We have developed our own built-in, multi-feature next generation Computerized Adaptive Testing (CAT) engine. We’ve leveraged state-of-the-art CAT techniques from the field of education and expanded the methods to our application. For example, item-response theory (IRT) enables dynamic assessments where the questions asked are based on the subject’s performance in the test up to that point.
To create CHOIR CAT, we first collaborated with the NIH and Northwestern University, integrating the IRT features of the NIH’s Patient Reported Outcomes Measurement Information System (PROMIS) into CHOIR CAT. These are the features that allow PROMIS to efficiently assess various psychometric domains, such as Depression, Anxiety, Anger, Pain Interference, Pain Behavior, Fatigue, and Physical Function. We then built upon this foundation and added other state-of-the-art CAT techniques, such as multiple objectives and constrained optimization.
Novel dynamic visualization features help clinicians and researchers better understand the inner workings of IRT and CAT.
We find substantial savings in patient burden with the use of modern patient-reported outcomes (PRO). As an example, the below chart compares the numbers of items asked via a modern patient question system (NIH’s PROMIS) with the numbers of items used on instruments based on classic (legacy) testing theory.
Why not just use EMR?
A frequently asked question is: Why produce a new system? Why not just use [insert favorite EMR]? Computational Complexity of Modern Patient Reported Outcomes (PRO) It’s true that we could have just translated paper questionnaires into simple web forms. However, a database like CHOIR enables data to be collected much more efficiently, and for much more sophisticated outcomes to be reported. The computational complexity and demands of modern patient reported outcomes (PROs) are beyond what can be provided by traditional EMR. They are not simple, static web forms and require substantial computational resources.