Analysis of MR log files drives process improvement

Wednesday, December 2 | 3:10 p.m.-3:20 p.m. | SSM12-02 | Room S102D
In this talk, a University of Washington team will describe how mining log files stored on MR scanners paved the way for efficiency gains.

Radiology departments today face continuing pressure to improve the quality of their services while at the same time maintaining cost levels and delivering better value, according to presenter Dr. Martin Gunn. Making informed decisions to achieve quality and efficiency improvements requires good data, however.

The institution collaborated with Philips Healthcare to unearth a gold mine of data stored on log files, which are lists of events that are saved on most advanced imaging equipment.

"It is a bit like a flight data recorder on an aircraft," Gunn said.

When accessed and decoded using software tools, the data in these log files contain much more detail about imaging examinations than is available in the RIS or PACS, he said.

The researchers made use of the modality log file data for an MRI process improvement project. After identifying the most commonly used protocols on the radiology department's MR scanners, the team was able to calculate the average scan duration and variability for each body imaging protocol. These results were then used to refine the protocols.

After these changes were implemented, the department saw a decrease in both average scan duration and variability.

"Our data has also shown us that MRIs take longer on Fridays than on Mondays, leading us to do a root-cause analysis to determine why," Gunn said. "We can also monitor the utilization of our ultrasound equipment, and we have found that there are times during the day when not much scanning is occurring, which has helped us to relook at staff scheduling."

The researchers are also currently investigating the reasons behind aborted and repeated MRI sequences.

"Overall, we see huge applications for this data source for operational improvement in our department," Gunn said.

Learn more by attending this talk on Wednesday afternoon.

Page 1 of 775
Next Page