While hospitals are trying to rein in antibiotic use, outpatient settings are on the frontier of sorts in the effort to stop the rise of multidrug-resistant bacteria and their possible consequence: untreatable infections.
In an analysis using the new ICD-10-CM (international classification of diseases-clinical modification) codes, investigators showed misuse of antibiotics in ambulatory settings while unveiling a study model that could aid future research.
“Among all outpatient antibiotic prescription fills by 19,203,264 privately insured U.S. children and non-elderly adults in 2016, 23.2% were inappropriate, 35.5% were potentially appropriate, and 28.5% were not associated with a recent diagnosis code,” they concluded.
Lead author Kao-Ping Chua, MD, assistant professor of pediatrics at the University of Michigan, explained the important findings of the study in the following interview with Hospital Infection Control & Prevention.
HIC: Can you elaborate on the methodological framework you developed to assess outpatient antibiotic appropriateness using ICD-10-CM diagnosis codes? You conclude it may be an approach that could be adopted by others.
Chua: What prompted this study1 was the lack of methods to assess antibiotic appropriateness using ICD-10-CM diagnosis. Prior methods were all based on ICD-9-CM, which was replaced by ICD-10-CM in the U.S. in October 2015. To address this gap, we individually classified each and every one of the 91,000-plus ICD-10-CM diagnosis codes as being “never,” “sometimes,” or “always” indications for antibiotics. For each antibiotic prescription fill, we then looked back three days to see what types of ICD-10-CM diagnosis codes were present on claims during that period. If, for example, there were only codes that never justified antibiotics during this time period (such as acute upper respiratory infection) we classified the antibiotic as “inappropriate.”
This general framework could be used with any administrative data set that uses ICD-10-CM diagnosis codes, including Medicare and Medicaid claims, electronic medical records, and national surveys. The vast majority of developed countries have been using some form of ICD-10 for decades, so in theory, our methods could also be adapted to other countries, assuming they have data on antibiotic prescription fills and the associated diagnosis codes.
HIC: Could you comment on the finding that only 12.8% of antibiotic prescriptions were appropriate?
Chua: It’s important to look at the distribution of prescriptions across the four categories. This distribution indicates the degree to which we can be confident that an antibiotic prescription was appropriate. For 12.8% of prescriptions (“appropriate”), we can be reasonably confident. For 35.5% (“potentially appropriate”), we are not really that confident. For 23.2% (“inappropriate”), we are not confident. And for 28.5% (not associated with a recent diagnosis), we really don’t have any idea.
We are not claiming that only 12.8% of antibiotics were necessary, since some of the antibiotic prescribing in the “potentially appropriate” and “not associated with a recent diagnosis” cases could, in fact, have been necessary. However, we are claiming that we can be reasonably confident that antibiotics were appropriate just 12.8% of the time.
HIC: Is the level of inappropriate prescriptions in line with prior research and your general expectations?
Chua: The most widely cited estimate of antibiotic appropriateness came from a study2 in the Journal of the American Medical Association [JAMA] in 2016. That study estimated that 30% of outpatient antibiotic prescriptions were inappropriate.
It’s difficult to make an apples-to-apples comparison with our estimate of 23.2% for a few reasons. First, the JAMA study analyzed data from emergency departments and office-based visits in 2010 and 2011. In contrast, we used data from privately insured patients in 2016 that captured prescribing across a broader variety of settings.
In addition, our figure of 23.2% only refers to antibiotics written for conditions that don’t justify their use. The JAMA authors’ figure of 30% reflects both these antibiotics as well as their best estimates of the rate of inappropriate prescribing among antibiotics that are written for conditions that only sometimes justify antibiotics, like sinusitis.
To get at the latter, they used estimates of bacterial prevalence and also examined regional variation in prescribing for particular conditions, assuming that the lowest-prescribing U.S. Census region reflected the rate of appropriate prescribing.
HIC: Nearly two-thirds of outpatient antibiotic prescriptions are only potentially appropriate or not associated with a recent diagnosis code. Can you comment on what may be happening there?
Chua: Overall, 35.5% of antibiotics were classified as potentially appropriate because they were associated with diagnoses that only sometimes justify antibiotics, like sinusitis, sore throats, and ear infections. We know that there is a lot more prescribing for these conditions than you would expect based on the percentage of times these conditions have a bacterial cause. Therefore, there is likely to be a lot of unnecessary prescribing in this category.
The 28.5% of antibiotic prescriptions that were not associated with a recent diagnosis code could have several explanations. The first is that our medical claims database captured all antibiotic prescriptions but not necessarily all visits, including dental visits (which would not be reimbursed through medical insurance) and retail clinic or urgent care center visits that are paid out of pocket. However, we know that inappropriate prescribing occurs in these settings as well, particularly retail clinic and urgent care centers.
Another possibility is that these prescriptions could be refills of antibiotics for people who take them for long periods of time, such as patients with acne. We found that about one-sixth of the 28.5% were, in fact, refills.
A final possibility is that antibiotic prescriptions occurred after a telephone or online “visit” that was not billed to insurance — that is, a prescription that occurred in the absence of examining a patient in person. While such prescriptions could be reasonable in certain circumstances, it’s generally important to be able to perform a physical exam to determine whether a patient is likely to have a bacterial illness requiring antibiotics.
HIC: Many hospitals have adopted antibiotic stewardship programs. Are some of the outpatient settings adopting programs to control antibiotic use? Are you aware of any approaches that are proving effective?
Chua: There have been decades of efforts to educate the public and physicians about antibiotic overprescribing, as well as many efforts to improve outpatient antibiotic stewardship. Some of the more effective approaches have come in research performed by the senior author of this paper, Jeff Linder, MD.
In one study,3 he and his colleague doctors at primary care practices received an e-mail that indicated whether they were a “top performer” in terms of avoiding inappropriate prescribing among their peers. That simple intervention reduced inappropriate prescribing drastically, potentially because many doctors are naturally competitive and recoil at the idea of being worse clinicians than their peers.
- Chua KP, Fischer MA, Linder JA. Appropriateness of outpatient antibiotic prescribing among privately insured US patients: ICD-10-CM based cross sectional study. BMJ 2019;364:k5092 doi: https://doi.org/10.1136/bmj.k5092.
- Fleming-Dutra KE, Hersh AL, Shapiro DJ, et al. Prevalence of Inappropriate Antibiotic Prescriptions Among US Ambulatory Care Visits, 2010-2011. JAMA 2016;315(17):1864-1873.
- Meeker D, Linder JA, Fox CR, et al. Effect of Behavioral Interventions on Inappropriate Antibiotic Prescribing Among Primary Care Practices: A Randomized Clinical Trial. JAMA 2016;315(6):562-570. doi:10.1001/jama.2016.0275.
