Health status and quality of life outcomes are an important category of PROs. Published papers often use the terms ‘quality of life’ (QOL), ‘health status’, ‘functional status’, ‘health-related quality of life’ (HRQOL) and ‘well-being’ loosely and interchangeably, despite clear definitions of terms (see Table 17.3.a).
Different types of instruments are available for measuring health status and quality of life (see Table 17.3.b). These may yield an overall score or indicator number (representing impact of the intervention on physical or emotional function, for instance), an index number (again an overall score, but weighted in terms of anchors of death and full health), a profile (individual scores of dimensions or domains), or a battery of tests (multiple outcome assessing different concepts): see Table 17.3.b.
HRQOL can be measured using generic or specific instruments, or a combination of both. If investigators were interested in going beyond the specific illness and possibly making comparisons between the impact of treatments on HRQOL across diseases or conditions, they may have chosen generic HRQOL measures that cover all relevant areas of HRQOL (including, for example, mobility, self-care, and physical, emotional, and social function), and are designed for administration to people with any kind of underlying health problems (or no problem at all). These instruments are sometimes called health profiles; the most commonly used health profiles are short forms of the instruments used in the Medical Outcomes Study (Tarlov 1989, Ware 1995). Alternatively (or in addition) randomized trials and other studies may have relied on instruments that are specific to function (e.g. sleep or sexual function), a problem (e.g. pain), or a disease (e.g. heart failure, asthma, or irritable bowel syndrome).
Elicitation of concepts and items for a PRO questionnaire should come from qualitative research with patients, family members, clinical experts, and the literature. For a guide to using qualitative methods, see Chapter 20. Involvement of patients in PRO questionnaire development is essential to ensure content validity. The concepts that are included and measured in an included study can only be determined by examining the actual content of items or questions included in an instrument claiming to measure quality of life or health-related quality of life. The concept is the ‘thing’ being measured. Concepts may relate to an individual item or to a subset of items that refer to the same concept, often referred to as domains. For example, an item measuring pain, a sensation known only to the patient, would be a symptom and the symptom concept that is being measured can be labelled as pain. An item assessing difficulty walking up stairs would be a concept related to physical functioning and might be labelled walking up stairs or as part of physical function. The labelling of concepts varies widely among researchers and there is no agreed-upon classification of concepts. Nonetheless, each item, subdomain, domain, or overall score addresses one or more concepts, which authors can identify from the content, e.g. language, used in the label for an item, domain, or overall score.
Review authors may gain considerable insight from what the authors of the original PRO development studies write about the nature or sources of items chosen for inclusion in a specific instrument. Unfortunately review authors will often find themselves reading between the lines of published clinical trial results to try and get a precise notion of the concepts or constructs under consideration. They may, to gain a full understanding, have to make at least a brief foray into the articles that describe the development and prior use of the PRO instruments included in the primary studies.
For example, authors of a Cochrane review of cognitive behavioural therapy (CBT) for tinnitus included quality of life as an outcome (Martinez-Devesa 2007). Quality of life was assessed in four trials using the Tinnitus Handicap Questionnaire, in one trial the Tinnitus Questionnaire, and in one trial the Tinnitus Reaction Questionnaire. The original sources are cited in the review. Citations to articles on the psychometric properties are also available in MEDLINE for all three instruments and could easily be identified with a search using the Google search engine. Information on the items and the concepts measured are contained in these articles, and review authors were able to compare the content of the instruments.
Another issue to consider in understanding what is being measured is how the PRO instruments are weighted. Many specific instruments weight items equally when producing an overall score. Utility instruments designed primarily for economic analysis put great stress on item weighting, attempting to present HRQOL as a continuum anchored between death and full health. Readers interested in the issues we have laid out in the previous paragraph can look to an old, but still useful summary (Guyatt 1993).