SEARCH

SEARCH BY CITATION

Keywords:

  • bereaved relatives;
  • experiences;
  • hospital care;
  • nursing;
  • patients;
  • stroke;
  • survey;
  • VOICES questionnaire

Abstract

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

Title. Experiences of hospital care reported by bereaved relatives of patients after astroke: a retrospective survey using the VOICES questionnaire.

Aim.  This paper is a report of a study conducted to explore the determinants of satisfaction with health and social care services in the last 3 months and 3 days of life as reported by bereaved relatives of those who died from a stroke in an institutional setting.

Background.  There is limited research about how best to meet the needs of those who die from stroke. A thorough understanding of the determinants of satisfaction with end of life care is crucial for effective service provision to increase awareness of the needs of dying patients.

Methods.  During a six-month period in 2003, a population-based survey of bereaved relatives of patients after stroke was conducted using a stroke-specific version of the Views of Informal Carers Evaluation of Services postal questionnaire (183 informants, response rate 37%). The sub-sample included those informants who reported that the deceased person had died in an institutional setting (91%, = 165). The analysis was divided into two phases: univariate (Pearson chi-square test) and multivariate phase (logistic regression).

Results.  Logistic regressions showed that discussing any worries about the treatment of the deceased person and feeling that the doctors and nurses knew enough about their condition were predictors of satisfaction with doctors and nurses in the last 3 months of life. Meeting the personal care needs of the deceased person, being involved in decisions and feeling that the deceased person died in the right place were predictors of satisfaction with care in the last 3 days of life.

Conclusion.  End of life care needs to address the individual needs of patients who die from stroke and those close to them. This study shows that individualised end of life care increases satisfaction and, although the data reported in this paper reflect care in 2003, there is no more recent evidence that contradicts this important overall finding.


What is already known about this topic

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References
  • The concept of satisfaction remains ill-defined but satisfaction with or ratings of health care continue to play an important role in the evaluation and development of healthcare services.
  • Proxy accounts of health care play an important part of evaluations of care at the end of life.
  • Little is known about the needs of patients who died following a stroke, and their needs may differ from those who have cancer.

What this paper adds

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References
  • The predictors of rating end of life stroke care as excellent were similar to those reported among other groups and have been reflected in the literature on a ‘good death’.
  • Relatives who were actively involved in discussions with doctors and nurses about the condition of the deceased person and feel that staff know enough were more likely to rate care as excellent.
  • Addressing the personal care needs of patients and involving relatives in decisions about treatment in the last 3 days of life were likely to increase proxy ratings of overall care.

Implications for practice and/or policy

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References
  • End of life care should have a strong focus in clinical practice and clinical teams should work to integrated communication and management plans which should be clearly documented and updated.
  • Communication skills are paramount in end of life care and clinical staff should have access to high quality education and support in this area.
  • Clinical staff should strive to provide individualized care to meet the personal care needs of patients who are dying.

Introduction

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

Measures of patient satisfaction are now an integral component of quality assurance in many healthcare systems, including the National Health Service (NHS) in the United Kingdom (UK) (Department of Health (DoH) (1999). To date there has been little published research concerning the determinants of satisfaction with end of life care (Aspinal et al. 2003), in contrast to the large and growing literature focusing on general medical care (Bikker & Thompson 2006, Jackson et al. 2001). Measures developed for use in general medical care are unlikely to be suitable for use in palliative or end of life care (Heyland et al. 2006). A thorough understanding of the determinants of satisfaction with end of life care is crucial for effective service provision to increase awareness of the needs of dying patients.

Satisfaction surveys

The last 20 years have seen growing interest worldwide in the use of measures of satisfaction in the assessment of health care (Alazri & Neal 2003, Carr-Hill 1992, Fakhoury 1998, Fan et al. 2005, Jenkinson et al. 2002, Mead et al. 2002). Patient satisfaction measures have been used to evaluate healthcare systems (Bikker & Thompson 2006), quality of care (Suhonen et al. 2005) and to address particular aspects of service provision (Alazri & Neal 2003), and in consequence have influenced the development of health care and policy. Despite the wide use of patient satisfaction measures, satisfaction continues to be an ill-defined concept (Aspinal et al. 2003). Research suggests that a number of factors are inherently related to patient satisfaction including expectations and perceived value of services and quality of care (Rogers et al. 2000, Wilson et al. 2002), communication with healthcare professionals (Mead et al. 2002), a patient’s mental state (Williams et al. 1998) and beliefs about how services can be improved (Jackson et al. 2001). Socioeconomic status, age, gender and area of residence have also been shown to influence satisfaction (Fakhoury et al. 1997).

Recently, and largely as a result of the work of the Picker Institute (http://www.pickereurope.org/), attention has begun to be focussed on patient experiences rather than overall satisfaction as a key indicator of quality or performance in health care. Measuring a patient’s experience is likely to give a clearer picture about how services can be improved and what patients or those close to them feel is important to deliver a better standard of care. This approach is used by the Picker Institute in its national patient surveys, commissioned by the DoH in the UK, to evaluate services delivered by the NHS from the perspective of patients. These surveys cover eight dimensions of care: access; patients’ values, preferences and expressed needs; co-ordination and integration of care; information, communication and education; physical comfort; emotional support and alleviation of fear and anxiety; involvement of family and friends; and transition and continuity.

End of life care

In 2008, the first national End of Life Care Strategy was launched in England and Wales by the DoH (2008). This recommends a care pathway approach both for commissioning services and for delivery of integrated care for individuals to ensure that everyone who dies, regardless of cause or setting, receives appropriate, high quality care. The strategy builds on previous UK policy documents such as the 2007 interim report ‘Our NHS, Our Future’, which highlights that all health care professionals are responsible for the care provided to dying patients and that such care needs to become more personal, to ensure that the needs and preferences of individual patients are met (DoH 2007a, Ellershaw 2007). The need for improved end of life care is evidenced by the fact that the Healthcare Commission (2007) highlighted that the most common complaints in acute settings concern relationships between healthcare staff and family members, lack of attention to basic clinical records and the care given to dying patients. The End of Life Care Strategy discusses the importance of the availability of good information on end of life care to, for example, evaluate the impact of the strategy and enable comparisons between different parts of the UK countries. It emphases the importance of including information on patient and carer experience of care, and proposes that this information is collected via a national survey of bereaved relatives. This highlights the importance of understanding the determinants of bereaved informants’ satisfaction with care.

End of life care for patients after stroke

Stroke is one of the leading causes of death in the UK and is a major cause of severe disability (Office for National Statistics 2006). In recent years, there have been a number of policy documents about stroke care services and access to appropriate hospital stroke care, most notably the National Service Framework for Older People (2001) and the Intercollegiate Stroke Working Party (2004, 2005, 2007). The Healthcare Commission reported that, although patients after a stroke are generally positive about the care and treatment received in hospital, they want to be more involved in decisions, would like more information about their care and to receive more help with personal care such as eating, washing and going to the toilet (Healthcare Commission 2005, 2006).

In 2007, the DoH (2007b) developed a best practice guidance ‘National Stroke Strategy’ in response to lack of awareness of stroke and the need to improve care for stroke survivors in England and Wales. This included recognition of the need to raise the quality of end of life care received for people with stroke (Healthcare Commission 2007, Leatherman et al. 2008). To date, there has been little published research in this area (DoH 2007c, Stevens et al. 2007).

The NHS End of Life Care Programme (http://www.endoflifecare.nhs.uk/eolc) places emphasis on providing high quality end of life care for all patients regardless of diagnosis. In 2003, we conducted a population-based survey of bereaved relatives after stroke to provide information on the experiences of people who die following a stroke and the perceptions of their families and friends to help improve the care of these patients. In this paper, we examine the determinants of satisfaction with hospital care in the last 3 months of life and health and social care in the last 3 days of life to identify priorities for service improvement.

The study

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

Aim

The aim of the study was to explore the determinants of satisfaction with health and social care services in the last 3 months and 3 days of life as reported by bereaved relatives of those who died from a stroke in an institutional setting.

Design

A retrospective survey of bereaved relatives’ views of health and social services at the end of life was conducted, using a revised Views of Informal Carers Evaluation of Services (VOICES) postal questionnaire (for a detailed description of the full study, see Addington-Hall et al. 2006a, Young et al. 2008). The data presented in this paper are a sub-sample from the larger study who responded that their relatives had died in an institutional setting. Exploratory analyses were carried out to identify determinants of satisfaction with care by health and social care services.

Sample

The Office for National Statistics was used to select a random sample of informants who had registered a stroke death in four primary care trusts (primary care provider organizations) in London in 2003. A total of 493 informants were selected from specific ICD codes (A00-B99, C00-D48, D55-D59, E00-E90, F00-F99, G00-G99, I00-I99, J00-J99, K00-K93, M00-M99 and N00-N99), where stroke was given as the main or underlying cause of death (full details can be found in Addington-Hall et al. 2006a, Young et al. 2008). Data were collected over a six-month period in 2003. Three to 9 months after the death, potential respondents were sent a questionnaire pack containing a letter of invitation, information sheet, VOICES II questionnaire, refusal form and a stamped addressed envelope. Two reminder mailings were sent, one approximately 3 weeks after the initial mailing and another after a further 3 weeks.

Questionnaire

The VOICES questionnaire was initially developed following completion of the Regional Study of Care of the Dying (Addington-Hall & McCarthy 1995) and was demonstrated in a randomized controlled trial to be an acceptable alternative to interviews in post-bereavement surveys (Addington-Hall et al. 1998). VOICES is a survey questionnaire rather than an outcome measure with fixed content, and its exact content varies between surveys, depending on the precise aims and objectives of the survey; principles of survey questionnaire design and of testing survey instruments have been applied at each stage of its development (Addington-Hall et al. 2007). The stroke-specific VOICES questionnaire used in the survey reported here (VOICES-II) was developed following a literature review and interviews with 21 health and social care professionals and six bereaved carers. It contains questions on the patient’s or bereaved relative’s experiences of care in seven domains: help at home, general practitioners, nursing and residential home care, last hospital admission, symptoms and treatment, care in the last 3 days of life and circumstances surrounding the death. It also contains questions on demographics and informants’ views on completing the questionnaire.

In this paper, we focus on the responses given about the last hospital admission in the last 3 months and care received by health and social services in the last 3 days of life.

Ethical considerations

The study was approved by the relevant research ethics committee. All those sampled were given the research team’s contact details, together with those of local and national bereavement support agencies in case they wanted to discuss any issues raised by the questionnaire. Consent was assumed when participants returned their completed VOICES-II questionnaires.

Questions from VOICES-II used in the analysis

Satisfaction variables

The VOICES-II questionnaire included three categorical questions to assess informants’ overall satisfaction with care:

  • • 
    Overall, do you feel that the care they got from doctors in hospital was: excellent, good, fair or poor?
  • • 
    Overall, do you feel that the care they got from nurses in hospital was: excellent, good, fair or poor?
  • • 
    Overall, do you think that the care they received in the last 3 days of life from health and social services was: excellent, good, fair or poor?

Because of the relatively small sample, and because our primary aim was to investigate predictors of high satisfaction, these questions were dichotomized for analysis into excellent care and not excellent care (good, fair or poor).

Socio-demographic variables

Five socio-demographic variables were included in the analyses. Two of these were the age and gender of the deceased person and the remaining three concerned the informant’s age, gender and relationship to the deceased person (Table 1).

Table 1.   Characteristics of the sub-sample
Characteristicsn (%)*
  1. *Percentages rounded up to the nearest whole number.

Age of the deceased person at death (years)
 50–593 (2)
 60–6918 (11)
 70–7933 (20)
 ≥80109 (66)
 Do not know/missing 2 (1)
Sex of the deceased person
 Male65 (39)
 Female100 (61)
Ethnic group of the deceased person
 White British136 (82)
 White Irish7 (4)
 White other6 (4)
 Asian Indian1 (1)
 Black Caribbean10 (6)
 African4 (2)
 Chinese1 (1)
Age of the informant (years)
 20–291 (1)
 30–3910 (6)
 40–4928 (17)
 50–5957 (34)
 60–6944 (27)
 70–7919 (12)
 ≥806 (4)
Sex of the informant
 Male67 (41)
 Female95 (58)
 Missing data3 (2)
Ethnic group of the informant
 White British143 (87)
 White Irish4 (2)
 White other2 (1)
 Asian Indian2 (1)
 Black Caribbean7 (4)
 African4 (2)
 Black other1 (1)
 Chinese1 (1)
 Missing data1 (1)
Relationship of the informant and deceased
 Spouse/partner24 (15)
 Son/daughter90 (55)
 Brother/sister10 (6)
 Son-in-law/daughter-in-law7 (4)
 Parent6 (4)
 Non-relative5 (3)
 Missing data23 (14)
Hospital service variables

Six variables were used to assess overall care during the last hospital admission in the last 3 months of life (Table 3). If the deceased person had been admitted to hospital on more than one occasion during the last 3 months of life, the informant was asked to report about the last hospital admission only.

Table 3.   Univariate analyses of the care received from doctors and nurses in the last 3 months of life of those that had a hospital admission during this time-period
VariableOverall rating of the care from doctors in hospitalOverall rating of the care from nurses in hospital
Excellent, n (%)*Other, n (%)*P valueExcellent, n (%)*Other, n (%)*P value
  1. *Percentages rounded up to the nearest whole number.

  2. P value relates to a chi-squared test where the responses to the questions have been dichotomized into the most positive response compared with all other possible responses for that question (hence relate to a 2 × 2 table).

Were you able to discuss with hospital staff any worries or fears you might have had about the condition, treatments or tests of the deceased?
 Yes, I discussed them as much as I wanted28 (80)31 (34)≤0·00128 (76)31 (34)≤0·001
 I had no worries or fears to discuss1 (3)6 (6)1 (3)6 (7)
 Yes, I discussed them but not as much as I wanted5 (14)43 (47)6 (16)42 (47)
 No, although I tried to discuss them1 (3)9 (10)2 (5)8 (9)
 No, but I did not try to discuss them0 (0)3 (3)0 (0)3 (3)
Were there enough nurses on duty to care for the deceased in hospital?
 There were always or nearly always enough nurses on duty26 (74)33 (36)≤0·00131 (84)28 (31)≤0·001
 There were sometimes enough nurses on duty8 (23)27 (29)5 (13)30 (33)
 There were rarely or never enough nurses on duty0 (0)23 (25)1 (3)22 (25)
 Do not know1 (3)9 (10)0 (0)10 (11)
Did the doctors and nurses know enough about the condition of the deceased?
 All of them knew enough26 (76)12 (13)≤0·00125 (67)13 (15)≤0·001
 Most of them knew enough4 (12)32 (35)9 (24)27 (31)
 Only some knew enough2 (6)32 (35)1 (3)33 (37)
 None knew enough1 (3)6 (7)1 (3)6 (7)
 Do not know1 (3)9 (10)1 (3)9 (10)
How much of the time was the deceased treated with respect and dignity by hospital staff?
 Always29 (83)24 (26)≤0·00132 (86)21 (23)≤0·001
 Most of the time4 (11)36 (39)4 (11)36 (40)
 Some of the time1 (3)18 (20)1 (3)18 (20)
 Never0 (0)6 (6)0 (0)6 (7)
 Do not know1 (3)8 (9)0 (0)9 (10)
Were you involved I decisions about the treatment and care of the deceased as much as you wanted?
 Involved as much as I wanted27 (78)38 (42)0·00128 (76)37 (41)0·001
 Would have liked to be more involved4 (11)42 (46)5 (13)41 (46)
 Would have liked to be less involved0 (0)0 (0)0 (0)0 (0)
 No decisions were needed4 (11)6 (7)4 (11)6 (7)
 Do not know0 (0)5 (5)0 (0)5 (6)
Was there any decision made about the care or treatment of the deceased that they would not have wanted?
 No28 (82)40 (45)0·00126 (74)42 (48)0·003
 Yes2 (6)26 (29)1 (3)27 (31)
 Do not know4 (12)23 (26)8 (23)19 (21)
Relationship to the deceased
 Spouse or partner10 (29)10 (13)0·0389 (26)11 (14)0·152
 Other71 (24)67 (87)26 (74)65 (86)
Age of the informant
 Under 6018 (51)32 (35)0·08616 (43)34 (38)0·567
 Over 6017 (49)60 (65)21 (57)56 (62)
Care received in the last three days

Eight variables were included to assess the characteristics of care received in the last 3 days of life of the deceased person (Table 4).

Table 4.   Univariate analyses of the care received from health and social services in the last 3 days of life of those that had spent time in hospital during this time-period
 Overall rating of the care received from health and social services
Excellent, n (%)*Other, n (%)*P value
  1. *Percentages rounded up to the nearest whole number.

  2. P value relates to a chi-squared test where the responses to the questions have been dichotomized into the most positive response compared with all other possible responses for that question (hence relate to a 2 × 2 table)

Was there enough help available to meet the personal care needs of deceased?
 Yes all of the time42 (91)38 (39)≤0·001
 Yes some of the time 1 (2)34 (35)
 No0 (0)13 (14)
 She did not need help with these things3 (7)3 (3)
 Do not know0 (0)9 (9)
Was there enough help available to meet the nursing care needs of the deceased?
 Yes all of the time42 (90)43 (44)≤0·001
 Yes some of the time 2 (4)31 (32)
 No0 (0)11 (11)
 She did not need help with these things3 (6)1 (1)
 Do not know0 (0)12 (12)
Were you involved in decisions about the treatment and care of the deceased as much as you wanted?
 Involved as much as I wanted41 (84)42 (44)≤0·001
 Would have liked to be more involved3 (6)37 (39)
 Would have liked to be less involved0 (0)0 (0)
 No decisions were needed5 (10)10 (11)
 Do not know 0 (0)6 (6)
Was there any decision made about the care or treatment of the deceased that they would not have wanted?
 No36 (74)47 (49)0·037
 Yes3 (6)17 (18)
 No decisions were needed4 (8)9 (10)
 Do not know6 (12)22 (23)
Was there any decision made about the care or treatment of the deceased that you would not have wanted?
 No42 (86)56 (58)0·013
 Yes3 (6)19 (20)
 No decisions were needed2 (4)10 (11)
 Do not know2 (4)10 (11)
How much of the time was the deceased treated with respect and dignity by the staff who cared for her?
 Always49 (100)35 (36)≤0·001
 Most of the time 0 (0)36 (37)
 Some of the time 0 (0)12 (12·5)
 Never0 (0)2 (2)
 Do not know 0 (0)12 (12·5)
Were you told that she was likely to die shortly?
 Yes43 (88)58 (61)0·001
 No6 (12)37 (39)
Do you feel that the deceased died in the right place?
 Yes, it was the right place46 (94)65 (68)0·004
 No, it was not the right place1 (2)22 (23)
 Not sure 2 (4)8 (8)
 Do not know0 (0)1 (1)

Data analysis

Statistical analyses were performed using spss for Windows software (version 14). The analysis was divided into two phases: univariate (phase I) and multivariate (phase II). For phase I, all variables associated with the characteristics of the service provided to the deceased person (e.g. the hospital care received in the last 3 months and the care received in the last 3 days of life variables) were tested for their association with satisfaction with care (excellent/not excellent) using the Pearson chi-square test.

All independent variables that were found to be statistically significantly associated with satisfaction with care from the chi-square test at the 10% level were subjected to a multicollinearity check prior to inclusion in the logistic regression analyses (phase II).

The selection of the variables included for phase II was based on a lax criterion (10%) because some of the variables might contribute to multivariate models in unforeseen ways because of complex interrelationships among themselves (Kirkwood & Sterne 2003).

For simplicity, and because of the small sample and response for some categories, all variables were dichotomized for chi-square tests and for inclusion in the logistic regression into the most positive response to the question vs. all other responses. Relationship of the informant to the deceased person was dichotomized into ‘spouse or partner’ and ‘other’ because of previous work by Seale (1995) and Fakhoury et al. (1996). Categorical variables with responses of ‘Do not know’, ‘No decisions were needed’ or ‘The deceased did not need any help’ were included in the ‘all other responses’ category. Frequencies across all categories of response are presented in Tables 1–4, and the P values relate to the dichotomized variable.

Table 2.   Ratings of satisfaction with care and place of death as reported by the informant
 n (%)*
  1. *Percentages rounded up to the nearest whole number.

The care the deceased person got from doctors in hospital was
 Excellent35 (21)
 Good49 (30)
 Fair34 (21)
 Poor9 (5)
 Missing data38 (23)
The care the deceased person got from nurses in hospital was
 Excellent37 (22)
 Good47 (29)
 Fair30 (18)
 Poor13 (8)
 Missing data38 (23)
The care the deceased person received in the last three days of life was
 Excellent49 (30)
 Good55 (33)
 Fair25 (15)
 Poor18 (11)
 Missing data18 (11)
Place of death
 Nursing/residential care home38 (23)
 Hospital – stroke unit33 (20)
 Hospital – elder care ward43 (26)
 Hospital – other51 (31)

Multivariate backward stepwise logistic regression analysis was performed to explore the relationship between the characteristics of services and overall satisfaction with care [excellent/not excellent (Menard 2002)]. A backward stepwise regression was chosen because the analysis begins with a full model of variables and eliminates the independent variables in an iterative process. When no more variables can be eliminated from the model, the analysis is complete. As a result of the small sample size, it was not possible to assess interactions in the logistic regression modelling process.

Results

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

Of the 493 randomly selected informants, 183 informants returned completed questionnaires, a response rate of 37%. The sub-sample included in the analyses presented here was based on those informants who reported that the deceased person had died in an institutional setting such as a hospital or nursing or residential care home (91%; = 165). This sub-sample was used to explore informants’ experiences of hospital care in the last 3 months and health and social care services in the last 3 days of life. Of those who were reported to have died in an institutional setting, 76% (= 126) were reported to have spent time in hospital during the last 3 months of life. During the last 3 days of life, 64% (= 105) were reported to have been in hospital, 25% (= 41) in a nursing or residential care home and 5% (= 9) were at home. The sub-sample reported in this paper did not differ statistically significantly from the main sample data in terms of the age, sex, relationship of the informant to the deceased person or the duration of illness of the deceased person.

Characteristics of the sub-sample

Deceased persons

Sixty-six per cent of the deceased persons were reported to be aged over 80 years at the time of death, 39% were male and 82% were White British. During the last 3 days of life, 26% (= 43) of informants reported that the deceased person was unconsciousness all of the time. Forty-six per cent (= 76) reported that the deceased person had been ill for more than 6 months, in contrast to the 3% (= 5) who were reported to have died suddenly.

Informants

Sixty-one per cent of informants were aged between 50 and 69 years, 41% were male and 87% were White British. Fifty-five per cent were sons or daughters of the deceased persons and 15% were spouses or partners (Table 1).

Within the sub-sample, 20% were reported to have died on a stroke unit, 26% on an elder care ward, 31% elsewhere in the hospital and 23% in a nursing or residential care home. Although 67% (= 111) of informants were told that the deceased person was likely to die shortly, only 30% (= 33) were with the deceased person when they had died. Sixty-one per cent (= 100) reported that they had enough privacy when they were told that the deceased person was going to die shortly, and 76% (= 125) reported that they felt that the deceased person had died in the right place (Table 2).

Rating satisfaction with care

Informants were asked to report their overall satisfaction with care by giving a rating of excellent, good, fair or poor. Twenty-one per cent rated overall care received from doctors as excellent compared with 22% for nurses during the last 3 months of life of the deceased person. Thirty per cent reported that the care received in the last 3 days of life from health and social care services was excellent. Tables 3 and 4 show the frequencies of response to each of the questions by satisfaction of care (excellent/not excellent). They also shows the P value relating to the univariate chi-squared test.

Multicollinearity analyses

Hospital care received in the last three months

For overall satisfaction with hospital doctors during the last 3 months of life, multicollinearity analyses indicated that responses to four questions were highly correlated (>0·50): enough nurses on duty to care for the deceased person, they knew enough about the condition of the deceased person, they treated the deceased person with respect and dignity and were involved in decisions about the treatment and care of the deceased person. This could be as a result of the relatively small sample size and relatively large number of variables, but this meant that all variables could not all be included in the regression model because they were likely to account for similar variance in the outcome. The dignity and respect variable was eliminated from further analyses due to having the lowest tolerance level 0·533 (1 − 0·458). After controlling for this, all other predictor variables had a tolerance level >0·574 (1 − 0·426), including age of the informant and relationship to the deceased person.

Although identical variables were found to be highly correlated with the overall satisfaction with hospital nurses variable, further analyses were required to determine which additional variables contributed to the collinearity problems. After controlling for the respect and dignity variable (tolerance level 0·512, 1 − 0·399), three further variables were still highly correlated: able to discuss any worries or fears about the treatment or tests of the deceased person, enough nurses on duty to care for the deceased person and knew enough about the condition of the deceased person. To determine which variable was more likely to contribute to the collinearity problem, linear regressions were conducted. The procedure revealed that ‘enough nurses on duty to care for the deceased person’ variable had the lowest tolerance level 0·627 (1 − 0·327). After controlling two of the six variables, all other predictor variables had a tolerance level >0·695 (1 − 0·305).

Care received in the last three days

For overall satisfaction with care received in the last 3 days of life, multicollinearity analyses revealed high correlations between enough help available for the deceased person with personal care and nursing care and treated the deceased person with respect and dignity (>0·50). As a result of the high correlation between personal care and nursing care, the nursing care variable was eliminated from the model (tolerance level 0·233, 1 − 0·417). After eliminating this, treated with respect and dignity was still found to be highly correlated (tolerance level 0·582, 1 − 0·418). After eliminating the dignity and respect variable, decisions made about the care or treatment of the deceased person that the informant (tolerance level 0·650, 1 − 0·296) or deceased (tolerance level 0·646, 1 − 0·296) did not want were found to be highly correlated. As a result of the high correlation between these two variables, a decision was made to eliminate the informant variable rather than the deceased person variable. After controlling for these variables, all other predictor variables had a tolerance level >0·705 (1 − 0·295), including the age of the informant.

Logistic regression analyses of satisfaction with care

Hospital doctor

Of the 11 variables entered into the univariate analysis, 8 variables were associated with satisfaction for the overall rating of the care received from doctors in the last hospital admission at < 0·10 (including age of the informant and relationship to the deceased person). One of these variables was eliminated from the model because of multicollinearity, and therefore seven variables were entered into the logistic regression model. Of the seven remaining variables, three remained in the final logistic regression model predicting high satisfaction with care received from hospital doctors. The model showed that high satisfaction with the overall care from hospital doctors was best predicted by whether informants were able to discuss with hospital staff any worries or fears they had about the condition, treatment or tests of the deceased person [odds ratio (OR) 5·97, 95% confidence interval (CI) = 1·75–20·37], whether they felt that the doctors and nurses knew enough about the deceased person’s condition (OR 16·39, 95% CI = 5·09–52·84) and whether they were the spouse or partner of the deceased person (OR 4·27, 95% CI = 1·01–18·10; Table 5).

Table 5.   Logistic regression models of service provider variables predicting informants’ satisfaction with hospital doctors and nurses
VariableRegression coefficient (se)P valueOdds ratio (95% confidence interval)
  1. *For doctors: selection criteria: backward stepwise; number of cases = 104; cases correctly classified = 66·3%; adjusted R2 = 0·426.

  2. For nurses: selection criteria: backward stepwise; number of cases = 120; cases correctly classified = 29·2%; adjusted R2 = 0·305.

Regression one hospital staff – doctors*
 Constant−3·31  
Doctor service characteristics
 Able to discuss worries and fears with hospital staff about deceased condition, treatment or tests (discussed as much as I wanted vs. all other categories)1·79 (0·63)0·0045·97 (1·75–20·37)
 Doctors and nurses knew enough about the deceased’s condition (they knew enough vs. all other categories) 2·80 (0·60)≤0·00116·39 (5·09–52·84)
 Relationship to the deceased (spouse/partner vs. all other categories)1·45 (0·74)0·0494·27 (1·01–18·10)
Regression two hospital staff – nurses
 Constant−2·55  
Nurse service characteristics
 Able to discuss worries and fears with hospital staff about deceased condition, treatment or tests (discussed as much as I wanted vs. all other categories)1·41 (0·51)0·0064·11 (1·51–11·20)
 Doctors and nurses knew enough about the deceased’s condition (they knew enough vs. all other categories)2·17 (0·49)≤0·0018·74 (3·32–23·04)
Hospital nurse

By comparison, six out of the 11 variables were associated with satisfaction with hospital nurses at < 0·10. Of these, two were eliminated due to multicollinearity. As a result, four variables contributed to the logistic regression model and two were found to predict high satisfaction with hospital nurses. The regression model indicated that high satisfaction with the overall care from hospital nurses was best predicted by whether informants were able to discuss with hospital staff any worries or fears they had about the condition, treatment or tests of the deceased person (OR 4·11, 95% CI = 1·51–11·20), and felt that doctors and nurses knew enough about the condition of the deceased person (OR 8·74, 95% CI = 3·32–23·04; Table 5).

Care received in the last three days

Nine out of 13 variables were associated with satisfaction with care received in the last 3 days from health and social services at < 0·10 (including age of the informant). Three variables were eliminated because of multicollinearity. Six variables contributed to the logistic regression model and three were found to predict high satisfaction with health and social care in the last 3 days of life. The model indicated that high satisfaction with care provided by health and social services was best predicted by whether informants reported that there was enough help available to meet the personal care needs of the deceased person all of the time (OR 10·74, 95% CI = 3·37–34·24), those who were involved as much as they wanted in decisions about the treatment and care of the deceased person (OR 3·96, 95% CI = 1·48–10·59) and those who felt that the deceased person died in the right place (OR 5·49, 95% CI = 1·07–28·11; Table 6).

Table 6.   Logistic regression model of service sector variables predicting informants’ satisfaction with health and social services in the last 3 days
VariableRegression coefficient (se)P valueOdds ratio (95% confidence interval)
  1. Selection criteria: backward stepwise; number of cases = 132; cases correctly classified = 62·9%.; adjusted R2 = 0·295.

Health and social care service
 Constant−4·68  
Health and social care service characteristics
 Enough help available to help with personal care needs (yes all of the time vs. all other categories)2·37 (0·59)≤0·00110·74 (3·37–34·24)
 Involved in decisions about the deceased treatment and care (involved as much as I wanted vs. all other categories)1·38 (0·50) 0·0063·96 (1·48–10·59)
 Felt that the deceased died in the right place (yes vs. all other categories)1·70 (0·83) 0·0415·49 (1·07–28·11)

Discussion

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

Our findings suggest that bereaved relatives of people who died following stroke were more likely to rate care as ‘excellent’ if they considered that the deceased person and/or their own individual needs were met during the last 3 months and last 3 days of life. This included being actively involved in discussions with doctors and nurses about the condition of the deceased person, being able to discuss their worries and fears, perceiving that the doctors and nurses know enough about their condition and having enough help available to help with personal care needs.

Study limitations

There were a number of limitations to this study. The data used in this paper are now over five years old, reflecting the vagaries of short term research funding which has meant the write up of this study has been completed in the lead authors own time. While the data reflect care provided in 2003 we would argue that our findings are still of relevance. No more recent studies have been published in this area and while the National Stroke Strategy includes quality, markers for End of Life Care to date there are no detailed plans on how these might be achieved. Our conclusions however, reflect the aims of the National End of Life Care Strategy with its emphasis on patient led care. The response rate was less than that achieved in similar surveys (Addington-Hall et al. 1998, 2006b, Addington-Hall & McCarthy 1995). There are also concerns about the use of proxies in evaluating health care, particularly in relation to recall and the impact of bereaved informants’ emotions on their views. The limited available evidence suggests that informants’ views on satisfaction with care and observable symptoms accord reasonably closely with those of patients (McPherson & Addington-Hall 2003). Whilst there is some evidence that bereaved people do change their views during bereavement, these views appear to be based on their knowledge of the patient, and not on their own distress (McPherson & Addington-Hall 2004). With regard especially to care in the last 3 days of life, it is unlikely that patients themselves would have been able to engage in a research interview or complete a questionnaire (Addington-Hall et al. 2007), either because they were too unwell or because they might not have been identified as likely to die; this makes proxy evaluations the only source of information. The three regression models developed give an indication of the main factors which are associated with excellent care, but they may not be the only factors. The models were developed using a relatively small sample and a relatively large number of variables were considered for inclusion in the model. This was highlighted by the fact that a number of questions had to be excluded from regression analyses because of multicollinerarity problems. The choice of variable to exclude is not straightforward and can lead to misjudgements and therefore mis-specification of model parameters. If two variables are highly correlated and each accounts for similar variance in the outcome, then it is not possible to tell which is important and they could be used interchangeably. Information from a larger number of participants would allow the complex relationships between these variables and excellent care to be assessed more thoroughly.

Stroke care at the end of life

These findings support our earlier qualitative work which showed that complaints from bereaved relatives about end of life care reflect instances in which the deceased person or they themselves had not been treated as individuals and in which they had felt ‘devalued’, ‘dehumanized’ or ‘disempowered’ (Rogers et al. 2000). Importantly, instances in which the expected ‘rules’ of patient/doctor communication were breached were also an important cause of dissatisfaction or complaint (Addington-Hall et al. 1995). Sources of dissatisfaction or complaint with end of life care seem to be similar to those with health care in general (Coyle 1999, Williams et al. 1998).

The UK Health Care Commission has recently highlighted the need for open and on-going communication between healthcare staff and patients’ relatives in end of life care (Healthcare Commission 2006, 2007). This review of second stage complaints about the NHS care showed that relatives were often unaware that their loved one was likely to die, were often not told about their relatives’ impending death soon enough to be with the patient at the time of their death, and were particularly distressed if given conflicting or contradictory information. Additionally, the report highlighted the need for relatives to know that their loved ones’ basic care needs were met and, importantly, that their care had been documented. Our findings indicate the importance of these issues for bereaved relatives regarding overall satisfaction with end of life care of those who died following stroke. Our findings are similar to those reported by Addington-Hall et al. (1995), where people who died following stroke did not receive sufficient help or information from healthcare professionals.

Concerns about the standard of care for dying people in the UK have led to a number of government initiatives, including introduction of the Liverpool Care Pathway in the acute hospital environment (Jack et al. 2003). The Pathway aims to promote best practice in the last 3 days of life by requiring healthcare professionals to address the individual medical, nursing, psychosocial and spiritual needs of patients and those close to them. The Pathway encourages open discussion between professionals, patients and relatives, and requires four- hourly assessment of the patient to be documented. The pathway may go some way to ensuring that individual needs are met and that relatives see evidence of this.

Towards the end of life, many people lack the capacity to make major decisions relating to their health care; this is especially true in stroke, which often renders patients unable to communicate meaningfully. The Department for Constitutional Affairs (2007) requires clinicians to consider the opinions of relatives and friends of such patients when making decisions to withhold or withdraw treatment (BMA 2007). However, clinicians are not bound by the opinions of relatives and must act in what they feel are the best interests of the patient. In some instances, these may be at odds with the views of relatives and therefore disagreements about end of life decisions will probably continue to be a source of dissatisfaction. It is too early to assess the impact of advance directives on the provision of end of life care, but they may reduce the number of instances in which clinicians and relatives disagree (DoH 2007d).

To date, there has been little literature concerning the needs of dying patients after stroke and those close to them (Jack et al. 2004, Shah et al. 2005, Stevens et al. 2007). However, we do know that during the last days of lives their care needs are likely to be similar to those of people dying from other conditions (Addington-Hall 2005, Ellershaw 2007, Mirando et al. 2005). Until this time, however, patients after stroke, especially those who remain unconscious for longer than a week, are likely to present particular challenges to those caring for them. Assessing individual needs, responding to patients as individuals and engaging with those close to them may be especially difficult when patients are unable to engage in normal social interactions and about whom nursing staff may have little or no prior knowledge (Rogers & Addington-Hall 2005). Such patients are easily forgotten in busy acute environments, where the focus of specialist stroke care tends to be on rehabilitation and recovery (Coleman Gross et al. 2001, Hilderley et al. 2002, Nir et al. 2004).

As a result of the acute onset of stroke and the need for prompt medical intervention following appropriate assessment, people with stroke are likely to be admitted to hospital. Our data show that 76% of our sample were admitted to hospital in the last 3 months of life. Up to a third of people experiencing a first stroke will have died within a year. The UK End of Life Care Strategy aims to increase the number of people being cared for and dying in the setting they choose, including their own homes (DoH 2008). Therefore, service providers will increasingly need to commission a range of services and service configurations that are flexible enough to meet individual needs and address the objectives set out in this strategy. To support these, The Stroke Association (2008) has recently launched a campaign entitled ‘Lost without Words’, highlighting some of the key points from the National Stroke Strategy (DoH 2007b). Continuing to highlight the need for people with stroke to have access to services and support reinforces the challenges faced by the current delivery of the ‘Our NHS, Our Future: NHS Next Stage Review’ (2008).

Despite concerns about the theoretical robustness of the concept of satisfaction and patient and proxy ratings of health care, such ratings continue to be an important component of user involvement and service evaluation. Having a clear understanding of the ways in which patients and proxies rate care will be essential to service improvement and development.

Conclusion

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

Our findings suggest that in the last 3 months of life individualized holistic care for both patients and their families, such as that inherent in palliative care, would do much to improve the care of this group. Addressing individual needs is an essential component of nursing care and when caring for dying patients such care should extend its focus to include the needs of those close to the patient, particularly when decisions about end of life care are made.

Author contributions

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References

JAH and AR were responsible for the study conception and design. AR performed the data collection. AY performed the data analysis. AY was responsible for the drafting of the manuscript. AR, JAH and AY made critical revisions to the paper for important intellectual content. AY and LD provided statistical expertise. JAH obtained funding. AY provided administrative, technical or material support. JAH and AR supervised the study.

References

  1. Top of page
  2. Abstract
  3. What is already known about this topic
  4. What this paper adds
  5. Implications for practice and/or policy
  6. Introduction
  7. The study
  8. Results
  9. Discussion
  10. Conclusion
  11. Acknowledgement
  12. Conflict of interest
  13. Funding
  14. Author contributions
  15. References
  • Addington-Hall J. (2005) Extending palliative care to chronic conditions. European Journal of Palliative Care 12(2), 1417.
  • Addington-Hall J. & McCarthy M. (1995) Regional study of care for the dying: methods and sample characteristics. Palliative Medicine 9, 2735.
  • Addington-Hall J., Lay M., Altmann D. & McCarthy M. (1995) Symptom control, communication with health professionals, and hospital care of stroke survivors in the last year of life as reported by surviving family, friends and officials. Stroke 26(12), 2242.
  • Addington-Hall J., Walker L., Jones C., Karlsen S. & McCarthy M. (1998) A randomised controlled trial of postal versus interviewer administration of a questionnaire measuring satisfaction with, and use of, services received in the year before death. Journal of Epidemiology Community Health 52, 802807.
  • Addington-Hall J., Rogers A., Saleem T. & Young A. (2006a) Development of a Postal Questionnaire for Use in a Population Based Survey of the Experiences of People Who Die from Stroke and Their Families, from the Perspective of Bereaved Relatives. University of Southampton for Department of Health, London.
  • Addington-Hall J., Shipman C., Burt J., Ream E., Beynon T. & Richardson A. (2006b) Evaluation of the Education and Support Programme for District and Community Nurses in the Principles and Practice of Palliative Care. University of Southampton for Department of Health, London.
  • Addington-Hall J., Eduardo B., Higginson I. & Payne S. (2007) Research Methods in Palliative Care. Oxford University Press, London.
  • Alazri M. & Neal R. (2003) The association between satisfaction with care provided in primary care and outcomes in type 2 diabetes mellitus. Diabetes Medicine 20, 486490.
  • Aspinal F., Addington-Hall J., Hughes R. & Higginson I. (2003) Using satisfaction to measure the quality of palliative care: a review of the literature. Journal of Advanced Nursing 42(4), 324339.
  • Bikker A. & Thompson A. (2006) Predicting and comparing patient satisfaction in four different modes of health care across a nation. Social Science & Medicine 63, 16711683.
  • BMA (2007) Withholding and Withdrawing Life-prolonging Medical Treatment. Blackwell Publishing, Oxford.
  • Carr-Hill R. (1992) The measurement of patient satisfaction. Journal of Public Health Medicine 14(3), 236249.
  • Coleman Gross J., Goodrich S., Kain M. & Faulkner E. (2001) Determining stroke rehabilitation inpatients’ level of nursing care. Clinical Nursing Research 10(1), 4051.
  • Coyle J. (1999) Exploring the meaning of ‘dissatisfaction’ with health care: the importance of ‘personal identity threat’. Sociology Health & Illness 21(1), 95124.
  • Department for Constitutional Affairs (2007) Mental Capacity Act 2005: Code of Practice. The Stationery Office, London.
  • Department of Health (1999) Health Act 1999. The Stationery Office, London.
  • Department of Health (2001) National Service Framework for Older People. Department of Health, London.
  • Department of Health (2007a) Our NHS Our Future: NHS Next Stage Review Interim Report. Department of Health, London.
  • Department of Health (2007b) National Stroke Strategy. Department of Health, London.
  • Department of Health (2007c) A New Ambition for Stroke: A Consultation on a National Strategy. Department of Health, London.
  • Department of Health (2007d) Advance Care Planning: A Guide for Health and Social Care Staff. Department of Health, London.
  • Department of Health (2008) Our NHS Our Future: NHS Next Stage Review Leading Local Change. Department of Health, London.
  • Ellershaw J. (2007) Care of the dying: what a difference an LCP makes! Palliative Medicine 21, 365368.
  • Fakhoury W. (1998) Satisfaction with palliative care: what should we be aware of? International Journal of Nursing Studies 35, 171176.
  • Fakhoury W., McCarthy M. & Addington-Hall J. (1996) Determinants of informal caregivers’ satisfaction with care for dying cancer patients. Social Science & Medicine 42(5), 721731.
  • Fakhoury W., McCarthy M. & Addington-Hall J. (1997) The effects of clinical characteristics of dying cancer patients on informal caregivers’ satisfaction with palliative care. Palliative Medicine 11, 107115.
  • Fan V., Burman M., McDonell M. & Fihn S. (2005) Continuity of care and other determinants of patient satisfaction with primary care. Journal of General Internal Medicine 20(3), 226233.
  • Healthcare Commission (2005) Survey of Patient 2005: Stroke. Healthcare Commission, London.
  • Healthcare Commission (2006) Survey of Patients 2006: Caring for People After They Have had a Stroke. Healthcare Commission, London.
  • Healthcare Commission (2007) Caring for Dignity: A National Report on Dignity in Care for Older People While in Hospital. Healthcare Commission, London.
  • Heyland D., Dodek P., Rocker G., Groll D., Gafni A. & Pichora D. (2006) What matters most in end-of-life care: perceptions of seriously ill patients and their family members. Canadian Medical Association Journal 174(5), 627633.
  • Hilderley L., Iwamoto R. & Knobf T. (2002) Quality of life through rehabilitation at end of life. Cancer Practice 10(4), 174178.
  • Intercollegiate Stroke Working Party (2004) National Clinical Guidelines for Stroke. Royal College of Physicians, London.
  • Intercollegiate Stroke Working Party (2005) National Sentinel Stroke Audit 2004. Royal College of Physicians, London.
  • Intercollegiate Stroke Working Party (2007) National Sentinel Stroke Audit 2006. Royal College of Physicians, London.
  • Jack B., Gambles M., Murphy D. & Ellershaw J. (2003) Nurses’ perceptions of the Liverpool care pathway for the dying patient in the acute hospital setting. International Journal of Palliative Nursing 9(9), 375381.
  • Jack C., Jones L., Jack B., Gambles M., Murphy D. & Ellershaw J. (2004) Towards a good death:the impact of the care of the dying pathway in an acute stroke unit. Age and Ageing 33, 625635.
  • Jackson J., Chamberlin J. & Kroenke K. (2001) Predictors of patient satisfaction. Social Science & Medicine 52, 609620.
  • Jenkinson C., Coulter A., Bruster S., Richards N. & Chandola T. (2002) Patients’ experiences and satisfaction with health care: results of a questionnaire study of specific aspects of care. Quality & Safety in Health Care 11, 335339.
  • Kirkwood B. & Sterne J. (2003) Essential Medical Statistics, 2nd edn. Blackwell Publishers, Oxford.
  • Leatherman S., Sutherland K. & Airoldi M. (2008) Bridging the Quality Gap: Stroke. The Health Foundation, London.
  • McPherson C. & Addington-Hall J. (2003) Judging the quality of care at the end if life: can proxies provide reliable information? Social Science & Medicine 56, 95109.
  • McPherson C. & Addington-Hall J. (2004) How do proxies’ perceptions of patients’ pain, anxiety and depression change during the bereavement period? Journal of Palliative Care 20(1), 1219.
  • Mead N., Bower P. & Hann M. (2002) The impact of general practitioners’ patient-centredness on patients’ post-consultation satisfaction and enablement. Social Science & Medicine 55, 283299.
  • Menard S. (2002) Applied Logistic Regression Analysis, 2nd edn. Sage, California.
  • Mirando S., Davies P.D. & Lipp A. (2005) Introducing an integrated care pathway for the last days of life. Palliative Medicine 19(1), 3339.
  • Nir Z., Zolotogorsky Z. & Sugarman H. (2004) Structured nursing intervention versus routine rehabilitation after stroke. American Journal of Physical Medicine & Rehabilitation 83(7), 522529.
  • Office for National Statistics (2006) Heart Disease Leading Cause of Death in England and Wales. National Statistics, London.
  • Rogers A. & Addington-Hall J. (2005) Care of the dying stroke patient in an acute setting. Journal of Research in Nursing 10(2), 153167.
  • Rogers A., Karlsen S. & Addington-Hall J. (2000) “All the services were excellent. It is when the human element comes in that things go wrong”: dissatisfaction with hospital care in the last year of life. Journal of Advanced Nursing 31, 768774.
  • Seale C. (1995) Dying alone. Sociology Health & Illness 17(3), 376392.
  • Shah S.H., Gambles M., Jack B. & Ellershaw J. (2005) The Liverpool care pathway: its impact on improving the care of the dying. Age and Ageing 34(2), 197199.
  • Stevens T., Payne S., Burton C., Addington-Hall J. & Jones A. (2007) Palliative care in stroke: a critical review of the literature. Palliative Medicine 21, 323331.
  • Suhonen R., Valimaki M. & Leino-Kilpi H. (2005) Individualized care, quality of life and satisfaction with nursing care. Journal of Advanced Nursing 50(3), 283292.
  • The Stroke Association (2008) Lost Without Words: Campaign Report. The Stroke Association, London.
  • Williams B., Coyle J. & Healy D. (1998) The meaning of patient satisfaction: an explanation of high reported levels. Social Science & Medicine 47(9), 13511359.
  • Wilson A., Wynn A. & Parker H. (2002) Patient and carer satisfaction with ‘Hospital at Home’: quantitative and qualitative results from a randomised controlled trial. British Journal of General Practice 52(474), 913.
  • Young A., Rogers A. & Addington-Hall J. (2008) The quality and adequacy of care received at home in the last three months of life by people who died following a stroke: a retrospective survey of surviving family and friends using the VOICES questionnaire. Health and Social Care in the Community 16(4), 419428.