Dotting the i's and crossing the t's: Improving the quality of RIS information

When patient information transcribed into a facility’s radiology information system (RIS) is inaccurate or incomplete, it can have a negative impact on both patient care and workflow. According to a recent study published by the Journal of the American College of Radiology, institutions can significantly improve the quality of information provided to radiologists on image requisitions through education and the implementation of a thorough checklist.

Steven J. Baccei, MD, UMass Memorial Medical Center in Worcester, Mass., and colleagues studied 500 random radiology reports in August 2015, looking at “the degree of concordance with their corresponding clinical order indications” and tracking grammatical errors. Technical staff members and referring physicians were then provided with in-depth education, and a checklist to help the radiology scheduler was put into place.

“The checklist prompted staff members to ask themselves questions necessary to accurately complete an indication in the RIS on the basis of the order received from the clinical provider,” the authors wrote. “These questions guide the radiology scheduler in obtaining appropriate and accurate information from the referring physician’s office when absent.”

In November 2015, after the education was complete and the checklist had been implemented, an additional 500 random radiology reports were studied, allowing the researchers to observe the impact of their quality improvement efforts.

Overall, the number of RIS indications matching the corresponding clinical order verbatim increased from more than 46 percent to more than 62 percent. In addition, the number of “partially concordant” matches due to subpar RIS indications decreased from more than 32 percent to more than 22 percent.

“By educating radiology order entry staff members through one-on-one training and supplying them with a checklist when transcribing information from image requisition to the RIS, we significantly improved the clarity of patient information presented to radiologists,” the authors wrote.

The number of partially concordant matches due to inadequate clinical order indications, however, more than doubled, jumping from more than 4 percent to more than 11 percent.

This number surprised the authors, so they looked closely at each instance and found that technologists had actually obtained additional information to supplement the order form in a majority of the cases.

“By our established definitions, these reports were originally classified as ‘partially concordant (inadequate RIS),’ but because this phenomenon provided an unexpected benefit to our radiologists, we relabeled the cases in which this occurred as ‘concordant plus,’” the authors wrote.

Baccei et al. added that the number of grammatical errors was “relatively unchanged” after the countermeasures. There were 267 incorrect capitalizations and 49 misspellings/misused punctuation marks in the original 500 reports, and 322 incorrect capitalizations and 19 misspellings/misused punctuations marks in the second sample of 500 reports.

“The problem of grammatical errors persisted,” the authors wrote. “Part of the problem was the continued lack of a spell-check function in the RIS during our follow-up data collection phase. Also, the checklist did not specifically include a reminder stressing the importance of proper spelling and grammar during the clinical order indication entry process. Exploring the possibility of a spell-check function in our RIS in addition to adding this key element to the checklist will be the focus of subsequent quality improvement efforts.”