Description
At the point of care, clinicians seek information to understand their patients such as ‘What is this person’s HbA1C goal?’ or ‘Did this person have an adverse reaction to metformin?’. Today, finding answers to these questions often requires foraging through progress notes from past encounters in Electronic Health Records (EHRs). Our goal is to improve the clinician’s understanding of their patient by making it easier to find answers to their prototypical questions. We used a previously developed artificial intelligence natural language processing (AI-NLP) method to extract relevant information from unstructured data in clinician notes. We present a user-centered design approach to support the creation of a user interface for a primary care clinician at the point of care. The design approach was informed by a mixed methods approach used to identify both the nature of physician’s prototypical questions as well as the order in which they are asked. In this presentation, we report how the ethnographic research helped to pinpoint both the structured and unstructured data required by physicians to understand their patients and how the design supported these needs.
Learning Objective: Understand the role of ethnographic research in the design of a software product to be used at the point of care by primary care physicians. Specifically, how this research can help to pinpoint both the structured and unstructured data required by physicians to understand their patients. As well as, how natural language processing can aid in the extraction of disease-relevant insights.
Authors:
Jeff Sokolov (Presenter)
IBM
Morgan Foreman, IBM
Yoonyoung Park, IBM
Sarah Miller, IBM
Amar Das, IBM
Paul Tang, IBM
Presentation Materials: