Abstract
BackgroundRecently, a growing number of Item-Response Theory (IRT) models has been published, which allow estimation of a common latent variable from data derived by different Patient Reported Outcomes (PROs). When using data from different PROs, direct estimation of the latent variable has some advantages over the use of sum score conversion tables. It requires substantial proficiency in the field of psychometrics to fit such models using contemporary IRT software. We developed a web application (http://www.common-metrics.org), which allows estimation of latent variable scores more easily using IRT models calibrating different measures on instrument independent scales.ResultsCurrently, the application allows estimation using six different IRT models for Depression, Anxiety, and Physical Function. Based on published item parameters, users of the application can directly estimate latent trait estimates using expected a posteriori (EAP) for sum scores as well as for specific response patterns, Bayes modal (MAP), Weighted likelihood estimation (WLE) and Maximum likelihood (ML) methods and under three different prior distributions. The obtained estimates can be downloaded and analyzed using standard statistical software.ConclusionsThis application enhances the usability of IRT modeling for researchers by allowing comparison of the latent trait estimates over different PROs, such as the Patient Health Questionnaire Depression (PHQ-9) and Anxiety (GAD-7) scales, the Center of Epidemiologic Studies Depression Scale (CES-D), the Beck Depression Inventory (BDI), PROMIS Anxiety and Depression Short Forms and others. Advantages of this approach include comparability of data derived with different measures and tolerance against missing values. The validity of the underlying models needs to be investigated in the future.
Highlights
A growing number of Item-Response Theory (IRT) models has been published, which allow estimation of a common latent variable from data derived by different Patient Reported Outcomes (PROs)
Our goal is to enable researchers to compare data obtained with different measures, for example if in Study A the Patient Health Questionnaire 9 (PHQ-9) has been used for the measurement of depression, but in Study B the Beck Depression Inventory (BDI) was the measure of choice
Compared to traditional IRT software the major strength of our approach by providing a web application is that theta estimation from different PROs does not require detailed knowledge on IRT modeling nor estimation techniques
Summary
A growing number of Item-Response Theory (IRT) models has been published, which allow estimation of a common latent variable from data derived by different Patient Reported Outcomes (PROs). Different methods yielding comparable results have been applied to link measures, such as fixed-parameter estimation or concurrent estimation with subsequent linking [12, 13, 18] Those IRT models have been frequently used to develop sum score conversion tables between measures [7, 8, 10, 12, 15] since it is possible to derive latent trait estimates solely from the sum score [19]. It is possible to estimate the latent trait directly from the response pattern This approach has some advantages over the use of sum score conversion tables since it takes into account differences in the response pattern, yielding more accurate results [12, 13] than converted sum scores. It is favorable in case of missing item response, since estimation of the latent variable is still viable under that condition [12, 13]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
