Clinical Query Intent – Give Doctors What They Want, Not What They Ask For

The goal of our approach is to give the clinical doctor what makes sense from what they type in. We describe the various tools and approaches at our disposal: automated knowledge creation/adjustment, the knowledge graph for knowledge representations, text mining processes for turning strings into concepts, human ranked tooling to leverage an expert (doctor), and additional algorithms to understand just what the doctor is looking for. Starting from the query, we apply these algorithms to determine the core clinical concept(s), the refinement(s), and which concept(s) should be kept, and which concepts may be replaced with better choices, all to produce relevant clinical search results to be evaluated and improved upon.