A conceptual comparison of well-being measures used in the UK

There is significant political interest in the UK in measuring subjective well-being (SWB) and the possibility of incorporating such measures into policy, including health policy [All-Party Parliamentary Group on Well-being Economics, 2014]. A number of different, yet related, measures of well-being and health are used across government departments. This includes four summary subjective (personal) well-being questions which ask about life satisfaction, happiness yesterday, anxiety yesterday and worthwhileness adopted by the Office of National Statistics (ONS) under its Measuring National Well-being Programme (referred to here as the ONS-4). They have also adopted the use of the short Warwick Edinburgh Mental Well-being Scale (S-WEMWBS) and the General Health Questionnaire (GHQ-12) which is a mental health screening measure that has been used in well-being measurement. In addition to the measures used within the ONS framework, the National Institute for Health and Care Excellence (NICE) currently rely upon the EQ-5D, a measure of health-related quality of life (HRQoL), in the assessment of medical technologies and public health interventions,[NICE 2012; 2013a] while social care guidance includes measures of capability and need, ICECAP-A and Adult Social Care Outcomes Toolkit (ASCOT).[NICE 2013b]

There is limited evidence on how the GHQ-12, ONS-4, S-WEMWBS, EQ-5D, ASCOT and ICECAP-A relate to each other, which causes difficulty in the comparison of results across datasets and evaluations containing different measures and for informing decisions across sectors. Given that these measures are used to inform policy making throughout Government, it is important to better understand how these measures compare. For example, it is unclear whether the GHQ-12 is redundant if the S-WEMWBS is included alongside the ONS-4 within the ONS well-being indicator set.

In order to assess how these measures are related, it is important to establish what construct they seem to measure. Well-being has no agreed upon meaning. There are five main theories that offer a possible explanation of what it is for a life to go well:

1. Objective list theories: Life goes well if an individual has certain goods or characteristics, regardless of whether they want or enjoy them.

4

2. Preference satisfaction theories: Life goes well if an individual gets what she wants. This has been the dominant paradigm within economics.

3. Hedonist theories: Life goes well if an individual experiences pleasure and not pain; therefore the key focus is feelings.

4. Evaluative theories: Life goes well if an individual evaluates her life positively.

5. Flourishing theories: Life goes well if an individual has ideal qualities for a human being.

There are a number of other approaches that focus on what they take to be policy relevant outcomes, including HRQoL, capabilities and the fulfillment of basic needs. Though the theories are distinct, there is considerable overlap in the constructs that they cover; consequently, questions asked across measures assessing the different theories may be very similar.

The instruments

The 6 main measures used in the UK were designed for different purposes and have one or more theory of well-being as the main focus.

1. GHQ-12 contains 6 positive and 6 negative worded items that cover both feelings and psychological functioning. It was designed as a screening tool for mild mental health problems, but has been used (mostly for convenience) as a measure of SWB. The response options differ between the positive and negative items and in both cases are referenced to ‘usual’, making it problematic to use in order to identify stationary health or SWB states. There has been intense disagreement on whether the 12 items in the GHQ act as a single domain.

2. WEMWBS contains 14 positively worded items designed to measure positive mental health and tap into both feelings and psychological flourishing. A shortened 7-item version (S-WEMWBS) was developed using rasch analysis and has been found to have a single domain.

3. ONS-4 contains 4 (experimental) questions, each aiming to represent a different aspect of SWB, including feelings (happiness and anxiety), life evaluation (life satisfaction) and flourishing (worthwhile). However, with only single items it does not provide a complete picture of these concepts.

5

4. ICECAP-A aims to measure capability through prefixing its 5 items with the words ‘able to’ or ‘can’, for example, ‘can have a lot of enjoyment and pleasure’. The dimensions cover aspects of psychological well-being or flourishing along with a more hedonic item (enjoyment). The content was developed from in-depth interviews with adults to define attributes of quality of life and the capabilities that are important to people in their lives.[Al-Janabi et al, 2012]

5. ASCOT is a measure of social-care related quality of life that is designed to assess the extent to which an individual’s social-care needs and wants are being met. It also draws upon Sen’s theory of capabilities, with the level of met needs in each of the 8 domains being disaggregated into having needs met and an ‘ideal’ state in which the level of functioning in the domain is consistent with individual preferences. The instrument can be anchored to the QALY scale.[Netten et al, 2012] Its content overlaps with some aspects of psychological well-being and health, but it is not concerned with the extent that an individual can meet their needs or wishes on their own but with informal and formal care.

6. EQ-5D is a measure of HRQoL rather than well-being or full quality of life. It has two components: five questions describing health and a visual analogue rating scale. The five item descriptive system generates a utility score based on a valuation exercise to extract preferences of a general population sample.

Comparison of measures

When we look at the currently used measurement instruments, most do not fit neatly into any particular theoretical concept of well-being. Measures of health and well-being may draw from different approaches yet have similar items within their measure. Despite this potential over-lap, it is clearly useful to understand the theoretical foundations behind the well-being measurement tools and their conceptual differences. Our six instruments draw from a number of competing conceptual theories and measure different overall constructs: mental health and psychological well-being (GHQ-12, WEMWBS), social-care needs (ASCOT), capabilities (ICECAP-A), evaluative, affect and flourishing (ONS-4), and health-related quality of life (EQ-5D). Despite drawing on specific conceptual theories, there is also considerable overlap in questions asked across the measures. GHQ-12, WEMWBS, ONS-4, ICECAP-A and ASCOT all have questions covering constructs that are important for flourishing; GHQ-12, WEMWBS and EQ-5D

6

also include questions on feelings. One of the distinguishing features between the measures is the inclusion/exclusion of positive and negative items; GHQ-12 has a mix while WEMWBS and ICECAP-A have only positively worded questions, which has implications for potential ceiling and floor effects of the instrument. Where items draw upon more than one theory of well-being, the question of whether this aims to represent a very broad view of a single construct of well-being or whether they are to be treated as separate domains is not clear. Empirical evidence suggests that GHQ-12 may have more than one domain, though the evidence is inconclusive. There is some (limited) evidence that the full WEMWBS may have more than one domain. ICECAP-A, ASCOT and EQ-5D have questions which are intended to measure different domains but they have relative weights for each domain to explicitly combine them. EQ-5D and ASCOT indices also can be combined with survival to provide information that can be used in economic evaluation. There are a number of other differences between the measures including issues to do with the top and bottom anchors of the measures as well as the recall period which may have an impact on what is measured and hence their comparability.

Summary and conclusion

The conceptual assessment of four of the measures (GHQ-12, WEMWBS, ONS-4 and ICECAP) used in the UK indicate that whilst the instruments derive from different backgrounds and a number of other differences (such as recall periods), there may still be considerable empirical overlap between the instruments and possible redundancy when using more than one of these instruments to measure well-being. The EQ-5D and ASCOT have a narrower focus and overlap rather less. Empirical evidence comparing these measures is limited and the next stage will be to compare the measures using existing datasets to examine the degree of over-lap.