Abstract
Background
Assessing the appropriateness of antibiotic prescribing in ambulatory care generally relies on the accuracy of diagnosis codes, which is uncertain. It is also uncertain if documented history and physical findings support antibiotic indications (AI). We completed a retrospective study of pediatric primary care (PPC) encounters to determine: A) if documented findings supported documented AI; and B) whether diagnosis codes captured documented AI (figure).
Methods
We conducted point-prevalence audits of the 9 PPC clinics in our healthcare system, randomly selecting one weekday per month to review all visits between 9/2017 and 4/2018. We included only encounters with antibiotic prescribing. We reviewed clinician notes, orders, laboratory results, and ICD-10 diagnosis codes. We recorded demographics; visit date/location; AI as documented in notes; history, examination, and laboratory findings; and diagnosis codes. We used national guidelines to determine whether documentation supported AI. We calculated the sensitivity of diagnosis codes using documented AI as the gold standard.
Results
The sample included 452 encounters. The most common AI were acute otitis media (AOM), pharyngitis, and sinusitis. For AOM, 163 of 168 encounters (97.0%) had an appropriate diagnosis code; for pharyngitis, 127 of 138 (92.0%); and for sinusitis, 68 of 75 (90.7%). For AOM, 160 of 168 encounters (95.2%) had adequate documentation of supportive findings. For sinusitis, 44 of 75 encounters had adequate supporting history and/or examination findings (58.7%). For pharyngitis, while 135 of 139 (97.1%) had a positive streptococcal test, 104 of 139 (74.8%) had history and examination findings to support testing.
Conclusion
By chart review, we identified each AI and evaluated whether findings supported those AI. The sensitivity of diagnosis codes for AI ranged from 90.7–97.0% for common conditions; this result can inform the design of ambulatory stewardship programs. Only 74.8% of children treated for pharyngitis and 58.7% of children treated for sinusitis had sufficient supporting documentation. Use of discrete data elements alone (Figure 1) may result in overestimates of the proportion of children for whom antibiotics are appropriate. Further research is needed across healthcare settings.
Disclosures
All authors: No reported disclosures
IFN-λ4 is associated with increased risk and earlier occurrence of several common infections in African children
