Neonatal PACS snafu leads to notification software

Neonatal imaging exams are not the studies you want slipping through the cracks. After a PACS software upgrade caused a spike in report turnaround times for neonatal imaging studies, MedStar Georgetown University Hospital developed notification software to alert radiologists that these exams were ready to be read.

The internally developed notification tool, which paged attending pediatric radiologists, on-call residents, and IT staff, was able to bring report turnaround times back to near-baseline levels.

"Our newborn exam notification system importantly and positively impacted care of a vulnerable population," said Dr. Ross Filice, chief of imaging informatics.

He presented the group's experience during a scientific session at the recent Society for Imaging Informatics in Medicine (SIIM) annual meeting.

PACS-driven workflow

Like many radiology departments, the one at MedStar Georgetown University Hospital uses a PACS-driven workflow in which imaging studies are opened from a PACS-provided worklist. Radiologists can apply worklist filters based on variables such as specialty and modality, as well as pediatric or adult patients.

Things can slip through the cracks, though, when filters are being used. In 2013, for example, pediatric radiologists discovered that exams performed on neonates were not being displayed on pediatric worklists that were filtered by patient age. This omission had led to some embarrassing situations in which priority exams performed on neonates were not read in a timely manner, Filice said.

"As far as we know, there has been no effect on outcomes, fortunately, but at the very best it's very embarrassing, and obviously we felt very lucky [that it did not have a negative effect on outcomes]," he said.

As its PACS vendor was not able to fix the problem quickly, the department set out to develop its own solution.

Using Mirth Connect (Mirth) as its HL7 engine, the institution currently receives real-time HL7 feeds of all information relevant to radiology department workflow on a secure server. The department knows when an exam is completed and can filter exam completion messages to focus on where the completion date is the same as the patient's birth date. The group was able to develop a software tool within a few days that could identify these messages and automatically send pages to pediatric attending radiologists, the on-call resident's pager, and IT staff, according to Filice.

Improved turnaround times

After adopting the tool, the researchers analyzed final turnaround times for neonatal imaging studies performed on the same day of birth from March 2012 to September 2014.

Looking at exams read by an attending pediatric radiologist, initial turnaround times ranged from four to 193 minutes until April 2013, when they spiked to as high as 1,093 minutes.

"The time where it went from very small turnaround times to a lot of very large turnaround times, up to over a day, corresponded exactly with a PACS upgrade that we underwent, so that's when we think that the PACS filter issue was introduced," Filice said.

After the notification software was implemented in May 2014, most turnaround times improved dramatically and returned to near-baseline levels below 200 minutes. The researchers did notice a few outliers, however, that persisted after the PACS notification tool was implemented.

"Those outliers a lot of times were either people who were not pediatric attendings, so [it was] some specialty neuro exam that was done and that a neuroradiologist read or some other exam that was not a stat exam," he said. "So most of those exams can be explained and we are pretty comfortable that the vast majority of those exams that should have been read quickly were being read quickly because of our PACS notification system."

Filice noted that having access to and being familiar with their radiology departmental data was crucial in the process.

"Understanding your data allows you to respond quickly," he said.

Page 1 of 775
Next Page