More than 23% of all antibiotic prescriptions filled in 2016 were medically unnecessary, and another 36% were questionable, according to an analysis of prescribing data for 19.2 million children and nonelderly adults.
Based on the diagnosis codes for 15.5 million prescriptions filled that year, at least 3.6 million (23.2%) were “inappropriate” – prescribed for conditions for which an antibiotic is almost never recommended, such as acute upper respiratory conditions – and 5.5 million (35.5%) were “potentially inappropriate” – conditions such as acute sinusitis or otitis media, for which an antibiotic is only sometimes recommended, Kao-Ping Chua, MD, PhD, of the University of Michigan, Ann Arbor, and his associates reported in the BMJ.
Only 12.8% of filled prescriptions for the 39 oral antibiotics assessed were classified as “appropriate” under the investigators’ scheme, which assigned an antibiotic appropriateness level to all 91,738 diagnostic codes in the 2016 ICD-10-CM. Finally, 28.5% of antibiotic fills were not associated with a recent diagnosis code, suggesting that they involved phone consultations that did not result in claims or visits that were paid out of pocket and did not make it into the Truven MarketScan Commercial Claims and Encounters database used in the study, the investigators said.
The three highest levels of inappropriate fills were 70.7% in office-based settings, 6.2% in urgent care centers, and 4.7% in emergency departments.
Prescriptions for children, who represented almost a quarter of all antibiotic fills, were less likely to be inappropriate than those for adults aged 18-64 years. Proportions for children were 17.1% inappropriate, 48.7% potentially inappropriate, and 17.0% appropriate, compared with 25.2%, 31.4%, and 11.4%, respectively, for adults, Dr. Chua and his associates said.
“This study shows how data and analytics can help us identify and understand important challenges facing the American health care system,” said Gopal Khanna, director of the Agency for Healthcare Research and Quality, which funded the study. “We now need to use these data to spur change in the prescribing of these very common medications.”
SOURCE: Chua K-P et al. BMJ. 2019;364:k5092. doi: 10.1136/bmj.k5092.