Résumé
Large health system databases have revolutionized our understanding of the epidemiology of adverse drug reactions and immunologically mediated drug hypersensitivity. Population-based background rates of newly reported drug intolerance with each therapeutic exposure could not have been determined without large health system databases. Large databases have increased our understanding of multiple drug intolerance syndrome. Large health care systems, such as Kaiser Permanente, with a single electronic medical record system that covers all inpatient, outpatient, and pharmacy interactions, are particularly valuable in understanding the population-based incidence of severe and nonsevere adverse drug reactions, the risks of delayed-onset adverse drug reactions, such as those caused by Clostridiodes difficile, which can occur months after antibiotic exposures, and the risks and benefits associated with “allergy” delabeling, specifically penicillin allergy delabeling, which may accrue in the years after the delabeling. There currently are limitations to using electronic data, specifically billing code data, when studying adverse drug reactions. It is critical to audit electronic health records, which have temporally associated the use of a drug and an adverse reaction because of high rates of miscoding or lack of true cause and effect. Pending improvements in drug hypersensitivity coding in International Classification of Diseases, Eleventh Revision may make large databases even more useful.