Pages 531-538, Language: EnglishAnderson, Jim D. / Szalai, John P.Purpose: The objective was to develop a patient-based outcome measure of condition-specific quality of life that would minimize measurement error related to the instrument when used with patients requiring extraoral craniofacial prostheses.
Materials and Methods: An item pool of potential questionnaire items covered 10 clinical/technical and social/psychologic domains. They sought how frequently the issue in the item affected patients and how important the problem in the item was. The 139 items were administered to 94 treated patients in 5 centers in the United States, Canada, and the United Kingdom. Items were eliminated using relevance (frequency × importance), frequency of answer endorsement, Cronbach's alpha (internal consistency), and correlation of items on the same subject. International cultural agreement was tested using analysis of variance and Tukey comparisons within each domain. Scoring was transformed to a scale (0 to 100).
Results: The final instrument contained 52 items yielding a mean quality-of-life score of 72.5% and a standard deviation of 17.9. Very high internal consistency was demonstrated with a final Cronbach's alpha of 0.967. No international cultural disagreement was found in 9 of the 10 domains.
Discussion: The relative weight of each of the domains is (partially) based on the relevance to the patients. Of the 52 items, 29 were identified that do not mention a prosthesis. This subscale has a Cronbach's alpha of 0.976. These items may therefore be useful where within-patient change is of interest.
Conclusion: A patient-based outcome measure of condition-specific quality of life has been developed with control of bias and demonstrated performance characteristics.
