Cumulative Probability of Receiving an Antibiotic Prescription over Time

Abstract

To the Editor: The therapeutic benefits of antibiotics must be balanced against their risks, which include drug-related toxic effects, Clostridium diffi-cile infection, and selection for antibiotic resistance. 1 Understanding the cumulative probability of receiving an antibiotic prescription over time is necessary to inform clinical practice, antibiotic stewardship, and public policy. However, estimates of population-level prescribing from cross-sectional , deidentified, aggregate pharmacy records cannot evaluate these outcomes. 2 We previously used an individual-level database of enrollees in health insurance plans in the United States to show that approximately one third of enrollees filled an antibiotic prescription at an outpatient pharmacy in a given year. 3 Here, we take a similar approach to quantify the cumulative probability of filling an antibiotic prescription over time. We performed a Kaplan-Meier survival analysis using the Truven Health MarketScan Research Databases 4 for 2011 through 2014, which cover 62 million enrollees in health insurance plans and 100 million claims for antibiotic prescriptions at outpatient pharmacies. We performed separate analyses stratified according to antibiotic class, sex, age group, and U.S. Census region. (Details are provided in the Methods section in the Supplementary Appendix, available with the full text of this letter at NEJM.org.) The probability of filling an antibiotic prescription at an outpatient pharmacy was 33% (95% confidence interval [CI], 30 to 36) over 1 year, 47% (95% CI, 44 to 49) over 2 years, 55% (95% CI, 53 to 58) over 3 years, and 62% (95% CI, 60 to 64) over 4 years (Fig. 1). Similar patterns were observed when the data were stratified according to drug group, sex, Census region, and age group. In nearly all cases, demographic groups with a higher annual volume of consumption 2 had a higher cumulative probability of antibiotic use at all time points (Fig. 1 in the Supplementary Appendix). For example, women had a higher probability of antibiotic use than men, and enrollees in southern states had a higher probability than those in other regions. These results show that the probability that an enrollee filled an antibiotic prescription is high, with approximately half the study population filling a prescription over 2 years. There was also heterogeneity in the frequency of antibiotic use: roughly one third of enrollees filled an antibiotic prescription during a given year, whereas another third did not fill a prescription over a 4-year span. Our study is limited by the sample, which includes approximately 20% of the U.S. population, by the 4-year time frame, and by the fact that we could not address the appropriateness of prescribing. Because antibiotic use is widespread, with this week's letters 1872 Cumulative Probability of Receiving an Antibiotic Prescription over Time 1873 Imputability of Adverse Events to Anticancer Drugs 1875 Tafenoquine versus Primaquine to Prevent Relapse of Plasmodium vivax Malaria 1876 Outcomes after PCI Strategies in Cardiogenic Shock 1878 Marine n-3 Fatty Acids and Vitamin D Supplementation and Primary Prevention 1880 Dapagliflozin and Cardiovascular Outcomes in Type 2 Diabetes e36 Lactobacillus for Gastroenteritis in Children The New England Journal of Medicine Downloaded from nejm.org on November 22, 2020. For personal use only. No other uses without permission.