SEATTLE – Blood cultures were ordered for over half of pediatric skin infection encounters across 38 children’s hospitals, with rates varying from about 20% to 80% between hospitals, according to a review of almost 50,000 encounters in the Pediatric Health Information System database.
It was a surprising finding, because current guidelines from the Infectious Diseases Society of America do not recommend blood cultures as part of the routine evaluation of uncomplicated pediatric skin and soft-tissue infections (SSTIs), meaning infections in children who are otherwise healthy without neutropenia or other complicating factors.
Just 0.6% of the cultures were positive in the review, and it’s likely some of those were caused by contamination. After adjustment for demographics, complex chronic conditions, and severity of illness, culture draws were associated with a 20% increase in hospital length of stay (LOS), hospital costs, and 30-day readmission rates.
“Our data provide more evidence that [routine] blood cultures for children with SSTI represents low-value practice and should be avoided,” said lead investigator John Stephens, MD, a pediatrics professor and hospitalist at the University of North Carolina at Chapel Hill.
Dr. Stephens became curious about how common the practice was across hospitals after he and a friend penned an article about the issue for the Journal of Hospital Medicine’s “Things We Do for No Reason” series. The single-center studies they reviewed showed similarly high rates of both testing and negative cultures (J Hosp Med. 2018 Jul;13:496-9).
Dr. Stephens and his team queried the Pediatric Health Information System database for encounters in children aged 2 months to 18 years with the diagnostic code 383, “cellulitis and other skin infections,” from 2012 to 2017, during which time “there really wasn’t a change” in IDSA guidance, he noted. Transfers, encounters with ICU care, and immunocompromised children were excluded.
Hospital admissions were included in the review if they had an additional code for erysipelas, cellulitis, impetigo, or other localized skin infection. The rate of positive cultures was inferred from subsequent codes for bacteremia or septicemia.
Across 49,291 encounters, the median rate of blood culture for skin infection was 51.6%, with tremendous variation between hospitals. With blood cultures, the hospital LOS was about 1.9 days, the hospital cost was $4,030, and the 30-day readmission rate was 1.3%. Without cultures, LOS was 1.6 days, the cost was $3,291, and the readmission rate was 1%.
Although infrequent, it’s likely that positive cultures triggered additional work-up, time in the hospital, and other measures, which might help account for the increase in LOS and costs.
As for why blood testing was so common, especially in some hospitals, “I think it’s just institutional culture. No amount of clinical variation in patient population could explain” a 20%-80% “variation across hospitals. It’s really just ingrained habits,” Dr. Stephens said at Pediatric Hospital Medicine.
“The rate of positive blood culture was really low, and the association was for higher cost and utilization. I think this really reinforces the IDSA guidelines. We need to focus on quality improvement efforts to do this better,” he said, noting that he hopes to do so at his own institution.
“I’d also like to know more on the positives. In the single center studies, we know more than half of them are contaminants. Often, there’s more contamination than true positives,” he said at the meeting sponsored by the Society of Hospital Medicine, the American Academy of Pediatrics, and the Academic Pediatric Association.
Instead of routine blood culture, Dr. Stephens recommended in his article to send pus for a Gram stain and culture and sensitivity, while noting that blood cultures remain reasonable for complicated infections, immunocompromised patients, and neonates.
There was no external funding, and Dr. Stephens didn’t report any disclosures.