David K. Warren
An intervention that included changing the electronic ordering system to encourage urinalysis reflex testing for suspected urinary tract infections led to a 45% reduction in urine cultures ordered at a St. Louis hospital without missing any UTIs, according to research published in Infection Control & Hospital Epidemiology.
“Over-testing for UTIs drives up health care costs and leads to unnecessary antibiotic use, which spreads antibiotic resistance,” David K. Warren, MD, MPH, a professor of medicine in the division of infectious diseases at Washington University School of Medicine and a hospital epidemiologist at Barnes-Jewish Hospital, both in St. Louis, said in a news release.
“Ordering tests when the patient needs them is a good thing. But ordering tests when it’s not indicated wastes resources and can subject patients to unnecessary treatment.”
Warren and colleagues made two changes to how urine testing is conducted at Barnes-Jewish Hospital. As explained in the news release, they first sent an email to clinicians to explain the rationale behind ordering urine dipstick tests for signs of a UTI before ordering a bacterial culture. They also changed the electronic ordering system to make urine dipstick — and not culture — the default test, leaving clinicians with the option of still ordering a culture test by itself if they wanted, albeit after opening an additional screen on their computers.
For their study, Warren and colleagues included adults who had at least one urine culture performed during their stay. They compared urine culture rates from January 2015 to April 2016 — before the 1-month intervention conducted in April 2016 — to rates from May 2016 to August 2017, after the intervention.
Warren and colleagues reported that 18,954 patients (median age, 62 years; 52.3% women) had 24,569 urine cultures ordered; 27% of the cultures were positive. According to the findings, there was a significant decrease in the urine culturing rate during the postintervention period for any specimen type compared with the preintervention period (38.1 per 1,000 patient days vs. 20.9 per 1,000 patient days), clean catch (30.0 vs. 18.7) and catheterized urine (7.8 vs. 1.9; P <.001, all). They also reported that urine culture rates decreased for all specimen types when an interrupted time series model was used.
“By encouraging clinicians to use urinalysis reflex testing, rather than urine cultures for hospitalized patients, we significantly reduced the number of urine cultures performed among these patients,” Warren told Infectious Disease News. “At the same time, we saw no change in catheter-associated UTI rates, suggesting that we were not missing UTIs.”
Warren added that the study showed the importance of how computerized physician order entry systems are designed.
“Our study shows that relatively minor changes in these systems can have a major impact on testing practices, which can limit unnecessary testing and potentially unnecessary antibiotic use,” he said. – by Bruce Thiel
Disclosures: The authors report no relevant financial disclosures.