Abstract
Background: In a previous study of adverse drug events (ADEs) in the ambulatory setting, we found that 42% of the serious, life-threatening, and fatal ADEs were preventable. We identified 7 types of errors in medication prescribing and monitoring that were the proximal causes of these events:
prescribing a medication for which the patient has a history of allergy,
a medication that conflicts with recent laboratory values or the patient’s medical conditions,
a medication that interacts with other medications the patient is taking or
an excessive dose,
inadequate monitoring or
not acting on the results of monitoring, or
the patient does not use prophylaxis for known side effects of the medication.
The aim of this study is to assess the feasibility of applying fault tree analysis and probabilistic risk assessment to the analysis of errors contributing to ADEs.
Methods: We applied probabilistic risk assessment techniques borrowed from the safety engineering industry to identify system-level interventions for reducing the rate of ADEs. Interdisciplinary clinical including primary care and specialist physicians, advanced practice nurses, nurse managers, and pharmacists were convened to construct fault trees (graphic representations of events contributing to an undesired outcome) for the 7 types of errors in prescribing and monitoring medications. We are using probabilistic risk assessment software (SAPHIRE) to evaluate the fault trees, develop cut-sets, and identify and analyze potential system-level interventions to prevent ADEs.
Results: Participants learned to develop complex and thorough fault trees with little difficulty. The clinicians quickly became comfortable detailing the systemic and individual processes that lay behind serious medication errors. The fault trees reveal an array of opportunities to implement interventions with a high probability of improving patient safety.
Conclusions: Techniques for risk reduction from industry appear to be of value in assessing and improving safety problems related to medication prescribing and monitoring.
- Received September 11, 2008.
