Building Clinical Dashboards
I am a data scientist that works closely with a team of nurses to reduce hospitalizations for patients with kidney failure who are receiving dialysis. My team built and deployed a predictive model to provide the highest risk patients to a team of nurses. This worklist started as a minimum viable product (MVP) excel sheet.
We learned quickly that the excel sheet was difficult to display individual patient data. As a result, we have transitioned to using dashboards that include a worklist and a patient summary. To build this dashboard, we utilized agile techniques to iteratively improve each section of the dashboard. We began with a simple dashboard (essentially a duplicate of the excel sheet) and then worked closely with a few nurses to iteratively improve the dashboard.
Over time we expanded the features contained on the dashboard through user feedback sessions and surveys. The dashboard has expanded to the entire team and nearly every nurse who uses this new system has reported an increase in job satisfaction. In summary, I believe working directly with a set of end users is critical to the development of dashboards and visualizations.
By: Andrew Long