The development of QoL measures

UK mental health policy continues to emphasise the routine measurement of outcomes as a means of enabling a more systematic approach to improving quality of care. However, I think that clinicians remain as reluctant to integrate outcome measurement into their practice as they were over a decade ago (Reference Gilbody, House and SheldonGilbody 2002), perhaps for one of the reasons that Gilbody et al suggested: the continuing lack of evidence of the value of the available measures.

This article addresses QoL measurement only, not the wider range of outcome measures in use, such as, for example, the Mental Health Recovery Star (Mental Health Providers Forum 2008). Quality of life is a descriptive term for people’s emotional, social and physical well-being, and their ability to function in carrying out the ordinary tasks of living. The World Health Organization (WHO) has developed an international QoL measure (WHOQOL-BREF; Box 1) and define QoL as:

Reference Ruggeri, Bisoffi and FontecedroRuggeri et al (2001) say that whereas the WHO definition is subjective, other QoL constructs include objective indicators of health, housing and other material circumstances.

Reference BasuBasu (2004) traces concepts of QoL in medicine to the early 1960s, with major work on developing systems of QoL measurement from the 1980s onwards, including mental health-specific measures.

The QoL concept can enable systematic investigation of the social, emotional and physical effects of illnesses and treatments on people’s daily lives, from objective and subjective perspectives. In theory, well-constructed, reliable, valid and responsive QoL measures can help doctors and healthcare providers, families and patients, to choose between different treatment approaches and monitor outcomes. The reality may not match the theory, particularly from the perspective of service users.

In their selective literature review of studies of QoL in severe mental illness, Reference Holloway and CarsonHolloway & Carson (2002) list four focuses of investigation:

1. 1 objective life circumstances

2. 2 the individual’s subjective appraisal of their life

3. 3 overall health status (health-related quality of life: HRQoL)

4. 4 health economics.

The second and third approaches are of concern in this article. Objective measures of QoL are generally integrated with subjective measures in tools developed since the 1980s and the health-economic approach (e.g. quality-adjusted life-years, or QALYs) is not covered here. It should be distinguished from measures of QoL.

This article traces some of the controversies and questions arising in the past 30 years of QoL measurement from a service user perspective.

Theoretical basis for QoL

It is difficult and controversial to define and measure quality of life, an essentially subjective and personal matter. The contested area of mental health and illness presents particular problems, such as whose values are these measurements based on, and what is being measured, by whom and for whom?

One of the problems with QoL measurement in mental health is the lack of an overt theoretical basis. This is particularly true of the health-related and disease-specific QoL measures:

Reference Barry and ZissiBarry & Zissi (1997) similarly critique the lack of attention to methodological and theoretical issues in QoL studies:

QoL and HRQoL in mental health

According to Reference Barry and ZissiBarry & Zissi (1997), the first large-scale mental health QoL studies were carried out in the USA, in the context of the transition to community care. Reference Lehman, Ward and LinnLehman et al (1982) linked objective domains with a subjective approach, developing the Quality of Life Interview (QoLI), a structured self-report tool. It contains a global measure of life satisfaction, plus measures of objective and subjective quality of life in the domains of living situation, daily activities, family and social relations, finances, work and school, legal and safety issues, and health. Information was obtained from respondents about their objective quality of life and then about their level of satisfaction in each life area, pairing objective and subjective results for each person.

Many QoL mental health measures, for example the Oregon Quality of Life Questionnaire (Reference Bigelow, McFarland and OlsonBigelow 1991), the Lancashire Quality of Life Profile (Reference Oliver, Huxley and PriebeOliver 1997) and Manchester Short Assessment of Quality of Life (Reference Priebe, Huxley and KnightPriebe 1999) were influenced by Lehman’s work. They are close in spirit to generic QoL measurements such as WHOQOL (WHOQOL Group 1994), although they are adapted for mental health and are often used alongside other measures.

Reference Holloway and CarsonHolloway & Carson (2002) describe this QoL approach as containing three domains: (1) personal characteristics, (2) objective life conditions and (3) subjective appraisal of life; and examining the impact of these on the global quality of life of the person or group. One aim of this approach is to find out how mental health services can improve people’s well-being as they move from institutional to community living.

The HRQoL approach takes the generic concept of QoL and adapts it to specific questions of disease. Reference LehmanLehman (1996) describes the purpose of one such HRQoL-type scale for schizophrenia as to assess the negative symptoms of schizophrenia and therefore not belonging to mainstream QoL measures. Reference Holloway and CarsonHolloway & Carson (2002) describe the HRQoL approach as containing the domains of: (1) psychosocial performance, (2) side-effects of medication and (3) symptoms; and assessing the impact of these on the global quality of life of the person or group.

Applying the QoL approach in mental health

Applying the HRQoL approach in mental health

Suitability and validity of QoL/HRQoL tools to measure outcomes

Reference Barry and ZissiBarry & Zissi (1997) argue that most studies (at that date) were cross-sectional, not longitudinal, and could not therefore demonstrate the role of QoL in measuring outcomes of service changes. In response, Reference Ruggeri, Bisoffi and FontecedroRuggeri et al (2001) carried out a 2-year longitudinal outcome study using the Lancashire Quality of Life Profile. They found difficulty in building predictive models around subjective outcomes because of the tendency towards ‘psychological adaptation or “response shift” that can occur over time in the subjective appraisal of a person’s current state … the multifactorial determinance of subjective outcomes and the diverse reaction of different individuals to the same circumstances’.

Reference Holloway and CarsonHolloway & Carson (2002) similarly find that QoL’s generic approach has poor sensitivity to change, which they say may result from ‘accommodation to adversity’ leading to lowered life expectations. They argue that subjective QoL will always be problematic because the salience of life domains will vary between individuals, and even within individuals over time, depending on people’s expectations, aspirations, self-appraisal, coping strategies and life experiences.

HRQoL measures also have their problems. Reference Gilbody, House and SheldonGilbody et al (2002) argue that they are designed to evaluate healthcare at a population level, not for making decisions about individual patients. Also, as mentioned above, HRQoL measures are rarely patient-centred, which Reference Carr and HigginsonCarr & Higginson (2001) argue reduces their value as outcome measures:

Can people give accurate self-reports when ill?

A frequent question about all forms of QoL measurement in mental health is whether patients’ self-reports of their quality of life are credible, when they may have depression, schizophrenia or mania, and whether more objective reports by others should be used:

Other sources argue that QoL is necessarily a primarily subjective measure and that the patient’s viewpoint should therefore be accepted as valid:

Reference Oliver, Huxley and PriebeOliver et al (1997) found on the basis of their data-set of more than 1500 patients that mental health problems (other than depression) do not significantly alter the results from QoL surveys. They agree with Reference Lehman, Ward and LinnLehman et al’s (1982) earlier verdict that clinical depression lowers subjective well-being scores in all domains:

Reference KuehnerKuehner (2002) studied the validity of subjective QoL measures with depressed patients, comparing patients at different stages of depression with control groups using the Lancashire Quality of Life Profile. She found, as did Reference Oliver, Huxley and PriebeOliver et al (1997), that global measures of subjective QoL (general life satisfaction and Cantril’s Ladder) were closely affected by concurrent levels of depression. Kuehner concluded that global assessments of quality of life mirror depression, remitting as symptoms remit. However, when patients and controls were compared on particular life domains, the results reflected changes in objective situations as patients reached different stages of depression:

Satisfaction with other aspects of life such as living conditions, security and finances, were not found to be affected by depression. This supports Kuehner’s argument that there is no evidence ‘for a globally biased negative evaluation of individual QoL domains’ by people with depression. She argues that the results of this study ‘should be regarded as indicators of valid self-assessments of symptoms and impaired functioning by depressed patients rather than as indicators of their distorted self-evaluation’ (Reference KuehnerKuehner 2002).

Kuehner concludes that the same issues apply to other commonly used QoL instruments, including the WHOQOL, and recommends that subjective QoL information should be interpreted on individual or subdomain levels rather than on aggregated levels.

A small study by Reference Khatri, Romney and PelletierKhatri et al (2001) using the Lehman et al Quality of Life Interview also looked at the validity of responses by patients, this time with schizophrenia, as compared with patients with cancer. Both groups were asked to choose proxies (a carer or someone who knew them well) to complete the interview as well, so that responses of the two groups on subjective and objective domains could be compared. As is commonly reported, the patients with schizophrenia showed a low correlation between objective and subjective life indexes, tending to be more positive about their lives than would appear objectively justified (unlike those with cancer). The proxies of the schizophrenia patients mirrored this discrepancy. The authors say that this suggests that ‘the reported difference between objective and subjective quality of life is a valid one’, perhaps because patients with schizophrenia (and presumably their families too) adapt to their circumstances and lower their expectations. They comment that these findings suggest that self-reports of patients with schizophrenia can be taken at face value.

Reference Pinikahana, Happell and HopePinikahana et al (2002) confirm this view, arguing that ‘it is becoming clear that people with schizophrenia can validly and reliably report their internal experiences and perceptions’. They also confirm that the mismatch between patient and clinician ratings of QoL is no longer seen as proof that patient perceptions are erroneous. Rather, they are different and complementary constructs, and ‘subjective assessment is now understood to be its own gold standard’.

Accountability of QoL and HRQoL measures to service users

There is little evidence in the literature of attempts to involve mental health service users as equal partners in creating QoL or HRQoL measures, although a number of the scales have been based on domains derived from interviews with patients of mental health services. Reference Carr and HigginsonCarr & Higginson (2001) say that few researchers creating measures have directly asked patients which factors they think constitute quality of life:

Reference Holloway and CarsonHolloway & Carson (2002) recommend re-reading the paper by Reference Lehman, Ward and LinnLehman et al (1982), which presented verbatim comments of people about the issues that affected their quality of life, and suggest this needs updating for the 21st century.

Reference Pinikahana, Happell and HopePinikahana et al (2002) argue that the divergence between the patient’s and the health professional’s perception of QoL ‘requires further exploration and indicates the need to develop an instrument with active input and involvement from consumers of mental health services’.

Inconsistent approaches to QoL

The Department of Health Outcomes Group report reviewed a range of outcome measures in mental health, including QoL measures, warning that ‘there is a lack of a clear consensus of what quality of life constitutes, and how it should be measured’ and that it is unclear how evaluations of quality of life by various different parties (including the patient) should be reconciled (Reference Schmidt, Garratt and FitzpatrickSchmidt 2000: p. 5). A more recent review for the Department of Health (Reference Fitzpatrick, Bowling and GibbonsFitzpatrick 2006: pp. 30–31) found that despite evidence that patients with chronic health problems benefit from increased involvement,

An exploration of the use of patient-reported outcome measures (Reference Greenhalgh, Long and FlynnGreenhalgh 2005), focused on HRQoL instruments, found that although clinicians report giving a high priority to these measures, in practice biomedical, symptom-related factors overrule HRQoL factors in their decision-making about patient care. The authors argue that the influence of HRQoL would improve if more patient-centred instruments were used and there was greater engagement with the clinicians involved in the patient’s care, fostering local ownership of the implementation of HRQoL measures. They advocate more research on how clinicians actually use outcome measures. Reference Gilbody, House and SheldonGilbody et al (2002) found no robust evidence that routine HRQoL measurement can improve quality of care or outcomes in psychiatry, and called for more systematic research to demonstrate the value of outcomes measurement.

QoL and policy

Since government policy is moving towards more effective service user involvement (Health and Social Care Act 2001, Section 11; National Health Service Act 2006, Section 242; Department of Health 2010: p. 13), quality of life measures need to move on from the generic welfarist well-being approach and the medicalised disease-specific approach to involve service users more directly in conceptualising life quality and creating QoL measures. The problems with measuring subjective QoL may be overcome with more sensitive measures and greater reference to emerging ecological concepts such as social capital (Reference Whitley and McKenzieWhitley 2005) and the capabilities approach (Reference HopperHopper 2007). The latter looks at how people have been disabled not only by their mental disorder but also by limitations on their ability to dream, think and act as full members of society that result from psychiatric treatments and social discrimination:

The ‘recovery’ values espoused in mental health policy documents make an implicit promise to restore full citizenship to people with long-term mental health problems. Quality of life measures could incorporate these aspirations rather than take currently restricted capabilities for granted as the best that people can expect. Likewise, the Royal College of Psychiatrists has made a commitment to involve service users in research, training and service delivery (Reference Fitch, Daw and BalmerFitch 2008).

A statement from a service user summarises the complexity of measuring quality of life outcomes, which (as many writers state) are and should remain individual and personal:

MCQs

Select the single best option for each question stem

1. 1 Which of the following types of measurement is discussed in detail in this article?

   1. a quality-adjusted life-years

   2. b quality circles

   3. c patient satisfaction

   4. d health-related quality of life

   5. e happiness measurement.

2. 2 Which of the following is a key domain of the Quality of Life Interview?

   1. a school attendance

   2. b criminal activity

   3. c living situation

   4. d hearing voices

   5. e patient satisfaction.

3. 3 The Lancashire Quality of Life Profile uses:

   1. a the Beck Inventory of Depression

   2. b SF-36

   3. c the Simpson–Angus Scale

   4. d Cantril’s Ladder

   5. e the Ladder of Citizen Participation.

4. 4 The Medical Outcomes Study was carried out in:

   1. a Japan

   2. b the USA

   3. c Sweden

   4. d the UK

   5. e Australia.

5. 5 The Department of Health Outcomes Group report covered outcome measures in:

   1. a dementia

   2. b cancer

   3. c intellectual disability

   4. d mental health

   5. e forensic services.