- Email: [email protected]
Dying in Hospital
Vol. 22 No. 6 December 2001
Journal of Pain and Symptom Management
1035
Original Article
Dying in Hospital: Medical Failure or Natural Outcome? Sue Middlewood, RN, BA, Glenn Gardner, RN, PhD, and Anne Gardner, RN, MPH Palliative Care Services (S.M.), The Canberra Hospital; and The Canberra Hospital and the University of Canberra Research Center for Nursing Practice (G.G., A.G.), Woden, Australian Capital Territory, Australia
Abstract The purpose of this study was to describe patterns of medical and nursing practice in the care of patients dying of oncological and hematological malignancies in the acute care setting in Australia. A tool validated in a similar American study was used to study the medical records of 100 consecutive patients who died of oncological or hematological malignancies before August 1999 at The Canberra Hospital in the Australian Capital Territory. The three major indicators of patterns of end-of-life care were documentation of Do Not Resuscitate (DNR) orders, evidence that the patient was considered dying, and the presence of a palliative care intention. Findings were that 88 patients were documented DNR, 63 patients’ records suggested that the patient was dying, and 74 patients had evidence of a palliative care plan. Forty-six patients were documented DNR 2 days or less prior to death and, of these, 12 were documented the day of death. Similar patterns emerged for days between considered dying and death, and between palliative care goals and death. Sixty patients had active treatment in progress at the time of death. The late implementation of end-of-life management plans and the lack of consistency within these plans suggested that patients were subjected to medical interventions and investigations up to the time of death. Implications for palliative care teams include the need to educate health care staff and to plan and implement policy regarding the management of dying patients in the acute care setting. Although the health care system in Australia has cultural differences when compared to the American context, this research suggests that the treatment imperative to prolong life is similar to that found in Americanbased studies. J Pain Symptom Manage 2001;22:1035–1041 © U.S. Cancer Pain Relief Committee, 2001. Key Words Terminal care, end-of-life decisions, dying in hospital
Introduction Much literature has emanated from North America regarding the care of dying patients within the acute care hospital setting. The
Address reprint requests to: Sue Middlewood, RN, BA, Palliative Care Services, The Canberra Hospital, PO Box 11, Woden, ACT 2606, Australia. Accepted for publication: March 7, 2001. © U.S. Cancer Pain Relief Committee, 2001 Published by Elsevier, New York, New York
common theme of this literature is that dying patients receive inadequate and inappropriate care because their needs do not fit with the ideology of large hospitals, where the main purpose of care is to cure using increasingly sophisticated interventions.1–4 In Australia, the Commonwealth Government has had a commitment to improving patient access to palliative care, whether that be in the hospital, community, or inpatient hos0885-3924/01/$–see front matter PII S0885-3924(01)00362-1
1036
Middlewood et al.
pice setting.5 In South Australia, the proportion of cancer deaths in metropolitan hospitals fell from 40% in 1981 to 28% in 1990.6 Cancer deaths in hospices over the same time period increased from 5% to 20%. The proportion of people dying at home (14%) remained the same. The remaining cancer patients died in country hospitals or nursing homes. A similar proportion of cancer deaths occurred at home in Victoria: 19% in 1988. Sixty percent of South Australians surveyed replied that in the event of a terminal illness they would prefer to die at home.6 Despite preference for home, the actual place of death is more likely to be a large hospital. Although Lickiss and colleagues5 note that palliative care units within teaching hospitals have the opportunity to act as agents of change, little research has been done on the management of dying patients in acute care settings in Australia. At present, studies involving observation of the last six days of life are underway in two teaching hospitals in South Australia. Findings suggest that “. . . the hospital environment reflects the business of life in society, which still denies the naturalness and inevitability of death.”7 Although Australian interest in end-of-life care is a recent development, many research and discussion papers have been generated since the SUPPORT study.8 This huge study, involving over 9000 patients and 5 teaching hospitals, not only described the problem of end-of-life care in hospitals but illustrated the inherent difficulties in changing established practice. Authors found that timely end-of-life decisions and discussion prior to death were uncommon, patients were subjected to aggressive treatment, and pain was not well managed for the majority of patients. Interventions aimed at enhancing physicians’ knowledge regarding the prognoses and preferences of their patients failed to have any impact on management. A retrospective study by Fins and colleagues9 investigated end-of-life care by auditing the charts of 200 consecutive hospital deaths. They reported that, along with the late timing of end-of-life decisions, there remained frequent interventions with intravenous antibiotics and blood tests, and to a lesser extent mechanical intervention, artificial nutrition, and hydration, for patients identified as dying. In this study, 28% of patients had completed a ‘health
Vol. 22 No. 6 December 2001
care proxy’ identifying a surrogate decisionmaker in case of incapacity. The presence of a health care proxy was significantly associated with Do Not Resuscitate (DNR) orders and comfort care plans. In Australia, there is no uniformity between the states and territories regarding legislation for advance directives.10 In the Australian Capital Territory (ACT), advance directives are not legally binding, and although ‘enduring power of attorney’ gives a surrogate the ability to choose treatment direction very few people make use of this legislation. Unlike the USA, little debate has occurred regarding legislature to clarify decision making issues, and patients, family members, and nurses are rarely involved in the DNR decision.11 The location for the current study was The Canberra Hospital, a 550-bed teaching hospital in the ACT. Mortality statistics for 1998 are not yet available; however, in 1997, there were 487 cancer deaths in the ACT.12 Of these, 163 cancer patients died in the inpatient hospice unit and 107 died at The Canberra Hospital (figures taken from medical records). With almost one-quarter of cancer deaths in the ACT occurring at The Canberra Hospital, investigation of the management practices of dying patients within this acute care setting is warranted.
Methods This study was designed to replicate the previously described American-based project by Fins and colleagues.9 Accordingly, the project used a retrospective research design to study the medical records of a sample of patients who died as inpatients at The Canberra Hospital. Research commenced in August 1999 with the approval of the hospital ethics committee. The project was guided by the research question: What are the patterns of medical and nursing practice in the care of patients dying of cancer in the acute care setting? The original study conducted by Fins et al.9 drew upon a sample of 200 patients. The chart abstraction tool used in the current study was validated in that project. With this tool, we judged that this study would be adequately conducted with a sample size of 100. Estimates of overall cancer deaths at the study hospital were 107 per year (1998 estimate). We therefore selected a sampling frame of one year and
Vol. 22 No. 6 December 2001
Dying in Hospital
surveyed the charts of 100 consecutive patients who died in hospital before August 1999 from an oncological or a hematological malignancy. The sampling error from this survey was small due to the homogeneity of the sample and the sample size in relation to the annual rate of cancer deaths. Permission was obtained to use the original chart abstraction tool (consisting of 83 items) and modify it for this study. Changes were made to accommodate the cultural and health services elements of the Australian context. Items surveying advance directives and health care proxies were excluded as patients in Australia rarely hold legal documents covering their wishes for end-of-life care. One item was added to ascertain the degree to which patients were subject to invasive investigations. A section was added to the general comments section for transcription from the chart of patterns of nursing care. These data would provide insight into the influences of nursing practice on the patients’ dying process. In addition, we limited the inclusion criteria to patients who died of cancer. The modified abstraction tool surveyed 60 items of information from the medical records. These items included demographics, diagnoses, reasons for admission, occurrence and timing of DNR orders, whether the patient was considered dying, the documentation of palliative care goals, the withdrawal or non-withdrawal of life sustaining treatment, and the use of diagnostic tests. Data collection was conducted by an experienced oncology nurse. An initial pilot process was conducted to ensure validity of altered items and the rater reliability of the collector. The first 10 charts were also examined independently by one or other principal researcher and the mean agreement between chart abstractors was 96%. The chief investigator assisted in clarifying any unclear responses. As in Fins et al.’s study,9 the question as to whether the patient was considered dying was answered yes if the following words appeared in the notes: end stage, dying, terminally ill, moribund, situation hopeless, prognosis grim. Whether a palliative care plan was in place was determined by the presence of the following language: comfort care, palliative care, supportive care. Data were analyzed using SPSS for Windows.
1037
Results The aim of the study was to draw upon the documented practices in the patient records to describe the medical and nursing care provided for patients who were dying in the acute care setting. The findings were drawn from both quantitative and qualitative data.
Patient Characteristics Patient characteristics are illustrated in Table 1. The median age of patients was 66 years (range 3–95 years). Seventy-four patients had an oncological malignancy, 22 had a hematological malignancy, and 4 patients had more than one primary diagnosis. The most common oncological malignancies were lung, breast, bowel, gynecological, and unknown primary. Seventy-eight percent of cancer patients were documented as having metastatic disease on admission. The most common hematological malignancies were non-Hodgkin’s lymphoma, acute leukemia, and chronic leukemia. A range of medical specialties were involved in patient care: 45 patients were under the care of oncologists, 25 under hematologists, and the remaining 30 were under the care of respiratory, surgical, neurosurgical, pediatric, aged care, gynecology, and renal physicians. Reasons for admission were ascertained from the notes of the admitting doctor. While two patients were admitted for elective surgery, two for chemotherapy, and two for radiotherapy, the remainder of patients were admitted with a cluster of symptoms. These included: pain, nausea and vomiting, dyspnea, lethargy, weakness, dehydration, constipation, diarrhea, confusion, falls, seizures, hypercalcemia, pneuTable 1 Patient Characteristics (N 100) Sex of Patient
Age of patient 0–20 21–40 41–60 61–80 81 Group Total Religion Religion stated No religion No documentation Group Total
Male
Female
Group Total
0 2 11 28 3 44
1 2 23 23 7 56
1 4 34 51 10 100
32 9 3 44
47 6 3 56
79 15 6 100
1038
Middlewood et al.
monia, and jaundice. Febrile neutropenia was documented in the admission notes of 4 patients.
Life Sustaining Management Sixty patients had active treatment in progress at time of death. Eighty-eight patients received IV fluids during admission and 49 patients continued to have IV fluids at time of death. Thirteen patients received artificial nutrition during admission (i.e., percutaneous endoscopic gastrostomy [PEG], nasogastric [NG] feeds, or total parenteral nutrition [TPN]). Five of these patients continued to have feeding in progress at time of death. Sixty-seven patients received IV antibiotics during admission. Antibiotics were not ceased prior to death for 27 patients. Of the 12 patients who were intubated during admission, 7 remained intubated at the time of death. Seventy-eight patients were subjected to diagnostic tests either within the last 48 hours of life or post-DNR order. Fourteen patients were admitted to the Intensive Care Unit (ICU) during admission and 7 of these patients died in ICU. Chemotherapy was administered to 17 patients.
Decisions for End-of-Life Care Of the 100 patients in the sample, 88 were documented DNR prior to death. The language used in 63 patients’ notes suggested they were dying. Seventy-four patients had evidence of a palliative care plan. Only 27% of patients had palliative care goals instigated coupled with a cessation of treatment and investigations more than 48 hours prior to death. Of the 12 patients who were not documented DNR, cardiopulmonary resuscitation (CPR) was attempted on 2. The median length of stay was 9 days (range 0–65 days).
Timing of Decisions Forty-six patients were documented DNR 2 days or less prior to death, and of these 12 were documented DNR on the day of death. Similar patterns emerged for the number of days between palliative care goals and death and between identified as dying and death (see Figure 1 and Table 2). Only 10 patients were considered dying, documented as DNR, and for palliative care on the same day. Six patients had no end-of-life decisions documented.
Vol. 22 No. 6 December 2001
Participants in Decisions According to the notes only 24 patients were involved in discussion regarding their documentation as DNR. In 38 cases, DNR was discussed with the patient’s family but not with the patient. There was no evidence of any discussion in 26 cases where the patient was documented DNR. Consistent with trends in health service research, data triangulation was used, with inclusion of qualitative data, to strengthen the research findings.13 Analysis of the qualitative data revealed that the focus of both nurses and doctors was almost exclusively on the body. Few records had documentation of discussion in which the patient’s and family’s priorities were evident or influenced patient care decisions. For example, the following is an early entry from a patient’s notes: Condition discussed with the patient and partner of 25 years—agreed to treat him conservatively. While not fully describing the nature of the discussion, this entry signals to the health care team that the management of care for this patient would be conservative in accord with the wishes of the patient and his partner. Despite this, the entries over the following 11 days were focused on the body and characterized by interventions. The records included entries from an occupational therapist assessing the patient’s ability to shower himself, nurses focusing on his urine output, a physiotherapist ordering that he should walk short distances with a frame over the weekend, a dietitian asking staff to monitor his food intake, and doctors wanting to transfer him to another hospital. There were also frequent references in the notes to the patient’s lack of cooperation with various treatment initiatives. This man’s condition deteriorated and he died without there being any evidence of palliative care management or a DNR order. However, there were exceptions to this and the following entry from nursing notes features a very different paradigm of care: Joanne believes that 15 years of nonHodgkin’s lymphoma is enough and is keen to maintain as much control as possible of future management. . . . she is wanting to see her mother and brother about speaking at her funeral. Then
Vol. 22 No. 6 December 2001
Dying in Hospital
1039
Table 2 Mean Days from Admission to Decision Points and Between Decision Points by Length of Stay (N 100) Admission to identification as dying
Admission to NFRa
Admission to palliative care goals
ID-PCGb
Length of Stay (days)
N
Mean
SD
N
Mean
SD
N
Mean
SD
N
Mean
SD
1–4 5–9 10–21 22
14 14 17 17 — 62
1.07 5.93 11.29 31.29
1.07 2.09 4.86 17.31
18 26 20 23 — 87
1.33 3.77 11.90 27.74
1.28 2.46 4.64 19.19
13 24 19 18 — 74
0.92 3.92 9.90 23.61
1.04 2.23 6.06 18.22
10 13 14 14 — 51
0.20 1.85 1.93 0.79
0.79 2.88 7.64 13.52
Totals: aNFR
Not for resuscitation order. Identified as dying to palliative care goals.
bID-PCG
she is considering having a syringe driver with morphine and midazolam to relieve symptoms ‘lung pain, dry mouth, general discomfort’. She is having difficulty swallowing medications at the moment. . . . Joanne asked for the commencement of a syringe driver today . . . same commenced. She asked that any distress be managed by increasing medications, therefore may need increasing over the weekend. In this entry, the end-of-life care is contextualized within the patient’s own perspective of her illness history. Throughout this record, the patient has primacy and the documented management decisions are in accord with her wishes. Furthermore, in contrast to the previous exam-
Fig. 1. Patterns of timing of end-of-life decisions (N 100).
ple, the patient’s stated requests were carried through to directives for subsequent care.
Discussion A high proportion of patients in our study had what Fins et al.9 described as the three outcome measures for end-of-life care, namely DNR orders, recognition of dying, and palliative care goals. However, the timing of these documented activities was typically close to the event of death and there was little evidence of a coordinated pattern of care. Ambiguity of decisions was also evident. In the majority of patients when end-of-life decisions were made, invasive tests and life sustaining treatment continued. The results suggest that the transition from acute care to a palliative approach is problematic in this Australian acute care setting. This is consistent with the reported conditions in other acute care settings. The original study by Fins et al.9 found that 72% of patients in their study were considered dying. Nonetheless, almost half the patients died while in an ICU and 30% of patients were receiving artificial nutrition and hydration. The SUPPORT team found in their comprehensive study that 38% of patients with advanced disease spent at least 10 days in an ICU.8 Our findings and the findings from other studies reported here suggest that in the acute care setting there is a pattern of health care that is dominated by a therapeutic imperative which is oriented towards curative interventions. Jorge Urzua, commenting on the role of technology in modern medicine, claimed that a feature of technology use is the predomi-
1040
Middlewood et al.
nance of objective over subjective values.14 Urzua14 suggested that the main criterion for success of technology-related interventions is objective results quantitatively measured. Survival, therefore, measured in terms of time, has primacy over unquantifiable considerations such as subjective feelings, patient comfort, perception of self, or welfare of the family. Subjecting dying patients to medical interventions no doubt adds to rather than reduces their suffering. The reality of so called ‘routine’ procedures, such as taking blood, ordering radiographs, cannulating for IV access, or inserting a nasogastric tube, is that they are often very traumatic for seriously ill patients. The data on timing of end-of-life care decisions indicate a lack of coordination of therapeutic goals, a reliance on euphemism to indicate approaching death, and a lack of recognition of the role of palliative care clinicians in the acute hospital environment. Short admission times prior to death for one-quarter of patients in the absence of pre-admission advance directives further complicate the Australian context. These problems are not confined to the hospital in this study. Reports from the SUPPORT study indicate that in the participating hospitals there was a lack of coordination in timing of DNR orders15 and many patients died with pain and other distressing symptoms.16 The Fins et al. study revealed a greater emphasis on advance planning and patient and family involvement in decision making.9 Only 5% of DNR orders were authorized by physicians alone, with no evidence of patient or family involvement, compared to 30% in our sample. Patients do not have pre-admission DNR orders in the ACT, whereas in Fins et al.’s study 13% had orders in place pre-admission. The lack of a comprehensive management plan for the care of dying patients means that no clear goals of care are documented. Additionally, the absence of a process for end-of-life decision making could lead to the kinds of experiences that patients in this study were subject to. Only 27% of patients had palliative care goals instigated, coupled with cessation of invasive treatments and investigations, more than 48 hours prior to death. This has serious implications for the patient, family, and health care professionals. Dieticians, physiotherapists, nurses, and out of hours doctors are confronted by situations where they have difficulty
Vol. 22 No. 6 December 2001
knowing whether care is geared towards improving health or managing a patient’s comfort. As in the record of Joanne, there may be no ambiguity about the focus of care, suggesting that the quality of documentation affects the decisions regarding the health care team’s interventions. The focus on acute care is also confusing for patients and families. The continuation of life-prolonging treatment can delay patients and their families in gaining an understanding about the terminal nature of their disease. The culture of practice in the acute care setting is inherently related to life supporting and life prolonging activities. Palliative care has presented an alternate philosophy to guide the care for dying patients. Current writings emphasize the need to integrate palliative care services into the acute care setting to assist the transition of goals from investigative and treatment oriented care to improving quality of life. As Fins et al.9 commented, it is important in the first instance for clinicians to recognize that patients are dying. This point will then provide the impetus for planning end-of-life care. The authors acknowledge that a retrospective study using medical records is reliant upon the quality of the notes and the inclinations of the doctors and nurses making these notes. Specifically, discussion may have taken place but not been charted. Nonetheless, the findings from this project describe patterns of practice for dying patients that reflect the curative focus of the acute care setting.
Conclusion Although there are inherent cultural differences between the health care systems in Australia and America, this Australian study found that end-of-life care in this setting is driven by the same treatment imperative oriented toward prolonging life. The late implementation of management plans and lack of clarity within these plans led to many patients being subjected to medical interventions and investigations up to the time of their death. This acute care focus deprived patients and their families of the valuable time preceding death when their individual needs should drive sensitive and compassionate care. The findings from this study can inform practice in the acute care setting in three distinct areas. Acute care clinicians will recognize
Vol. 22 No. 6 December 2001
Dying in Hospital
from reading this report the urgent need to incorporate palliative care goals into the care of patients who die in the acute hospital setting. There is, however, an urgent need for policy reform in the acute care setting to support medical and nursing clinicians in these therapeutic directions. An additional and related implication for practice is the need for clear unambiguous language in documenting care decisions and imperatives for patients who are dying in hospital. The tendency towards euphemistic language obscures the need for a systematic approach to palliative care. Finally, the implications for palliative care include the need to educate health care teams to plan and implement policy regarding the management of patients who are dying in the acute hospital setting. Ultimately, there is an urgent need to provide patients and families with information to empower their participation in decision-making and also to promote public debate in Australia to drive legislation for advance directives throughout the Australian states and territories. While retrospective studies have the benefit of hindsight, the sample in this study contained only patients who had died. A prospective study is urgently needed to test these results.
Acknowledgments The authors gratefully acknowledge The Canberra Hospital Salaried Specialists Private Practice Fund for financial support for this project.
References 1. Miller FG, Fins JJ. A proposal to restructure hospital care for dying patients. N Engl J Med 1996:334: 1740–1742. 2. Hanson LC, Tulsky JA, Danis M. Can clinical interventions change care at the end-of-life? Ann Intern Med 1997;126:381–388.
1041
3. Cassel EJ. The nature of suffering and the goals of medicine. N Engl J Med 1982;306:639–645. 4. Morrison RS, Siu AL, Leipzig RM, et al. The hard task of improving the quality of care at the endof-life. Arch Intern Med 2000;160:743–747. 5. Lickiss JN, Wiltshire J, Glare PA, Chye RWM. Central Sydney Palliative Care Service: potential and limitations of an integrated palliative care service based in a metropolitan teaching hospital. Ann Acad Med Singapore 1994;23:264–270. 6. Hunt R, Bonett A, Roder D. Trends in the terminal care of cancer patients: South Australia, 1981–1990. Aust NZ J Med 1993:23:245–251. 7. Brown M. Principles of palliative care are yet to be applied in acute care hospitals [letter]. BMJ 2000;320:1206. 8. The SUPPORT Principal Investigators. A controlled trial to improve care for seriously ill hospitalized patients. The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA 1995;274:1591–1598. 9. Fins JJ, Miller FG, Acres CA, et al. End-of-life decision making in the hospital: current practice and future prospects. J Pain Symptom Manage 1999;17: 6–15. 10. Schultz L. Not for resuscitation: two decades of challenge for nursing ethics and practice. Nurs Ethics 1997;4:227–238. 11. Manias E. Australian nurses’ experiences and attitudes in the “Do Not Resuscitate” decision. Res Nurs Health 1998;21:429–441. 12. Australian Bureau of Statistics. Australian social trends 1999: health—state summary tables. Available at http://www.abs.gov.au/ausstats/. 2000. 13. Bowling A. Research methods in health: investigating health and health services. Buckingham, UK: Open University Press, 1997. 14. Urzua J. Modern medicine and the rejection of death. Anaesth Intens Care 1991;19:400–403. 15. Hakim RB, Teno JM, Harrell FR, et al., for the SUPPORT Investigators. Factors associated with donot-resuscitate orders: patient’s preferences, prognoses, and physicians’ judgements. Ann Intern Med 1996;125:284–293. 16. Desbiens NA, Wu AW, Broste SK, et al. for the SUPPORT Investigators. Pain and satisfaction with pain control in seriously ill hospitalized adults: findings from the SUPPORT research investigations. Crit Care Med 1996;24:1953–1961.
Journal of Pain and Symptom Management
1035
Original Article
Dying in Hospital: Medical Failure or Natural Outcome? Sue Middlewood, RN, BA, Glenn Gardner, RN, PhD, and Anne Gardner, RN, MPH Palliative Care Services (S.M.), The Canberra Hospital; and The Canberra Hospital and the University of Canberra Research Center for Nursing Practice (G.G., A.G.), Woden, Australian Capital Territory, Australia
Abstract The purpose of this study was to describe patterns of medical and nursing practice in the care of patients dying of oncological and hematological malignancies in the acute care setting in Australia. A tool validated in a similar American study was used to study the medical records of 100 consecutive patients who died of oncological or hematological malignancies before August 1999 at The Canberra Hospital in the Australian Capital Territory. The three major indicators of patterns of end-of-life care were documentation of Do Not Resuscitate (DNR) orders, evidence that the patient was considered dying, and the presence of a palliative care intention. Findings were that 88 patients were documented DNR, 63 patients’ records suggested that the patient was dying, and 74 patients had evidence of a palliative care plan. Forty-six patients were documented DNR 2 days or less prior to death and, of these, 12 were documented the day of death. Similar patterns emerged for days between considered dying and death, and between palliative care goals and death. Sixty patients had active treatment in progress at the time of death. The late implementation of end-of-life management plans and the lack of consistency within these plans suggested that patients were subjected to medical interventions and investigations up to the time of death. Implications for palliative care teams include the need to educate health care staff and to plan and implement policy regarding the management of dying patients in the acute care setting. Although the health care system in Australia has cultural differences when compared to the American context, this research suggests that the treatment imperative to prolong life is similar to that found in Americanbased studies. J Pain Symptom Manage 2001;22:1035–1041 © U.S. Cancer Pain Relief Committee, 2001. Key Words Terminal care, end-of-life decisions, dying in hospital
Introduction Much literature has emanated from North America regarding the care of dying patients within the acute care hospital setting. The
Address reprint requests to: Sue Middlewood, RN, BA, Palliative Care Services, The Canberra Hospital, PO Box 11, Woden, ACT 2606, Australia. Accepted for publication: March 7, 2001. © U.S. Cancer Pain Relief Committee, 2001 Published by Elsevier, New York, New York
common theme of this literature is that dying patients receive inadequate and inappropriate care because their needs do not fit with the ideology of large hospitals, where the main purpose of care is to cure using increasingly sophisticated interventions.1–4 In Australia, the Commonwealth Government has had a commitment to improving patient access to palliative care, whether that be in the hospital, community, or inpatient hos0885-3924/01/$–see front matter PII S0885-3924(01)00362-1
1036
Middlewood et al.
pice setting.5 In South Australia, the proportion of cancer deaths in metropolitan hospitals fell from 40% in 1981 to 28% in 1990.6 Cancer deaths in hospices over the same time period increased from 5% to 20%. The proportion of people dying at home (14%) remained the same. The remaining cancer patients died in country hospitals or nursing homes. A similar proportion of cancer deaths occurred at home in Victoria: 19% in 1988. Sixty percent of South Australians surveyed replied that in the event of a terminal illness they would prefer to die at home.6 Despite preference for home, the actual place of death is more likely to be a large hospital. Although Lickiss and colleagues5 note that palliative care units within teaching hospitals have the opportunity to act as agents of change, little research has been done on the management of dying patients in acute care settings in Australia. At present, studies involving observation of the last six days of life are underway in two teaching hospitals in South Australia. Findings suggest that “. . . the hospital environment reflects the business of life in society, which still denies the naturalness and inevitability of death.”7 Although Australian interest in end-of-life care is a recent development, many research and discussion papers have been generated since the SUPPORT study.8 This huge study, involving over 9000 patients and 5 teaching hospitals, not only described the problem of end-of-life care in hospitals but illustrated the inherent difficulties in changing established practice. Authors found that timely end-of-life decisions and discussion prior to death were uncommon, patients were subjected to aggressive treatment, and pain was not well managed for the majority of patients. Interventions aimed at enhancing physicians’ knowledge regarding the prognoses and preferences of their patients failed to have any impact on management. A retrospective study by Fins and colleagues9 investigated end-of-life care by auditing the charts of 200 consecutive hospital deaths. They reported that, along with the late timing of end-of-life decisions, there remained frequent interventions with intravenous antibiotics and blood tests, and to a lesser extent mechanical intervention, artificial nutrition, and hydration, for patients identified as dying. In this study, 28% of patients had completed a ‘health
Vol. 22 No. 6 December 2001
care proxy’ identifying a surrogate decisionmaker in case of incapacity. The presence of a health care proxy was significantly associated with Do Not Resuscitate (DNR) orders and comfort care plans. In Australia, there is no uniformity between the states and territories regarding legislation for advance directives.10 In the Australian Capital Territory (ACT), advance directives are not legally binding, and although ‘enduring power of attorney’ gives a surrogate the ability to choose treatment direction very few people make use of this legislation. Unlike the USA, little debate has occurred regarding legislature to clarify decision making issues, and patients, family members, and nurses are rarely involved in the DNR decision.11 The location for the current study was The Canberra Hospital, a 550-bed teaching hospital in the ACT. Mortality statistics for 1998 are not yet available; however, in 1997, there were 487 cancer deaths in the ACT.12 Of these, 163 cancer patients died in the inpatient hospice unit and 107 died at The Canberra Hospital (figures taken from medical records). With almost one-quarter of cancer deaths in the ACT occurring at The Canberra Hospital, investigation of the management practices of dying patients within this acute care setting is warranted.
Methods This study was designed to replicate the previously described American-based project by Fins and colleagues.9 Accordingly, the project used a retrospective research design to study the medical records of a sample of patients who died as inpatients at The Canberra Hospital. Research commenced in August 1999 with the approval of the hospital ethics committee. The project was guided by the research question: What are the patterns of medical and nursing practice in the care of patients dying of cancer in the acute care setting? The original study conducted by Fins et al.9 drew upon a sample of 200 patients. The chart abstraction tool used in the current study was validated in that project. With this tool, we judged that this study would be adequately conducted with a sample size of 100. Estimates of overall cancer deaths at the study hospital were 107 per year (1998 estimate). We therefore selected a sampling frame of one year and
Vol. 22 No. 6 December 2001
Dying in Hospital
surveyed the charts of 100 consecutive patients who died in hospital before August 1999 from an oncological or a hematological malignancy. The sampling error from this survey was small due to the homogeneity of the sample and the sample size in relation to the annual rate of cancer deaths. Permission was obtained to use the original chart abstraction tool (consisting of 83 items) and modify it for this study. Changes were made to accommodate the cultural and health services elements of the Australian context. Items surveying advance directives and health care proxies were excluded as patients in Australia rarely hold legal documents covering their wishes for end-of-life care. One item was added to ascertain the degree to which patients were subject to invasive investigations. A section was added to the general comments section for transcription from the chart of patterns of nursing care. These data would provide insight into the influences of nursing practice on the patients’ dying process. In addition, we limited the inclusion criteria to patients who died of cancer. The modified abstraction tool surveyed 60 items of information from the medical records. These items included demographics, diagnoses, reasons for admission, occurrence and timing of DNR orders, whether the patient was considered dying, the documentation of palliative care goals, the withdrawal or non-withdrawal of life sustaining treatment, and the use of diagnostic tests. Data collection was conducted by an experienced oncology nurse. An initial pilot process was conducted to ensure validity of altered items and the rater reliability of the collector. The first 10 charts were also examined independently by one or other principal researcher and the mean agreement between chart abstractors was 96%. The chief investigator assisted in clarifying any unclear responses. As in Fins et al.’s study,9 the question as to whether the patient was considered dying was answered yes if the following words appeared in the notes: end stage, dying, terminally ill, moribund, situation hopeless, prognosis grim. Whether a palliative care plan was in place was determined by the presence of the following language: comfort care, palliative care, supportive care. Data were analyzed using SPSS for Windows.
1037
Results The aim of the study was to draw upon the documented practices in the patient records to describe the medical and nursing care provided for patients who were dying in the acute care setting. The findings were drawn from both quantitative and qualitative data.
Patient Characteristics Patient characteristics are illustrated in Table 1. The median age of patients was 66 years (range 3–95 years). Seventy-four patients had an oncological malignancy, 22 had a hematological malignancy, and 4 patients had more than one primary diagnosis. The most common oncological malignancies were lung, breast, bowel, gynecological, and unknown primary. Seventy-eight percent of cancer patients were documented as having metastatic disease on admission. The most common hematological malignancies were non-Hodgkin’s lymphoma, acute leukemia, and chronic leukemia. A range of medical specialties were involved in patient care: 45 patients were under the care of oncologists, 25 under hematologists, and the remaining 30 were under the care of respiratory, surgical, neurosurgical, pediatric, aged care, gynecology, and renal physicians. Reasons for admission were ascertained from the notes of the admitting doctor. While two patients were admitted for elective surgery, two for chemotherapy, and two for radiotherapy, the remainder of patients were admitted with a cluster of symptoms. These included: pain, nausea and vomiting, dyspnea, lethargy, weakness, dehydration, constipation, diarrhea, confusion, falls, seizures, hypercalcemia, pneuTable 1 Patient Characteristics (N 100) Sex of Patient
Age of patient 0–20 21–40 41–60 61–80 81 Group Total Religion Religion stated No religion No documentation Group Total
Male
Female
Group Total
0 2 11 28 3 44
1 2 23 23 7 56
1 4 34 51 10 100
32 9 3 44
47 6 3 56
79 15 6 100
1038
Middlewood et al.
monia, and jaundice. Febrile neutropenia was documented in the admission notes of 4 patients.
Life Sustaining Management Sixty patients had active treatment in progress at time of death. Eighty-eight patients received IV fluids during admission and 49 patients continued to have IV fluids at time of death. Thirteen patients received artificial nutrition during admission (i.e., percutaneous endoscopic gastrostomy [PEG], nasogastric [NG] feeds, or total parenteral nutrition [TPN]). Five of these patients continued to have feeding in progress at time of death. Sixty-seven patients received IV antibiotics during admission. Antibiotics were not ceased prior to death for 27 patients. Of the 12 patients who were intubated during admission, 7 remained intubated at the time of death. Seventy-eight patients were subjected to diagnostic tests either within the last 48 hours of life or post-DNR order. Fourteen patients were admitted to the Intensive Care Unit (ICU) during admission and 7 of these patients died in ICU. Chemotherapy was administered to 17 patients.
Decisions for End-of-Life Care Of the 100 patients in the sample, 88 were documented DNR prior to death. The language used in 63 patients’ notes suggested they were dying. Seventy-four patients had evidence of a palliative care plan. Only 27% of patients had palliative care goals instigated coupled with a cessation of treatment and investigations more than 48 hours prior to death. Of the 12 patients who were not documented DNR, cardiopulmonary resuscitation (CPR) was attempted on 2. The median length of stay was 9 days (range 0–65 days).
Timing of Decisions Forty-six patients were documented DNR 2 days or less prior to death, and of these 12 were documented DNR on the day of death. Similar patterns emerged for the number of days between palliative care goals and death and between identified as dying and death (see Figure 1 and Table 2). Only 10 patients were considered dying, documented as DNR, and for palliative care on the same day. Six patients had no end-of-life decisions documented.
Vol. 22 No. 6 December 2001
Participants in Decisions According to the notes only 24 patients were involved in discussion regarding their documentation as DNR. In 38 cases, DNR was discussed with the patient’s family but not with the patient. There was no evidence of any discussion in 26 cases where the patient was documented DNR. Consistent with trends in health service research, data triangulation was used, with inclusion of qualitative data, to strengthen the research findings.13 Analysis of the qualitative data revealed that the focus of both nurses and doctors was almost exclusively on the body. Few records had documentation of discussion in which the patient’s and family’s priorities were evident or influenced patient care decisions. For example, the following is an early entry from a patient’s notes: Condition discussed with the patient and partner of 25 years—agreed to treat him conservatively. While not fully describing the nature of the discussion, this entry signals to the health care team that the management of care for this patient would be conservative in accord with the wishes of the patient and his partner. Despite this, the entries over the following 11 days were focused on the body and characterized by interventions. The records included entries from an occupational therapist assessing the patient’s ability to shower himself, nurses focusing on his urine output, a physiotherapist ordering that he should walk short distances with a frame over the weekend, a dietitian asking staff to monitor his food intake, and doctors wanting to transfer him to another hospital. There were also frequent references in the notes to the patient’s lack of cooperation with various treatment initiatives. This man’s condition deteriorated and he died without there being any evidence of palliative care management or a DNR order. However, there were exceptions to this and the following entry from nursing notes features a very different paradigm of care: Joanne believes that 15 years of nonHodgkin’s lymphoma is enough and is keen to maintain as much control as possible of future management. . . . she is wanting to see her mother and brother about speaking at her funeral. Then
Vol. 22 No. 6 December 2001
Dying in Hospital
1039
Table 2 Mean Days from Admission to Decision Points and Between Decision Points by Length of Stay (N 100) Admission to identification as dying
Admission to NFRa
Admission to palliative care goals
ID-PCGb
Length of Stay (days)
N
Mean
SD
N
Mean
SD
N
Mean
SD
N
Mean
SD
1–4 5–9 10–21 22
14 14 17 17 — 62
1.07 5.93 11.29 31.29
1.07 2.09 4.86 17.31
18 26 20 23 — 87
1.33 3.77 11.90 27.74
1.28 2.46 4.64 19.19
13 24 19 18 — 74
0.92 3.92 9.90 23.61
1.04 2.23 6.06 18.22
10 13 14 14 — 51
0.20 1.85 1.93 0.79
0.79 2.88 7.64 13.52
Totals: aNFR
Not for resuscitation order. Identified as dying to palliative care goals.
bID-PCG
she is considering having a syringe driver with morphine and midazolam to relieve symptoms ‘lung pain, dry mouth, general discomfort’. She is having difficulty swallowing medications at the moment. . . . Joanne asked for the commencement of a syringe driver today . . . same commenced. She asked that any distress be managed by increasing medications, therefore may need increasing over the weekend. In this entry, the end-of-life care is contextualized within the patient’s own perspective of her illness history. Throughout this record, the patient has primacy and the documented management decisions are in accord with her wishes. Furthermore, in contrast to the previous exam-
Fig. 1. Patterns of timing of end-of-life decisions (N 100).
ple, the patient’s stated requests were carried through to directives for subsequent care.
Discussion A high proportion of patients in our study had what Fins et al.9 described as the three outcome measures for end-of-life care, namely DNR orders, recognition of dying, and palliative care goals. However, the timing of these documented activities was typically close to the event of death and there was little evidence of a coordinated pattern of care. Ambiguity of decisions was also evident. In the majority of patients when end-of-life decisions were made, invasive tests and life sustaining treatment continued. The results suggest that the transition from acute care to a palliative approach is problematic in this Australian acute care setting. This is consistent with the reported conditions in other acute care settings. The original study by Fins et al.9 found that 72% of patients in their study were considered dying. Nonetheless, almost half the patients died while in an ICU and 30% of patients were receiving artificial nutrition and hydration. The SUPPORT team found in their comprehensive study that 38% of patients with advanced disease spent at least 10 days in an ICU.8 Our findings and the findings from other studies reported here suggest that in the acute care setting there is a pattern of health care that is dominated by a therapeutic imperative which is oriented towards curative interventions. Jorge Urzua, commenting on the role of technology in modern medicine, claimed that a feature of technology use is the predomi-
1040
Middlewood et al.
nance of objective over subjective values.14 Urzua14 suggested that the main criterion for success of technology-related interventions is objective results quantitatively measured. Survival, therefore, measured in terms of time, has primacy over unquantifiable considerations such as subjective feelings, patient comfort, perception of self, or welfare of the family. Subjecting dying patients to medical interventions no doubt adds to rather than reduces their suffering. The reality of so called ‘routine’ procedures, such as taking blood, ordering radiographs, cannulating for IV access, or inserting a nasogastric tube, is that they are often very traumatic for seriously ill patients. The data on timing of end-of-life care decisions indicate a lack of coordination of therapeutic goals, a reliance on euphemism to indicate approaching death, and a lack of recognition of the role of palliative care clinicians in the acute hospital environment. Short admission times prior to death for one-quarter of patients in the absence of pre-admission advance directives further complicate the Australian context. These problems are not confined to the hospital in this study. Reports from the SUPPORT study indicate that in the participating hospitals there was a lack of coordination in timing of DNR orders15 and many patients died with pain and other distressing symptoms.16 The Fins et al. study revealed a greater emphasis on advance planning and patient and family involvement in decision making.9 Only 5% of DNR orders were authorized by physicians alone, with no evidence of patient or family involvement, compared to 30% in our sample. Patients do not have pre-admission DNR orders in the ACT, whereas in Fins et al.’s study 13% had orders in place pre-admission. The lack of a comprehensive management plan for the care of dying patients means that no clear goals of care are documented. Additionally, the absence of a process for end-of-life decision making could lead to the kinds of experiences that patients in this study were subject to. Only 27% of patients had palliative care goals instigated, coupled with cessation of invasive treatments and investigations, more than 48 hours prior to death. This has serious implications for the patient, family, and health care professionals. Dieticians, physiotherapists, nurses, and out of hours doctors are confronted by situations where they have difficulty
Vol. 22 No. 6 December 2001
knowing whether care is geared towards improving health or managing a patient’s comfort. As in the record of Joanne, there may be no ambiguity about the focus of care, suggesting that the quality of documentation affects the decisions regarding the health care team’s interventions. The focus on acute care is also confusing for patients and families. The continuation of life-prolonging treatment can delay patients and their families in gaining an understanding about the terminal nature of their disease. The culture of practice in the acute care setting is inherently related to life supporting and life prolonging activities. Palliative care has presented an alternate philosophy to guide the care for dying patients. Current writings emphasize the need to integrate palliative care services into the acute care setting to assist the transition of goals from investigative and treatment oriented care to improving quality of life. As Fins et al.9 commented, it is important in the first instance for clinicians to recognize that patients are dying. This point will then provide the impetus for planning end-of-life care. The authors acknowledge that a retrospective study using medical records is reliant upon the quality of the notes and the inclinations of the doctors and nurses making these notes. Specifically, discussion may have taken place but not been charted. Nonetheless, the findings from this project describe patterns of practice for dying patients that reflect the curative focus of the acute care setting.
Conclusion Although there are inherent cultural differences between the health care systems in Australia and America, this Australian study found that end-of-life care in this setting is driven by the same treatment imperative oriented toward prolonging life. The late implementation of management plans and lack of clarity within these plans led to many patients being subjected to medical interventions and investigations up to the time of their death. This acute care focus deprived patients and their families of the valuable time preceding death when their individual needs should drive sensitive and compassionate care. The findings from this study can inform practice in the acute care setting in three distinct areas. Acute care clinicians will recognize
Vol. 22 No. 6 December 2001
Dying in Hospital
from reading this report the urgent need to incorporate palliative care goals into the care of patients who die in the acute hospital setting. There is, however, an urgent need for policy reform in the acute care setting to support medical and nursing clinicians in these therapeutic directions. An additional and related implication for practice is the need for clear unambiguous language in documenting care decisions and imperatives for patients who are dying in hospital. The tendency towards euphemistic language obscures the need for a systematic approach to palliative care. Finally, the implications for palliative care include the need to educate health care teams to plan and implement policy regarding the management of patients who are dying in the acute hospital setting. Ultimately, there is an urgent need to provide patients and families with information to empower their participation in decision-making and also to promote public debate in Australia to drive legislation for advance directives throughout the Australian states and territories. While retrospective studies have the benefit of hindsight, the sample in this study contained only patients who had died. A prospective study is urgently needed to test these results.
Acknowledgments The authors gratefully acknowledge The Canberra Hospital Salaried Specialists Private Practice Fund for financial support for this project.
References 1. Miller FG, Fins JJ. A proposal to restructure hospital care for dying patients. N Engl J Med 1996:334: 1740–1742. 2. Hanson LC, Tulsky JA, Danis M. Can clinical interventions change care at the end-of-life? Ann Intern Med 1997;126:381–388.
1041
3. Cassel EJ. The nature of suffering and the goals of medicine. N Engl J Med 1982;306:639–645. 4. Morrison RS, Siu AL, Leipzig RM, et al. The hard task of improving the quality of care at the endof-life. Arch Intern Med 2000;160:743–747. 5. Lickiss JN, Wiltshire J, Glare PA, Chye RWM. Central Sydney Palliative Care Service: potential and limitations of an integrated palliative care service based in a metropolitan teaching hospital. Ann Acad Med Singapore 1994;23:264–270. 6. Hunt R, Bonett A, Roder D. Trends in the terminal care of cancer patients: South Australia, 1981–1990. Aust NZ J Med 1993:23:245–251. 7. Brown M. Principles of palliative care are yet to be applied in acute care hospitals [letter]. BMJ 2000;320:1206. 8. The SUPPORT Principal Investigators. A controlled trial to improve care for seriously ill hospitalized patients. The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA 1995;274:1591–1598. 9. Fins JJ, Miller FG, Acres CA, et al. End-of-life decision making in the hospital: current practice and future prospects. J Pain Symptom Manage 1999;17: 6–15. 10. Schultz L. Not for resuscitation: two decades of challenge for nursing ethics and practice. Nurs Ethics 1997;4:227–238. 11. Manias E. Australian nurses’ experiences and attitudes in the “Do Not Resuscitate” decision. Res Nurs Health 1998;21:429–441. 12. Australian Bureau of Statistics. Australian social trends 1999: health—state summary tables. Available at http://www.abs.gov.au/ausstats/. 2000. 13. Bowling A. Research methods in health: investigating health and health services. Buckingham, UK: Open University Press, 1997. 14. Urzua J. Modern medicine and the rejection of death. Anaesth Intens Care 1991;19:400–403. 15. Hakim RB, Teno JM, Harrell FR, et al., for the SUPPORT Investigators. Factors associated with donot-resuscitate orders: patient’s preferences, prognoses, and physicians’ judgements. Ann Intern Med 1996;125:284–293. 16. Desbiens NA, Wu AW, Broste SK, et al. for the SUPPORT Investigators. Pain and satisfaction with pain control in seriously ill hospitalized adults: findings from the SUPPORT research investigations. Crit Care Med 1996;24:1953–1961.