Developing methodology for deriving preference-based measures from condition specific quality of life instruments: Overactive Bladder (OAB-5D)
Methods were developed for deriving a multi-dimensional utility instrument from a condition specific quality of life instrument using a combination of Rasch analysis and psychometric methods. The methodology was developed using the overactive bladder questionnaire (OABq) to derive the preference-based OAB-5D.
The methodology used seven steps:
- Instrument dimensionality was explored using Factor analysis and Cluster analysis.
- Data were fitted to Rasch models as a tool for eliminating items from consideration in the preference based measure (item level ordering, differential item functioning and Rasch model fit criteria were used to eliminate items).
- Items were selected from each dimension identified in step 1 using a combination of item level fit, and spread of item scores from Rasch analysis and feasibility, internal consistency, distribution of item responses and responsiveness from psychometric analysis.
- A combination of Rasch and psychometric methods were used to explore the possibility of collapsing item levels to a manageable level for a valuation exercise.
- Steps 1 to 5 were repeated on further datasets to validate the results.
- A series of health states were selected and valued using time trade off (TTO) methods in a general population sample.
- Valuation results were modelled to produce utility values for all heath states.
The OAB-5D has 5 dimensions: Urge to urinate; urine loss; sleep; coping; concern. Each dimension has 5 severity levels, and the classification defines a total of 3125 health states. The OAB-5D has a range of utility values from 0.606 to 1.
Relevant Publications
- Yang Y, Brazier J, Tsuchiya A, Coyne K. Estimating a preference-based single index from the overactive bladder questionnaire. Value in Health 2009;12(1):159-66.
- Young T, Yang Y, Brazier JE, Tsuchiya A, Coyne K. The first stage of developing preference-based measures: Constructing a health-state classification using Rasch analysis. Qual Life Res 2009;18(2):253-65.
Further Information
If you would like further information about this work please contact t.a.young@sheffield.ac.uk