spectrum drugs also kill more of the good bacteria that live inside us, and can cause more side effects, said Jeffrey Gerber, an infectious-disease doctor at Children's Hospital of Philadelphia and the study leader.
He used the hospital's network of primary-care offices and electronic health records to show that a combination of education and individualized feedback can improve prescribing behavior. His team's report on its "antibiotic stewardship" program was published in last week's Journal of the American Medical Association.
"We found a relatively simple way to leverage an electronic health record to improve guide-line adherence," Gerber said.
The group focused on primary-care offices because 75 to 80 percent of antibiotic prescriptions for kids are written in outpatient settings.
In nine of the practices, an infectious-disease specialist taught pediatricians the latest prescribing practices. The study team then tracked prescriptions they wrote for pneumonia, sinusitis, and strep throat. Guidelines call for treating all of those infections with penicillin or amoxicillin.
Individual pediatricians received quarterly reports for a year on how well their medication choices matched guidelines. Results from those offices were compared with prescribing patterns at nine practices that had not received the intervention.
Overall, prescribing of broad-spectrum antibiotics for the three infections fell from 26.8 percent to 14.3 in the intervention group while it dropped by six percentage points in the control group. Broad-spectrum prescribing for pneumonia fell by 75 percent in the intervention group but changed little in the control group. In both groups, doctors rarely used antibiotics for viral infections.
Ear infections, another common reason for prescribing antibiotics, were not included because they were part of another study. Patients with allergies to antibiotics, 8 to 10 percent of the population, and patients who had recently taken antibiotics also were excluded.
Contact Stacey Burling at 215-854-4944 or email@example.com.