Session number: 1015
Abstract title: Appropriateness of Clinical Response to Adverse Drug Events
Author(s):
CR. Weir, Ph.D. - SLC VA GRECC and University of Utah
JF. Hurdle, MD, Ph.D. - SLC VA GRECC and University of Utah
JR Nebeker, MD - SLC VA GRECC and University of Utah College of Medicine
JM Hoffman, PharmD - SLC VA GRECC
BS Roth, PharmD - SLC VA GRECC
Objectives: Early detection and appropriate treatment could significantly minimize the negative impact of Adverse Drug Events (ADEs). The objectives of this study are to: 1) identify predictors of ADE detection by clinicians from a gold standard review and 2) assess the appropriateness of the clinician response to an ADE.
Methods: The study reported here was a subset of a larger prospective study assessing the incidence rate of ADEs through concurrent review (HSRD SAF 122-98). Nine hundred and thirty seven cases were randomly selected from inpatient admissions (41%) across a 4.5-month period. Cases were reviewed concurrently by a PharmD and discussed until agreement by an expert group panel composed of 2 physicians, 2 PharmDs and a nurse. A total of 521 ADEs were detected and characterized in terms of severity, preventability, causality, pharmacological type, and drug class. Detection was defined as either evidence of documentation in notes or documentation plus appropriate treatment of manifestations. Appropriate treatment was defined by group consensus using standardized protocols when available.
Results: Detection. Of the 521 ADEs detected by gold standard surveillance, providers failed to detect 218 (or 42%) of the cases using the documentation definition. Failure to document an ADE was associated with ADE severity, ADE manifestation, provider characteristics, drug class and ADE category, not the presence or absence of error. When the detection definition included either documentation or appropriate treatment, the number of non-detected ADEs dropped to 17%. Non-detected ADEs by this definition were significantly associated with severity, provider characteristics, drug class and error status, but not ADE category.
Clinical Response. Across all cases, 27% (n=137) were only partially treated and 8% (n=44) were not treated at all. Failure to treat appropriately was significantly associated with ADE severity, patient characteristics, error status and drug class. Failure to treat resulted in increased utilization of medical services.
Conclusions: Clinicians fail to both detect and appropriately treat a small, but significant number of ADEs. Both ADE and patient characteristics predict provider behavior.
Impact statement: Failure to appropriately detect and treat ADEs negatively impacts utilization of medical services and patient outcomes. Future interventions should target predictors of practice patterns.