Developing a common metric for commonly used instruments in Routine Outcome  Monitoring

Using a wide variety of clinical questionnaires might hamper the communication between clinicians and researchers in Routine outcome monitoring (ROM). Linking test scores to T scores as a common metric might be a solution. Item response theory (IRT) is the preferred way to estimate T-scores. However, required software and a large dataset is needed for that. In this thesis, two methods were used to arrive at a common metric for often used questionnaires and their subversions: the BSI, BSI-18, the DASS-42 and the DASS-21. As one method, we used an IRT approach to estimate theta-based T-scores, leading to crosswalk tables. For the second approach, conversion formulas were estimated for raw scores to calculated T-scores based on the best-fitting regression equation between the raw scores and the previously estimated theta-based T-scores. Crosswalk tables and conversion formulas are both alternative ways to link individual scores to a common metric. It was investigated if conversion...Show moreUsing a wide variety of clinical questionnaires might hamper the communication between clinicians and researchers in Routine outcome monitoring (ROM). Linking test scores to T scores as a common metric might be a solution. Item response theory (IRT) is the preferred way to estimate T-scores. However, required software and a large dataset is needed for that. In this thesis, two methods were used to arrive at a common metric for often used questionnaires and their subversions: the BSI, BSI-18, the DASS-42 and the DASS-21. As one method, we used an IRT approach to estimate theta-based T-scores, leading to crosswalk tables. For the second approach, conversion formulas were estimated for raw scores to calculated T-scores based on the best-fitting regression equation between the raw scores and the previously estimated theta-based T-scores. Crosswalk tables and conversion formulas are both alternative ways to link individual scores to a common metric. It was investigated if conversion formulas are valid by comparing calculated T-scores with theta-based T-scores. The agreement between calculated T-scores and theta-based T-scores were high for all questionnaires except the DASS-42. The agreement was also lower at the extreme ends of the questionnaires (T= < 50 and T= >75). All in all, conversion formulas seem to be a good alternative for estimating a common metric. Offering different options to calculate common metrics can help in improving the communication between professionals in the clinical field. Increased involvement and better communication in ROM is able to increase the quality of mental health care.Show less

Leiden University Student Repository

Documents

In Collections

Developing a common metric for commonly used instruments in Routine Outcome Monitoring