Sunday, November 29 | 11:35 a.m.-11:45 a.m. | SSA11-06 | Room S403A
A team from the Hospital of the University of Pennsylvania (HUP) will describe how adding qualitative and sharing features such as hashtags significantly enhanced resident performance analytics at HUP.While 24-hour coverage by attending physicians is becoming increasingly popular in the U.S., HUP continues to use an independent, resident overnight call system with backup by an attending radiologist, according to presenter Dr. Howard Chen.
In the HUP system, trainees create preliminary -- but full -- reports, which become part of the official medical record. Attendings later review the preliminary reports and assign one of four grades: agree, addition, minor change, or major change. Although they are not allowed to modify the preliminary report, the attendings assign the grading and their final interpretation as an addendum, Chen said.
Utilizing the more than 220,000 radiology reports produced by residents since July 2014, the researchers expanded an existing analytic tool to support daily analysis and easier review of call reports. In addition, residents can create searchable tags and comments to document their own cases, as well as share cases with a hashtag feature, Chen said. The program director can use an administrative analytic dashboard to assess emergency department (ED) callback rates.
After measuring usage patterns following implementation of the new system, the researchers found that more than 1,300 tags have been assigned, and there has been a more than threefold increase in average monthly usage by trainees. The trainees have also been very creative with their usage of hashtags, he said.
"The combination of analytics tool and manual tagging enables a mechanism to assess modality-specific quality improvement initiatives," Chen said. "For instance, we have identified ED callback due to [pulmonary embolism] studies and CT angiography as areas of potential improvement and are using the user tags as a way to investigate possible root causes to improve these callback rates."
