Patient-Reported Outcome Measures (PROMs) are increasingly used in health services research and clinical practice for the quantification of patient experiences, including quality of life, mood (e.g. depression), and satisfaction with services. Such PROMs usually take the form of questionnaires. The underlying measurement model is derived from psychometric theory, specifically Classical Test Theory (CTT). This model requires statistical analysis of questionnaire data to establish the quality of data so collected, with emphasis on the reliability (reproducibility) and validity (domain-specific measurement) of the data.
The widespread adoption of CTT by health service researchers and clinicians is potentially problematic because very few health service researchers receive training in psychometric methods. Researchers and clinicians are therefore unaware of the limitations of CTT and the assumptions of the statistical methods used. This has led to a number of theoretical and empirical problems, illustrated in this thesis, the common theme being that inappropriate methods applied uncritically yield data of questionable value.
|Date of Award||Jun 2009|