The UNITI project involved a large-scale randomized controlled trial (RCT) with extensive data collection, which was used to develop a Clinical Decision Support System (CDSS) with two cores. The CDSS is designed to recommend tailored examinations and optimal treatment strategies for individual tinnitus patients based on their profiles. The RCT dataset includes a comprehensive set of questionnaires, audiological measurements, genetic data and biomarker information that provided a diverse and robust foundation for the development of the CDSS.
The two cores comprising the CDSS are developed with complementary goals: The CDSS (core 1) builds a predictor for treatment improvement given the data collected in the UNITI and optimizes for predictive performance, while the design of the CDSS (Core 2) is inspired by the findings from the analysis of the historical center-level data, and also prioritizes the ranking of treatments in order of applicability to a patient in addition to optimizing for predictive performance. It was also found that supporting missing values during generation of predictions helps generate predictions for patients for whom data collection is still in progress (i.e., results from an audiological exam might not yet have come in) – which is closer to the use case in a live environment.
More information about the UNITI CDSS can be found here:
