Off-console automated artificial intelligence enhanced workflow enables improved emergency department CT capacity

Increasing CT capacity to keep pace with rising ED demand is critical. The conventional process has inherent drawbacks. We evaluated an off-console automated AI enhanced workflow which moves all final series creation off-console. We hypothesized the off-console workflow would 1) decrease overall ED CT exam begin to end times and decrease length and variability of time CT is occupied at the individual exam level.

Study population was identified retrospectively and included all CT exams done on all ED adult patients. 3 months of data was collected using the conventional workflow and 3 months of data was collected after implementation of the off-console workflow. Exam begin and the exam end timestamps were collected from the EMR. Additionally, 4 subgroups from the above conventional and off-console workflows were identified retrospectively with an Emergency Severity Index level 1, undergoing one of the four most common CT exam set(s) performed on ESI level 1 patients.

6,795 ED adult patients underwent ED CT in the 3 months immediately prior to implementation of the off-console workflow and 6,708 adult ED patients underwent CT in the 3 months after complete implementation. The off-console workflow demonstrated a 36% decrease in median exam begin to end times (P < 0.001). 4 subgroups demonstrated 56-75% decreases in median CT occupied time (P < 0.001) and decreases in variability in ¾ subgroups.

This off-console workflow enables increased CT capacity to meet rising ED demand. Similar improvements could be expected across most exam sets and imaging settings if broadly implemented.