage
This study was supported by Veterans Administration Northeast Regional HSR and D Field Program-Project #048. *Professor of Clinical Medicine, Dartmouth Medical School; Chief, Medical Service, White River Junction Veterans Administration Hospital. †Associate Professor of Clinical Psychiatry, Dartmouth Medical School; Chief, Chaplain Service, White River Junction Veterans Administration Hospital. ‡Project Coordinator, Quality of Death Study, Nurse Researcher, Dartmouth Medical School. This work was performed at the White River Junction Veterans Administration Hospital, White River Junction, Vermont. Reprint requests to: Peter Mogielnicki, MD, Medical Service, White River Junction Veterans Administration Hospital, White River Junction, Vermont, 05001.
the medical, emotional, spiritual, and social needs of the terminal patient.1'2 Prompted by several patients whose deaths were accompanied by particularly difficult circumstances, an informal review group began to meet to examine the quality of the deaths of patients at our hospital. In a manner analogous to a conventional "Death and Complications Conference" this "Quality of Death Review Group" reconstructed each case from the medical record and from personal accounts. This was done from the perspective that terminal care focused on the broad needs of a dying person rather than on the technical aspects of diagnosis and management of disease. The group then attempted to understand what could be expected from the hospital staff in facilitating a "good" quality of death. A recurring question during these discussions was what, after all, is a good quality death? The death and dying literature, since Elisabeth Kubler-Ross's landmark writings in the late 1960s can be divided into five main categories.3 The first category contains first-person accounts that describe personal experiences with an illness and care process as the author faced the reality of imminent death.4-5-6 Works in the second category focus on how patients are thought to experience the dying pro-
135
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
136
R. P. MOGIELNICKI et al.
cess. The third category addresses techniques of patient management. For example, Weisman describes what should be done for the terminally ill (especially the cancer patient) to promote "safe conduct" for a better death.12 This is a method of guiding the patient through the emotional distress and discomfort of dying. Cassem, in his instructive chapter in the MGH Handbook on Psychiatry, presents nine features that he believes to be essential for a good death.13 In comparing the management literature to our own experiences, the suggested styles of care seemed challenging for the busy clinician yet intuitively appropriate for a good death. Few articles gave quantitative insight into how often these problems can be expected to occur or, more importantly, how often they can be successfully managed without unsatisfactory trade-offs (ie, successful control of pain at the expense of clouded mental status).14"29 The fourth category of literature contains review articles that describe special circumstances such as children and dying.30-31'32 The last category focuses on the ethical dilemmas that often accompany the process of death.33'40 Except for one excellent study which examined trends in quality of life during the twomonth period prior to death from cancer, there was little to be found in the recent literature that dealt with a global assessment of the dying experience, or that integrated in some way the many aspects and potential trade-offs into some unified expression of the quality of the process.41 Also lacking were systematic descriptions of how frequently the specific problems actually occurred or how frequently the management programs resulted in successful outcomes.42 For these reasons we developed a project to study systematically a series of patients dying at our hospital in order to measure the frequency with which specific problems thought to be important in the dying process were present, and to determine how successfully they were managed. A secondary goal was to attempt to develop a summary technique that might be used to quantitatively define the quality of an individual's death.
METHODS For 40 consecutive patients dying on the acute care services of the White River Junction, Vermont, Veterans Administration Medical Center (WRJ, VAMC) an initial draft questionnaire was used to collect data from two nurses and the next of kin on the circumstances surrounding each patient's death. The information collected was analyzed with particular attention to conflicting responses and to questions that had been interpreted in more than one way. Based on these observations and comments, the questionnaire was extensively rewritten and is available from the authors upon request. The general areas covered by the questions are shown in Table 1. A second series of 40 consecutive patients who died in the hospital at least 72 hours after admission were then studied using this revised instrument. Hospital census sheets were screened each weekday morning, and the head nurse of each ward on which a death had occurred was contacted and asked to identify the two nurses—at either the RN or LPN level — who best knew the patient. Based upon the shift worked, amount of contact with the patient, and nursing degree status, one of these two was designated as the principal nurse for that patient. Within 72 hours of each patient's death, both of these identified nurses were interviewed and the questionnaire was administered by one of the authors (JD), preferably in person
Table 1. Factors examined as indicators of quality of death Oral intake Problems with urine Problems with stool Nausea Breathing difficulty Pain Mental alertness Mood Fear
Psychosocial/spiritual issues Hope Invasive procedures performed Explicitness of code status Communication with caregivers Communication with family/ friends Aloneness Control Global perceived quality of life during week before death Global perceived quality of dying moments
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
Dying process in elderly males
or by telephone. At least two and no more than three weeks thereafter, the next of kin was contacted, and (provided that consent to participate was granted) the questionnaire was administered by the same interviewer by telephone. After the interviewing process was completed, each patient's medical record was audited for demographic data, resuscitation status documentation in doctor's order sheets, and progress note and anesthesia dosing orders. Multiple regression analysis was performed with the assistance of the Dartmouth Medical School Biostatistics Shared Service. RESULTS Descriptive statistics All 40 patients were males. Their ages ranged from 41 to 91 years (average 67.2 years). Forty-five percent were married, 34% were divorced or separated, 10% were widowed and 10% were single. Sixty-five percent were retired, 7.5% were employed, and the remainder were either disabled or unemployed. Sixty percent were Protestant and 35% were Roman Catholic. Duration of hospitalization prior to death was more than 3 but less than 7 days for 20%, between 7 and 14 days for 32.5%, between 15 and 30 days for 30% and over 30 days for 17.5%. The primary diagnosis was cancer in 60% of the patients, disease of the circulatory system in 20%, nonmalignant disease of the respiratory system in 5% and other disease in 15%. Table 2 displays the frequency with which our series of patients experienced the elements postulated to affect the quality of the dying experience. Certain elements such as oral intake problems (vomiting, dysphagia, anorexia, aspiration, other problems sufficient to necessitate restricted oral intake or nasogastric tube placement) or invasive procedures (defined as including intravenous therapy and phlebotomy) were experienced by almost every patient. Others, such as problems with stooling, problems with urination, problems with alertness, problems with mood, fearfulness, not
137
Table 2. Prevalence of selected elements among forty consecutive patients during week before death
Element Oral intake problem Invasive procedures Urination problems (incontinence or catheterized) Lack of alertness Respiratory distress Important pain (score 2-5) Stooling problems (incontinence, constipation, or diarrhea) Mood problems Fearfulness No goodbye said Died alone Nausea Nonexplicit code status
Number (%) in which present 38 (94) 33 (83) 27 (68) 27 (68) 25 (62) 24 (60) 20 (50) 20 (50) 19 (48) 19 (48) 17 (43) 8(20) 2 (5)
having said goodbye to significant others, respiratory difficulty, and important pain were experienced by about half of our patients. Nausea, noted by some authors to be an important symptom in dying patients, was present only 20% of the time in our series. Finally, a nonexplicit code status was present in only 5% of our patients at the time of their deaths. Figure 1 demonstrates the distribution of pain scores assigned by the principal nurse to the 40 patients during the week before death. The question eliciting this information defined pain as that occurring with the patient immobile irrespective of analgesic dose. Thirty percent experienced no pain or unimportant pain (scores 0-1); 60% experienced important pain (scores 2-5); 27.5% experienced moderate to severe pain (scores 3-5); and 20% experienced severe pain (scores 4-5). The distribution of global assessments of the quality of life during the week prior to death as assigned by the principal nurse is shown in Figure 2 and that of global assessment of the quality of dying moments assigned by the same rater is shown in Figure 3. The high frequency of unknown responses in the latter reflects the fact that the principal nurse often was not present at the time of death and, therefore, could not assess the
138
R. P . MOGIELNICKI et al.
40-i c\i CO
30-
in
c\i CM
u c
20-
3
£ ul
in
10-
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
in
Unknown
0 None
Unbearable
Figure 1. Frequency distribution of pain scores assigned by principal nurse for 40 dying patients. Score represents degree of pain experienced by immobile patients during week prior to death. Scores reflect pain experienced, not amount of analgesia needed or used.
40-i
in
cvi CO
o CO
30in
u c
20-
3
s u.
10 in
in
0 1 (Excellent)
2
3 SCORES
4
5 Unknown (As poor as possible)
Figure 2. Frequency distribution of global quality of life during week before death scores assigned by principal nurse.
quality of the moments of death. Assignments of global assessment of quality of life during the week prior to death were unimodal and strongly skewed toward poor quality (mode = score of 4). Assignments of global assessments of the quality of dying moments were bimodal with one peak in the quite good
range (score =1) and another in the as poor as possible range (score = 5). Multiple regression analysis We employed multiple regression analysis to examine the relationships between the spectrum of individual variables explored in the
Dying process in elderly males
139
30CM
a*
o 20-
in
CM
o c « in
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
10 -
0 (Excellent)
1
2
3 SCORES
5 Unknown (As poor as possible)
Figure 3. Frequency distribution of global quality of dying moments scores assigned by principal nurse.
questionnaire and the responses to the two global score questions. This process yielded only three variables in our study with an identifiable relationship to the perceived quality of life during the week before death. Problems with oral intake score, pain score, and respiratory difficulty score, based on the modeling process, appeared to account for approximately 25% of the variation in global quality of life scores. Problems with oral intake dropped the global quality of life score an average of two points; pain and respiratory problems each dropped the score an average of one point. However, since this relationship depended on only two or three patients in the data set, it was felt to be insufficiently strong to represent a firm conclusion of the analysis. Similarly, three variables on initial runs appeared to have some effect on reported quality of the moments surrounding death. Depression, dying alone, and believing that one was going to leave the hospital alive all were associated with lowered scores in this area. However, these associations, too, depended upon only two or three patients in the data set and hence seem tenuous.
Resuscitation status Ninety-five percent of our patients had, at the time of their deaths, an explicit resuscitation status recorded in the doctor's order sheet. For 33 patients this status was "Do not resuscitate should cardiac or respiratory arrest occur (DNR)." For 5 patients, a specific order defining that the patient was to be resuscitated had been written (Not DNR). The average age for the DNR group was 65.4 years and for the Not DNR group was 73.8 years. Interrater concordance As expected, given the nature of many of the questions asked, there was a moderate amount of discordance in the responses of the two nurses and of the next of kin interviewed about each patient. For the most part, this discordance was fairly similar in degree between the two nurses and between the principal nurse and the next of kin. In addition, for the most part, there appeared to be no systematic differences between the two types of respondents. These generalizations were not true, however, in three specific areas. In questions about incontinence of both stool and urine, next of kin proved to be dis-
Downloaded by [Australian Catholic University] at 18:04 10 August 2017
140
R. P. MOGIELNICKI et al.
tinctly more likely to deny the existence of these conditions or, at least, to deny knowledge of their existence. Figure 4 displays the responses to several questions in a series of three-by-three charts which attempt to show the range of agreement and disagreement between the two nurses and between the principal nurse and the next of kin for all patients. Visual inspection reveals similar and symmetric scatter on both sides of concurrence— represented by the left to right downward diagonal in each chart—for responses to questions regarding oral intake trouble and the meeting of religious needs. This was the type of concurrence characteristic of all of the other questions asked as well, except for those regarding incontinence of stool and urine. Most of the time, when a nurse stated that incontinence was present, the next of kin either stated that incontinence was not present or did not know whether or not incontinence was present. The reverse situation—next of kin stating that incontinence was present and nurses denying knowledge of its presence or asserting its absence—was not
YES
NPO
YES RN
In the last decade or two it has become clear that other paradigms in addition to the biomedical one can productively be brought to bear on the care of the sick. Psychosocial, spiritual, behavioral, connexional, and func-
YES
NO
RN
NO
UNK
YES
NO
RN
NO
m
RN
NO
NPO
YES
NO
NO
\ \
UNK N.S.
N.S.
RN
RN
RN
RN YES
• : •
• • .^^
N.S.
N.S.
YES
UNK
UNK
YES
UNK
•
NO
YES
UNK
.-"
YES
.*
UNK
RN
RN
. . .^^ NPO
RELIGIOUS NEEDS MET
STOOL INCONTINENCE
RN
RN NO
DISCUSSION
URINARY INCONTINENCE
ORAL INTAKE PROBLEMS
YES
observed. Chi squared analysis confirmed a statistically significant difference between the responses of nurse and next of kin in comparison to the responses of the nurses at the
