Judging the quality of care at the end of life: can proxies provide reliable information?

https://doi.org/10.1016/S0277-9536(02)00011-4Get rights and content

Abstract

A major challenge in research into care at the end of life is the difficulty of obtaining the views and experiences of representative samples of patients. Studies relying on patients’ accounts prior to death are potentially biased, as they only represent that proportion of patients with an identifiable terminal illness, who are relatively well and therefore able to participate, and who are willing to take part. An alternative approach that overcomes many of these problems is the retrospective or ‘after death’ approach. Here, observations are gathered from proxies, usually the patient's next of kin, following the patient's death. However, questions have been raised about the validity of proxies’ responses. This paper provides a comprehensive review of studies that have compared patient and proxy views. The evidence suggests that proxies can reliably report on the quality of services, and on observable symptoms. Agreement is poorest for subjective aspects of the patient's experience, such as pain, anxiety and depression. The findings are discussed in relation to literature drawn from survey methodology, psychology, health and palliative care. In addition to this, factors likely to affect levels of agreement are identified. Amongst these are factors associated with the patient and proxy, the measures used to assess palliative care and the quality of the research evaluating the validity of proxies’ reports. As proxies are a vital source of information, and for some patients the only source, the paper highlights the need for further research to improve the validity of proxies’ reports.

Introduction

The provision of hospice and palliative care for patients at the end of life is growing. Recent estimates in the UK show that around 19% of patients with cancer die in a hospice and approximately 60% of cancer patients who die are cared for by palliative home care teams (The Hospice Information Service, 2000). Given the prevailing ethos of evidence-based practice and economic constraints, the provision of palliative care, like other areas of health care, has had to respond to demands for efficiency and effectiveness. Such demands have resulted in a move toward a greater evaluation of services and to clinical audit. Audit within palliative care serves various essential functions beyond economic concerns, including identifying patient and family needs, assessing palliative care outcomes such as quality of life, monitoring standards and identifying training needs (Higginson, 1993).

When evaluating services for patients at the end of life and their families, researchers are confronted with a number of methodological challenges. One problem is the difficulty of ascertaining the views and experiences of terminally ill patients. Apart from ethical concerns about interviewing patients, studies relying on the patient's accounts prior to death, the prospective approach, are fundamentally biased as they only represent the proportion of the population that is less impaired and therefore able, as well as willing, to participate. A decline in performance status is seen at the end of life in many cancer patients and is a good indicator of prognosis (Vigano, Dorgan, Buckingham, Bruera, & Suarez-Almazor, 2000). It is at this point in time that assessments of patients’ quality of life are most needed (Aaronson, 1991). Prospective studies are also limited to diseases such as cancer that have an identifiable terminal phase, and where patients are likely to be known to relevant services (Morris et al., 1986). Added to this, high attrition rates, small sample sizes and difficulties associated with predicting survival rates can also hamper attempts to conduct valid research (McWhinney, Bass, & Donner, 1994; Rinck et al., 1997).

An alternative approach that overcomes many of the problems associated with prospective studies is the retrospective or ‘after death’ approach. Here, observations are gathered from proxies acting on behalf of the patient, after the patient's death. Usually this is the patient's relative, friend or a health professional involved in the patient's care. This approach has been used in a number of influential studies on care of the dying (Cartwright & Seale, 1990; Seale & Cartwright, 1994; Addington-Hall & McCarthy, 1995). The largest survey of its kind in the United Kingdom is the Regional Study of the Care of the Dying (RSCD) (Addington-Hall & McCarthy, 1995). Being retrospective, the authors were able to identify those individuals who had died within a specific time period; in this case the last quarter of 1990. From these were extracted representative samples of 270 deaths from each of 20 different health districts in England. Selecting patients solely according to time and place of death avoided creating an inherent bias towards any particular cause of death. Thus, deaths from different causes could be represented (Addington-Hall, Lay, Altmann, & McCarthy, 1995; McCarthy, Lay, & Addington-Hall, 1996; McCarthy, Addington-Hall, & Altmann, 1997). The results of the RSCD show that current health care and social services in England are in need of significant improvement before they can fully meet the needs of terminally ill patients and their families. Similar methods have been used in other countries, for example in the SUPPORT study where bereaved relatives provided information on patients’ last three days of life (Lynn et al., 1997). These studies have provided important insights into death and dying in late 20th century.

However, collecting information indirectly from proxies at some later point in time introduces potential problems regarding the validity of reports (Higginson, Priest, & McCarthy, 1994; Hinton, 1996). Validity, in this context, refers to the degree to which proxy accounts accurately reflect the experiences of the individual they are representing. At present, little is known about the accuracy of proxies’ responses, despite the fact that inaccuracy may compromise validity and lead to erroneous conclusions. Given that proxies are a significant source of information and are often the only available source for a large proportion of dying patients, it is vital to understand how and why their accounts might differ from those of the patient. This paper provides a comprehensive and critical review of studies that assess the validity of proxies acting for patients at the end of life. The emphasis is on significant others as proxies for the patient. These include family members, friends or neighbours, but exclude health professionals. This is because the focus of retrospective surveys reported to date, rely on significant others acting as proxies1 rather than health professionals. Relevant literature from palliative and health care, survey methodology and psychology is drawn upon to examine factors likely to affect the validity of proxies’ reports.

Section snippets

Methods for testing the validity of proxies

A systematic review of the literature, using a recognised search strategy (Centre for Reviews and Dissemination, 1996), revealed a large volume of papers on patient–proxy agreement. This review only includes studies that investigate outcomes relevant to care at the end of life, where at least some of the subjects have an advanced disease or a terminal illness. Excluded from the review are all those studies that concern patient–proxy agreement involving children and adolescents and patients with

Levels of agreement between the patient and proxy

In spite of the heterogeneity of the research assessing the validity of proxies’ reports, there is some consistency across the literature. The evidence suggests that information that relies on concrete, observable phenomenon, such as service provision, service evaluation and factual information, tends to have good levels of agreement (Clipp & George, 1992; Farrow & Samet, 1990; Field et al., 1995; Higginson & McCarthy, 1993; O’Brien & Francis, 1988; Spiller & Alexander, 1993). This agreement is

Patient and proxy demographics

Evidence for a relation between levels of agreement between patients and proxies, and demographic information such as age, ethnicity, gender, socio-economic status and educational level have been inconsistent. Those studies that showed that demographics do have a bearing on agreement found the influence to be mixed (O’Brien & Francis, 1988) or a function of the symptom evaluated (Kristjanson et al., 1998). For example, older proxies were less accurate in reporting pain frequency than younger

Assessment instrument

Responding to a question is a complex cognitive process, particularly if it requires making judgements regarding other people's experiences. It consists not only of attending, perceiving and encoding information pertaining to an event, but also recalling material and judging its appropriateness in relation to the question and the response format. Errors in responding can occur at any stage, and some degree of inaccuracy in proxies’ reports can be attributed to the instrument itself. As Sudman

Factors associated with the validity studies

The research presented so far has summarised the many factors that can potentially affect the validity of proxies’ reports. This section deals with the quality of the studies that have investigated this aspect.

Summary and conclusions

This review provides encouraging evidence for the use of proxies at the end of life, for some aspects of the patient's experience, most notably for service provision and evaluation, and for symptoms that are more observable in nature. However, care must be taken when using proxies to report on aspects of the patient's experience that are more subjective, such as pain and affective states. The findings are consistent with Sprangers and Aaronson's (1992) review of proxies for patients with

Acknowledgements

This work was supported by a grant from the Joint Research Committee of King's College School of Medicine and Dentistry and King's College NHS Trust.

References (93)

  • J.N. Morris et al.

    The effect of treatment setting and patient characteristics on pain in terminal cancer patientsA report from the national hospice study

    Journal of Chronic Diseases

    (1986)
  • J. O’Brien et al.

    The use of next-of-kin to estimate pain in cancer patients

    Pain

    (1988)
  • K.C. Sneeuw et al.

    Comparison of patient and proxy EORTC QLQ-C30 ratings in assessing the quality of life of cancer patients

    Journal of Clinical Epidemiology

    (1998)
  • M.A. Sprangers et al.

    The role of health care providers and significant others in evaluating the quality of life of patients with chronic disease

    Journal of Clinical Epidemiology

    (1992)
  • R.J. Stancliffe

    Proxy respondents and quality of life

    Evaluation and Program Planning

    (2000)
  • A. Tversky et al.

    AvailabilityA heuristic for judging frequency and probability

    Cognitive Psychology

    (1973)
  • N.K. Aaronson

    Methodological issues in assessing the quality of life of cancer patients

    Cancer

    (1991)
  • J. Addington-Hall et al.

    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

    Stroke

    (1995)
  • J. Addington-Hall et al.

    Regional study of care for the dyingMethods and sample characteristics

    Palliative Medicine

    (1995)
  • S. Ahmedzai et al.

    Quality of death from lung cancerPatients’ reports and relatives’ retrospective opinions

  • A.M. Armstrong et al.

    Errors in memory for dietary intake and their reduction

    Applied Cognitive Psychology

    (2000)
  • A. Baddley

    Human memoryTheory and practice

    (1990)
  • L. Baider et al.

    Couples in crisisPatient–spouse differences in perception of interaction patterns and illness situation

    Family Therapy

    (1984)
  • S.S. Bassett et al.

    Reliability of proxy response on mental health indices for aged, community-dwelling women

    Psychology & Aging

    (1990)
  • P.H. Blaney

    Affect and memoryA review

    Psychological Bulletin

    (1986)
  • J.M. Blazeby et al.

    Observer variation in assessment of quality of life in patients with oesophageal cancer

    British Journal of Surgery

    (1995)
  • G.H. Bower

    Mood and memory

    American Psychologist

    (1981)
  • G.H. Bower

    How might emotion affect learning?

  • G. Bower et al.

    Scripts in memory for text

    Cognitive Psychology

    (1979)
  • D. Broadbent et al.

    Anxiety and attentional bias “State and trait”

    Cognition and Emotion

    (1988)
  • T.S. Brugha

    Depression in the terminally ill

    British Journal of Hospital Medicine

    (1993)
  • M. Burke et al.

    Autobiographical memory and clinical anxiety

    Cognition and Emotion

    (1992)
  • A. Cartwright et al.

    Life before death

    (1973)
  • A. Cartwright et al.

    The natural history of the surveyAn account of the methodological issues encountered in a study of life before death

    (1990)
  • S.-A. Christianson

    Emotional stress and eyewitness memoryA critical review

    Psychological Bulletin

    (1992)
  • S.-A. Christianson et al.

    Memory for traumatic events

    Applied Cognitive Psychology

    (1987)
  • CRD (NHS Centre for Reviews and Dissemination). (1996). Undertaking systematic reviews of research on effectiveness:...
  • E.C. Clipp et al.

    Patients with cancer and their spouse caregivers. Perceptions of the illness experience

    Cancer

    (1992)
  • J. Cohen

    A coefficient of agreement for nominal scales

    Educational and Psychological Measurement

    (1960)
  • J. Cohen

    Weighted kappaNominal scale agreement with provision for scaled disagreement or partial credit

    Psychological Bulletin

    (1968)
  • M.A. Conway

    A structural model of autobiographical memory

  • A.E. Curtis et al.

    Quality of life of oncology hospice patientsA comparison of patient and primary caregiver reports

    Oncology Nursing Forum

    (1989)
  • D.C. Farrow et al.

    Comparability of information provided by elderly cancer patients and surrogates regarding health and functional status, social network, and life events

    Epidemiology

    (1990)
  • D. Field et al.

    Terminal illnessViews of patients and their lay carers

    Palliative Medicine

    (1995)
  • Glickman, M. (1997). Making palliative care better: Quality improvement, multi-professional audit and standards....
  • I.H. Gotlib et al.

    Construct accessibility and clinical depressionA longitudinal investigation

    Journal of Abnormal Psychology

    (1987)
  • Cited by (299)

    • Measuring Goal-Concordant Care in Palliative Care Research

      2021, Journal of Pain and Symptom Management
    View all citing articles on Scopus
    View full text