Antibiotics are valuable drugs. Being able to quickly combat infections if very important. However, overuse of antibiotics is problematic as the infecting organism may mutate more rapidly, leading to antibiotic resistant strains of the disease. Thus, when there is evidence of an infection, we should prescribe antibiotics; when there is no evidence, we should not.
A paper in Health Affairs by Fischer et al. (2020) aims to examine whether there is evidence-based antibiotic prescribing. They use claims data (the Medicaid Analytic eXtract or MAX) between 2004-2013 covering 53 million Medicaid patients. They find that:
…among 298 million antibiotic fills (62 percent for children) for 53 million patients, 55 percent were for clinician visits with an infection-related diagnosis, 17 percent were for clinician visits without an infection-related diagnosis, and 28 percent were not associated with a visit. Non-visit-based antibiotic prescriptions were less common for children than for adults and more common in the West than in other US regions. Large fractions of antibiotic prescriptions are filled without evidence of infection-related diagnoses or accompanying clinician visits.
A key limitation of the study is that the authors assume that failing to document an infection implies there was no infection. It is possible, however, that the physicians documented the infection in the EMR but not on the submitted claim or there was an infection, but it was not documented at all. Despite this limitation, identifying areas where antibiotic over-use could be curbed is important.
- Michael A. Fischer, Mufaddal Mahesri, Joyce Lii, and Jeffrey A. Linder. Non-Infection-Related And Non-Visit-Based Antibiotic Prescribing Is Common Among Medicaid Patients. Health Affairs 2020 39:2, 280-288