Developing a descriptive system for a new preference based measure of health related quality of life for children

Aims The use of preference based measures (PBM) of health related quality of life is increasing in health care resource allocation decisions. Whilst there are measures widely used for this purpose in adults, research in the paediatric field is more limited. This paper reports on how the descriptive system for a new paediatric generic PBM of HRQoL was developed, including how the wording and order of the levels for each dimension was determined. The aim was to develop a descriptive system suitable for health state valuation, based on dimensions identified from previous work.

Methods The main constraint in designing a descriptive system for a PBM is that the health states defined by the system should be amenable to valuation. Ideally, each dimension needs to contain levels that are ordered within it to fit this criteria well. To develop these levels, the first stage is to determine whether they should be frequency or severity based. To do this, data from the original qualitative work for developing the dimensions was used, reviewing how children described the dimensions. Once this was determined, existing scales from the paediatric literature were reviewed for their suitability and scales were also developed empirically, based on the qualitative interview data from children. The empirical scales were developed by taking the adverbial phrases from the interview work and to confirm the ordinality of these, a ranking exercise was undertaken with children. The resulting scales were applied to the dimensions and a draft descriptive system developed.

Results For every dimension, severity arose as the predominant characteristic. No suitable scales were found in the paediatric literature for severity and so empirical scales based on the qualitative data were developed and used which resulted in seven different types of scale. Children were successfully able to rank these scales to determine the ordinality and the scales were applied to all the dimensions in order to form a draft descriptive system. The ordering of the statements resulting from the analysis made sense at face value.

Conclusions This work has empirically developed a descriptive system for the dimensions of HRQoL identified in the original interview work. As the methods were based on using data from children, the content validity of the final measure should be increased and the language and terminology is appropriate. Further research is needed to test the descriptive system on a paediatric population and to test the psychometric performance. In addition, due to the constraints of preference based measures, the number of dimensions will need to be reduced to be amenable to valuation. Further research is required to do this.