Designing a Human-Centered Early-Warning System for Oncology Inpatients

The Opportunity

It is well known that hospital patients are at high-risk for clinical instability, which is frequently presaged by unrecognized abnormal vital signs or lab results hours before overt decompensation. This project designed a human-early warning sign for clinical deterioration in cancer patients.

Our Approach

Dr. Pat Lyons and his team designed the Onc-EWS–an Electronic Health Record (EHR)-based real-time risk prediction model to alert clinicians to early clinical signs of deterioration, to allow for earlier interventions to prevent or mitigate deterioration. They conducted 86 user surveys, 22 interviews, and 8 user assessments of the data visualization prototype in the 4 sprints conducted.

The Solution

The initial design of an Onc-EWS user experience prototype was built using information (i.e., end-user motivations, values, and views) gathered through an ongoing qualitative research project. Principles of Human-Centered Design were used, as well as conducting iterative rapid-cycle prototype testing and improvement of the Onc-EWS user experience to create a product including a visual interface and active threshold-based alerts.