Symptom Control, Communication With Health Professionals, and Hospital Care of Stroke Patients in the Last Year of Life as Reported by Surviving Family, Friends, and Officials
Background and Purpose The needs of and appropriate service provision for patients dying from stroke have received little attention. The quality of care in the last year of life received by a population-based sample of stroke patients is described here, focusing on symptom control, communication with health professionals, and hospital care.
Methods Secondary analysis was made of data from the Regional Study of Care for the Dying, a retrospective interview survey in 20 nationally representative English health districts. Subjects were 237 persons who died from stroke in 1990. Of informants, 20% were spouses, 48% relatives, 11% friends or neighbors, and 20% officials.
Results More than half the patients were reported to have experienced pain (65%), mental confusion (51%), low mood (57%), and urinary incontinence (56%) in the last year of life. Pain control was inadequate: 51% of those treated for pain by hospital doctors and 45% of those treated by general practitioners were reported to have received treatment that relieved pain partially if at all. One third of respondents thought that hospital doctors had been too rushed (37%), and 25% thought that the patient had had insufficient choice about treatment. Two fifths had been unable to get all the information they had wanted about the patient’s condition.
Conclusions Improvements in symptom control and psychosocial support for patients who die from stroke are needed, as is better communication between health professionals and patients and their families. Education of doctors and nurses working with stroke patients in the principles of palliative care may help ensure that all dying stroke patients receive high-quality care.
The past 25 years have seen growing concern for care of the dying and rapid growth in the number of hospices and specialist palliative-care nurses.1 These service providers have focused almost entirely on cancer patients, and concern for cancer has tended to overshadow the needs of the remaining majority of people dying of other diseases. The 1992 UK Standing Medical Advisory Committee/Standing Nursing and Midwifery Advisory Committee report on palliative care2 argued that palliative-care services are appropriate and should be developed for patients dying of other causes, but little is known about the needs of or appropriate service provision for people dying of causes other than cancer.3 Cerebrovascular disease accounts for 12% of all deaths4 ; however, although there is substantial literature on the prevention of stroke and on the treatment and rehabilitation of stroke patients,5 the needs of and care received by those patients who die of stroke have received little attention. There is some evidence that general practitioners in England would like specialist palliative-care services to be extended to encompass stroke patients,6 but there are concerns that this could lead to existing services being overwhelmed. More important, it is not known to what extent cancer and stroke patients in the last months of life have the same needs for symptom control, open communication, and psychological support. It is therefore not clear how relevant the expertise built up within palliative medicine in the care of cancer patients would be to stroke patients. More information on the needs and experiences of stroke patients who die is needed if future service developments are to be evidence-based and if good quality care is to be provided for all stroke patients, regardless of outcome.
The Regional Study of Care for the Dying (RSCD) was established in 1990 to provide a contemporary account of dying and bereavement in 20 English health districts.7 From a random sample of deaths in each district, family and others who knew the circumstances of the patient’s last year of life were interviewed some 10 months after the death; questions were asked about the deceased’s health problems and restrictions, sources of formal and informal help, the respondent’s satisfaction with services, and their experience of bereavement. In this study, we used data from the RSCD to describe the symptoms reported to have been experienced by this population-based sample of stroke patients in their last year of life, the effectiveness of symptom control by hospital doctors and general practitioners (GPs), and the caregivers’ satisfaction with information received and the quality of hospital inpatient care.
Subjects and Methods
All district health authorities in England were invited to take part in the RSCD. Twenty did so, from a range of inner city, outer urban, and rural settings. Although self-selected, data from Department of Health Performance Indicators show that districts were nationally representative on indicators of deprivation, death rates, and health service provision.8 Within each district, the sample was drawn from death certificates of district residents dying in the last quarter of 1990. A target of 200 completed interviews per district was set to maximize statistical confidence within the limits of the cost of interviews. Since most district health authorities were particularly interested in people dying from cancer, cancer deaths were sampled disproportionately to ensure sufficient numbers to evaluate their care. The total sample size was 5375, 2915 (54%) of which were cancer deaths.
About 10 months after the death, a letter was sent to the deceased’s usual address introducing the study and informing the recipient that they would soon be contacted by an interviewer. Trained interviewers then contacted this address to begin their search for the best informant regarding the deceased’s final 12 months. A 69% response rate was achieved (3696/5375). There were small but statistically significant differences between the main cause of death of the interviewed sample and that of all adult deaths in England in 1990, although the proportion of deaths from circulatory disease did not differ (47% in both groups). Sex of the deceased did not differ from the 1990 national data, but age at death did; deaths of people aged 55 to 64 years were underrepresented, whereas deaths between the ages of 75 and 84 were significantly overrepresented.7 The median time between the deceased’s death and the interview was 44 weeks (10 months), with an interquartile range of 40 (9 months) to 50 weeks (12 months).
We were able to identify and interview a suitable respondent for 237 people for whom stroke was the main cause of death, a response rate of 69% (International Classification of Diseases codes 430 through 438, coded from death certificates following Office of Population Censuses and Surveys rules). Stroke patients for whom interviews were obtained were representative of all sampled stroke deaths in terms of their age and sex. Of the respondents for stroke deaths, 20% (48) had been married to the deceased; 37% (87) were sons, daughters, or siblings of the deceased; 11% (26) were sons- or daughters-in-law or more distant relatives; 11% (26) had been a friend or neighbor; and the remaining 20% were officials, usually members of staff in a nursing or residential home.
Interviewers conducted structured interviews using an adapted version of the schedule used by Cartwright and Seale.9 The schedule covered the last year of life. No attempt was made to distinguish health problems experienced as a consequence of the illness or event that led to the deceased’s death, or services used subsequent to this, from others experienced or used in the last year of life. The schedule contained questions about the deceased’s health problems and restrictions; sources of informal and formal care and the respondent’s experience of caring for the deceased; the deceased’s use of and respondent’s satisfaction with community nursing services, inpatient and outpatient hospital and hospice care, and GP and social services; information from and communication with health professionals; and the respondent’s experience of bereavement and bereavement care. Only results on symptoms, symptom control, hospital care, and communication with health professionals are reported here.
Characteristics of the Sample
Characteristics of stroke deaths for which interviews were obtained are listed in Table 1⇓. Women outnumbered men by a ratio of more than 2:1. The sample was predominantly elderly, with nearly four fifths aged 75 years or over. Nearly one in five (19%) had lived in a nursing home or residential home or had been in the hospital for the whole of their last year of life. Excluding these patients, two fifths of the sample had lived alone, 25% had lived with their spouse only, 8% with their spouse and others, and the remainder (23%) with others only. One third of the sample had had no living children and half had had no living siblings, suggesting that a sizeable proportion had few if any relatives to help care for them toward the end of their lives. Half the sample (48%) had been owner-occupiers, and one third had lived in council accommodations. The vast majority of the sample (232, 98%) were classified by the respondents as white.
Of the 237 stroke deaths, 11 (5%) had died suddenly without warning or time for care, 116 (49%) had had no community care after the stroke (they were either admitted to the hospital immediately after the stroke and died there or had a stroke while in the hospital), and the remaining 112 (47%) had had at least some care in the community in the last year of life. Two thirds died in the hospital.
Time From Stroke
The RSCD was not primarily a survey of stroke patients. Instead, it focused on describing the last year of life of a random sample of people who died of a variety of causes. Respondents were therefore not asked specifically about the deceased’s history of stroke or about the timing of the stroke that led to the death. The survey was entirely interview-based; therefore, no information was available from hospital or GP records. However, for 155 stroke deaths (65%) it was possible to ascertain the time between the patients’ first stroke and their death from respondents’ answers to open questions about what was wrong with the patient and about the circumstances of death.
According to respondents, one patient in 10 died within 24 hours of their first stroke, two fifths survived for longer than 1 day but less than 1 month, 15% for between 1 and 6 months, and the remaining third for longer than 6 months (Table 1⇑). It would therefore appear that the majority of patients had survived for a relatively short time after their first stroke. However, this information was lacking for one third of the sample, and indeed for 42 (18%), stroke was not mentioned at all on the interview schedule as having been suffered by the deceased and was recorded only on the death certificate.
Symptoms and Symptom Control
Respondents were asked whether the deceased had suffered the symptoms listed in Table 2⇓ during the last year, whether the symptom had been present in the last week of life, and how long the deceased had had the symptom.
Excluding sudden deaths, at least half of the patients in the sample were reported to have experienced pain, urinary incontinence, low mood, and mental confusion in the year before death. Breathlessness, insomnia, loss of appetite, constipation, and loss of bowel control were reported to have been experienced by at least one third. Many were reported to have had these symptoms for at least 6 months, including four fifths of those reported to have been in pain or to have been constipated, nine tenths of those who had suffered from insomnia or had had a persistent cough or were anxious, and three quarters of those who had been feeling low or were in pain.
The symptoms reported to have been most common in the last week of life were urinary incontinence (51%), pain (42%), mental confusion (41%), low mood (33%), and fecal incontinence (31%). These symptoms were less common in the last week of life than in the last year.
In addition to questions about the presence, duration, and severity of symptoms in the last year of life, respondents were also asked a series of questions about the control by the GP and hospital doctors of four symptoms: pain, breathlessness, nausea/vomiting, and constipation (Table 3⇓). More than three quarters of respondents thought that the GPs’ treatment for constipation and nausea/vomiting had relieved these symptoms “a lot/some” (88% and 79%, respectively), but somewhat smaller proportions reported this degree of control of pain and breathlessness (55% and 66%, respectively). More than four fifths of respondents thought that the GP had tried hard enough to control these symptoms.
Half of the patients who were treated for pain by hospital doctors were thought to have received treatment that relieved the symptom “a lot/some.” Control of breathlessness, nausea/vomiting, and constipation was thought to have been rather more effective (67%, 60%, and 84% of respondents reporting that the symptom was controlled a lot/some, respectively). The vast majority thought that the doctors had tried hard enough to control breathlessness (90%) and constipation (92%), but they were less sure that the doctors had tried hard enough to relieve pain (75%). (The numbers who suffered from nausea/vomiting were too small to draw meaningful conclusions.)
Use of and Satisfaction With Hospital Inpatient Care
Of patients who did not die immediately from a stroke, 79% had been admitted to the hospital during their last year of life. Nine percent had spent fewer than 3 nights in the hospital, 19% 3 nights to 1 week, 31% 1 week to 1 month, 27% 1 to 3 months, and 13% had been in the hospital for more than 3 months. Table 4⇓ describes the respondents’ satisfaction with this care. Overall care from nurses was rated higher than that from doctors: less than one third of respondents (29%) rated hospital doctors’ care as excellent, but almost half of the respondents (46%) rated hospital nurses this high. Fewer than one in 10 respondents felt that the doctors’ or nurses’ care had been poor. However, more than one third (37%) felt that doctors had been too rushed, and one quarter thought that the deceased had had too little choice about treatment. Hospital facilities were criticized by a sizeable minority, with one quarter feeling that the deceased had had insufficient privacy, and one fifth reporting that the deceased’s room had not been at all peaceful or quiet. Excluding officials, nearly one fifth of respondents felt that they had been treated with indifference or had even had a hostile reception when visiting the deceased.
Communication With Health Professionals and Deceased’s and Respondents’ Knowledge of Diagnosis and Prognosis
Respondents who were neither officials nor responding for sudden deaths were asked whether they had been able to get all the information they had wanted about the deceased’s medical condition, when they had wanted it. Two fifths (65, 39%) had been unable to do so. Respondents who were not satisfied with the information they had received were asked what they would have liked to have known about, or to have known about in more detail, or to have known about earlier. Nearly half had wanted more information about the likely outcome of the condition (46%), and a similar proportion had wanted to know more about what was wrong with the deceased. Thirty percent had wanted more information about how to cope with or care for the deceased, 20% had wanted to know more about the reasons for decisions about medical care or treatment, and 18% had wanted more information about medication or treatment.
The majority of those who had wanted more information had asked a health professional for more information (41, 63%). Three fifths of these respondents (25, 61%) had asked a hospital doctor, 41% a hospital nurse, and 34% had asked a GP. When asked why their informational needs had not been met despite asking for more information, 10% said that they had not understood the answer, 27% that the person they had asked had been too rushed, 12% that the person they had asked had been unwilling to give information, and 12% that the person asked had not known the answer. (These categories are not mutually exclusive because respondents could give several responses to this question: responses by fewer than 10% are omitted.)
According to respondents, one fifth of the deceased (23%) had definitely known that they were likely to die, and a similar proportion had probably known (19%) (patients who died suddenly are excluded) (Table 5⇓). Almost all of these were believed to have worked this out for themselves. The majority of these patients (67%) were described as being definitely accepting of their prognosis. One fifth of those who were believed to have known their prognosis were thought to have known for less than 1 week (20%), 63% for less than 1 month, and 86% for less than 6 months.
Eighty-one percent of respondents reported that they had definitely known what was wrong with the deceased, and a further 11% had suspected. A somewhat smaller proportion (58%) had definitely known that the deceased was dying, and a further 22% had suspected this. More than one third (36%) of those who knew or “half knew” the prognosis had worked this out for themselves. Of those respondents who had been told the prognosis by a health professional, most were told by a hospital doctor. The vast majority of those who knew the prognosis (86, 94%) thought that it was better that they did know. Of the respondents who had not known, just over one third (36%) would have liked to have known.
We believe this to be the first reported population-based study of the care of persons who die of stroke. The results suggest that many patients who die of stroke do not receive optimal symptom control, sufficient help to overcome psychological morbidity, or enough choice about their care. Many of their informal caregivers are dissatisfied with the information they receive from health professionals, and some are critical of hospital care of these patients.
The retrospective approach to collecting information about care for the dying used in the RSCD has limitations in that little is known about how and in what ways respondents’ views diverge from those of patients themselves or about the effects of bereavement on how care before the patient’s death is remembered and judged. Interpretation of the results of the small number of studies addressing these issues is hindered by differences between the time period assessed prospectively and that assessed retrospectively and by small sample sizes.10 11 12 The available evidence suggests that there may be better agreement between patients’ and carers’ assessments of physical symptoms than for psychological ones,12 and that carers’ views may become more polarized in bereavement.10 In the absence of more definitive studies, caution is needed in interpreting the results of retrospective studies such as the RSCD. However, adopting the alternative approach of interviewing patients themselves is likely to result in an unrepresentative sample because it will be restricted to those known to be dying, many of whom may be too ill to participate by the time their prognosis is known.13 The retrospective approach in the RSCD ensured that the sample was representative of all people who died, not just those who were known to be dying or who were receiving services.
The sample of stroke deaths for which interviews were obtained was representative of all stroke deaths in England for age and sex. The participating districts were representative of all English districts for deprivation, death rates, and health service provision. The results presented here are therefore likely to be representative of England as a whole.
Symptoms and Symptom Control
Our results show that stroke patients suffer from a number of long-lasting symptoms, notably pain, mental confusion, low mood, and urinary and fecal incontinence. Because it is clearly difficult to make accurate judgments about symptoms experienced by patients whose ability to communicate or level of consciousness have been affected by stroke, these results may underestimate the misery caused by poorly controlled symptoms. In addition, the list of symptoms did not include some likely to be of particular salience in stroke patients, such as dysphagia and paralysis, and again this may have led to the underestimation of symptom distress of these patients. It is also important to note that it is not possible in RCSD to distinguish between symptoms experienced as a consequence of a stroke, or even subsequent to it, from others that are unrelated to stroke. However, with the exception of breathlessness and anorexia, the most common symptoms reported by respondents are recognized effects of stroke.5 This provides some evidence for the validity of the data. Nevertheless, these data need replicating with a prospective study design in which the relative timing of symptoms and stroke can be determined.
There is some suggestion in the literature that help for stroke patients tapers off once active rehabilitation is withdrawn,5 and concern has been expressed that some patients are being discharged from hospitals to nursing homes too quickly without adequate attention being paid to their medical and nursing needs.14 Both the withdrawal of support after active rehabilitation and the early discharge of stroke patients may reflect a belief on the part of doctors, and perhaps nurses, that they have little more to offer once the best possible restitution of premorbid abilities has been achieved. Our findings suggest that this is unjustified: stroke patients need skilled attention to symptom control if unnecessary distress is to be avoided and if they are to be allowed as good a quality of life as possible. Skilled nursing care is indicated to help patients and caregivers cope with the effects of bladder and bowel incontinence and to prevent bedsores. Better pain control for stroke patients is also needed. This may require more systematic study of the cause and best treatment for pain in stroke patients.5 These results suggest that some stroke patients would benefit from the knowledge and expertise in symptom control developed within palliative medicine, although further research is needed to investigate whether treatments devised primarily to meet the needs of cancer patients are appropriate for stroke patients, whose symptoms have a quite different cause.
Ebrahim,5 in his review of the literature, concluded that about one in four 6-month survivors of stroke was clinically depressed. The percentage of respondents reporting that RSCD stroke patients had suffered from low mood in the last year of life was considerably higher (57%). This may reflect differences in the time periods studied and the various ways in which mood disorder was measured (respondents’ reports versus validated questionnaires or psychiatrists’ assessments). Some of the deceased who were reported to have had low mood may have demonstrated emotional lability rather than clinical depression.15 Nevertheless, as in some previous studies,16 although it is not possible to distinguish between existing mood disorders and those caused by physical or social effects of stroke, the results provide further evidence that depressed mood affects many stroke patients. Three quarters of those reported to have had low mood were reported to have had it for at least 6 months, suggesting that mood disturbances in stroke patients can be long lasting and may need careful treatment beyond the period of acute rehabilitation. Psychological support, perhaps provided primarily by nurses with training in basic counseling skills and backed up by psychiatric help where necessary, is needed to help reduce the number of patients experiencing low mood and anxiety.
Information From and Communication With Health Professionals
Levels of dissatisfaction with communication with health professionals similar to those levels reported here were found in a recent survey of surviving patients and their caregivers; that survey found that over half the carers wanted to know more about stroke, one fifth had been given confusing information from health professionals, and three quarters had had to ask for information.17 The authors concluded that more information needed to be made available (perhaps via literature, group meetings, and stroke family officers) and that efficient ways of improving understanding among carers needed to be found. Our results support the need for health professionals to communicate more fully with the families of stroke patients. Nearly two fifths had been unable to get all the information they wanted despite the majority having asked someone (usually a hospital doctor or nurse) for more information. Respondents did not always understand the information they were given, and understanding was probably not helped by the common perception that the health professionals were rushed. Although the provision of literature, volunteer groups, and special stroke officers17 might well help to improve matters, these should not be seen as substitutes for training doctors and nurses in communication skills, encouraging them to use their skills, and providing the necessary time and facilities (such as a quiet office) to allow them to do so. As many as one third of stroke patients die within 1 week and up to half within 3 weeks,18 which does not allow much time for patients and their carers to read and assimilate literature, to join a support group, or even to be contacted by a special stroke officer. Hospital doctors and nurses, and to a lesser degree GPs, are likely to remain the primary sources of information; they need to be aware of caregivers’ need for enough information to enable them to understand what has happened, to care for the patient, and to make plans for the future.
According to these results, twice as many carers as patients knew or suspected that the patient was likely to die. Almost all patients who knew or probably knew were believed to have worked this out for themselves, whereas two thirds of carers were told, usually by a hospital doctor. These results suggest a situation similar to that in oncology 30 years ago, when family members but not the patient were told the patient’s prognosis. It has been increasingly recognized that this approach hinders communication between family members and the cancer patient in the last weeks of life, and it has largely been abandoned in oncology in the UK. The single best predictor of early death in stroke patients has been reported to be level of consciousness at admission,5 and clearly it is not possible to discuss prognosis with an unconscious patient. However, the fact that two fifths of patients were thought to have at least suspected that they were likely to die suggests that the questions of when and how to break bad news, which have received much coverage in oncology, are not irrelevant to the care of stroke patients. Information on the likely prognosis can enable patients and their families to make decisions about the type and location of care during the time remaining and, in the case of imminent death, can enable other relatives to be summoned. Further research is needed to investigate the proportion of patients for whom prognostic information is available and the accuracy of this information. The demand from patients and families for prognostic information also needs further exploration, as do the consequences of giving prognostic information.
Despite the limitations of the data set, a number of important areas of concern have emerged from this analysis. The results suggest that many patients who die of stroke do not receive optimal symptom control or sufficient help to overcome psychological morbidity. Their informal caregivers experience difficulty in getting the information they need about the patient’s medical condition. Some of these issues, such as the need to help stroke patients overcome low mood, have been reported elsewhere.5 Others, such as the lack of good symptom control or adequate information from health professionals, have received less attention. Taken together, they demonstrate the importance of providing care for stroke patients and their families that encompasses their physical, emotional, and social needs and aims to improve the quality of life remaining. Further research is needed to explore the experiences of stroke patients who die and their families in more detail and to identify and evaluate the most effective ways of meeting their needs. In particular, the effects of educating doctors and nurses working with stroke patients, whether within the National Health Service or within the private nursing home sector, on the principles and practices of palliative care need to be explored to see whether, as has been suggested, this leads to a general rise in the quality of care offered to patients with chronic, life-threatening diseases.2
The analyses on which this article is based were funded by the Stroke Association of England and Wales. The Regional Study of Care for the Dying was funded by North East Thames Regional Health Authority, East Anglia Regional Health Authority, South East Thames Regional Health Authority, The Care Foundation Tunbridge Wells, The Stanley Luff Bequest Fund, and participating districts. The authors are grateful to the research interviewers and to Ailsa Cameron, Tim Cooper, Walid Fakhoury, and David Whitehead for their contributions to data collection and analysis. We are especially grateful to the many people who agreed to be interviewed, often recalling sad and painful experiences in the process.
- Received May 22, 1995.
- Revision received September 12, 1995.
- Accepted September 14, 1995.
- Copyright © 1995 by American Heart Association
Higginson IJ, McCarthy M. Evaluation of palliative care: steps to quality assurance. Palliat Med. 1989;3:267-274.
Standing Medical Advisory Committee, Standing Nursing and Midwifery Advisory Committee. The Principles and Provision of Palliative Care. London, UK: Department of Health; 1992.
Seale C. Death from cancer and death from other causes: the relevance of the hospice approach. Palliat Med. 1991;5:13-20.
Office of Population Censuses and Surveys. 1990 Mortality Statistics. London, UK: Her Majesty’s Stationery Office; 1990.
Ebrahim S. Clinical Epidemiology of Stroke. New York, NY: Oxford University Press; 1990:1-227.
Wilson IM, Bunting JS, Curnow RN, Knock J. The need for inpatient palliative care facilities for noncancer patients in the Thames Valley. Palliat Med.. 1995;9:13-18.
Addington-Hall JM, McCarthy M. Regional Study of Care for the Dying: methods and sample characteristics. Palliat Med.. 1995;9:27-35.
Addington-Hall JM, McCarthy M. Can national surveys be funded successfully from local NHS resources? Evidence from the RSCD. J Public Health Med.. 1995;17:161-163.
Cartwright A, Seale C. The Natural History of a Survey: An Account of the Methodological Issues Encountered in a Study of Life Before Death. London, UK: King’s Fund; 1990.
Higginson IJ, Priest P, McCarthy M. Are bereaved family members a valid proxy for a patient’s assessment of dying? Soc Sci Med. 1994;38:553-557.
Ahmedzai S, Morton A, Reid J, Stevenson R. Quality of death from lung cancer: patients’ reports and relatives’ retrospective opinions. In: Watson M, Greer S, Thomas C, eds. Psychosocial Oncology. Oxford, UK: Pergamon Press; 1988.
Field D, Douglas C, Jagger C, Dand P. Terminal illness: views of patients and their lay carers. Palliat Med. 1995;9:45-54.
Addington-Hall JM, MacDonald LD, Anderson HR, Chamberlain J, Freeling P, Bland JM, Raftery J. Randomised controlled trial of co-ordinating care for terminally ill cancer patients. BMJ. 1992;305:1317-1322.
House A, Dennis M, Molyneux A, Warlow C, Hawton K. Emotionalism after stroke. BMJ. 1989;298:991-994.
House A. Depression after stroke. BMJ. 1987;294:76-78.
Wellwood I, Martin S, Dennis CPW. Perceptions and knowledge of stroke among surviving patients with stroke and their carers. Age Ageing. 1994;23:293-298.
Weddell JM, Beresford SAA. Planning for Stroke Patients: A Four Year Descriptive Study of Home and Hospital Care. London, UK: Her Majesty’s Stationery Office; 1979.