Table 2 Development and validation of QoL measures

There are four key stages in instrument development:

▶ Identification of measurement model: QoL scales should be based on a stated model or theory of QoL.

▶ Generation of questionnaire content: Content of all QoL scales should be derived from interviews with relevant patients. Both the concerns and the wording used in the items should be generated during these interviews. Thirty to thirty-five interviews are usually sufficient to generate items. Qualitative analysis of the transcripts allows the construction of a QoL outcome model for the disease.

▶ Content refinement and item reduction: Content validity is assessed by comparing the issues covered by the items to the outcome model. Retained items should be clearly expressed, address only one issue, avoid duplication, be potentially capable of change and apply to all respondents. The draft measure should then be tested with a new set of patients to check comprehension, ability to answer the measure and ensure item relevance.

▶ Scaling and psychometric evaluation: Formal testing of dimensionality, reproducibility and construct validity should be achieved by means of a test-retest survey. In most European countries and North America, the survey can be conducted by post. A sample of 100 or more is preferable. It is strongly recommended that this stage should employ Item Response Theory techniques [42, 43].