Abstract
OBJECTIVES: To construct a prediction model of preference-adjusted health status (SF-6D) for Cushing's syndrome using a disease-specific health-related quality of life (HRQOL) measure (CushingQoL). METHODS: Data were obtained from the original multicenter, multinational study to validate the CushingQoL questionnaire. HRQOL was measured using the CushingQoL and the SF-36 questionnaires. SF-6D scores were calculated from responses on the SF-36. Sociodemographic and clinical data were also collected. Various predictive models were tested and the final one was selected on the basis of four criteria: explanatory power, consistency of estimated coefficients, normality of prediction errors, and parsimony. RESULTS: For the mapping analysis, data were available from 116 of the 125 patients included in the original validation study. Mean (SD) age was 45.3 (13.1) years and the sample was predominantly (83 %) female. Patients had a mean (SD) CushingQoL score of 52.9 (21.9), whereas the SF-6D (derived from SF-36) was skewed towards better health with a mean of 0.71 (median 0.74) on a scale of -0.704 to 1. Of the various models tested, a model which included the intercept (0.61), CushingQoL overall score, level one in CushingQoL item 2 (always have pain preventing me from leading a normal life), and level one in CushingQoL item 10 (my illness always affects my everyday activities) best met the four criteria for model selection. The model had an adjusted R (2) of 0.60 and a root mean square error of 0.084. CONCLUSIONS: Although the mapping function finally selected appears to be able to accurately map CushingQoL scores onto SF-6D outcomes at the group level, further testing is required to validate the model in independent patient samples.