The problem
A total count of EHR interactions does not describe how activity is distributed across a shift. Short bursts, quiet intervals, and recurring sequences can be lost when the data are aggregated.
Research approach
My work explores a shapelet-informed representation of nursing activity. A 30-minute window comprises six consecutive 5-minute intervals. Moving the window through a shift supports investigation of local patterns in their broader temporal context.
My focus
- Prepare shift-level time series from binned EHR interactions.
- Represent short activity windows and examine recurring patterns.
- Explore clustering and visual summaries for within-shift activity.
What this demonstrates
Applying time-series methods to clinical workflow data while keeping the meaning and limitations of the source data in view.
Research in progress. EHR interactions are a record of system use, not a complete measure of nursing workload or quality of care. This brief does not claim validated clinical outcomes.