… with alert overrides. … Villa Zapata L, Subbian V, Boyce RD, et al. Overriding drug-drug interaction alerts in …

Alert fatigue contributes to task burden and can threaten patient safety. In this study, researchers at one academic medical center found that machine learning techniques could enable intelligent filtering of medication alerts and reduce alert volume by 54%.

Computerized decision support alerts for drug-drug interactions are commonly overridden by clinicians. This study examined fifteen well-known drug-drug interactions and identified risk factors that could reduce risk in the majority of interactions (e.g., medication order timing, medication dose, and patient factors).

Improving diagnosis is a patient safety priority. Using data from patient safety incident reports, malpractice claims, morbidity and mortality reports, and focus group responses, this study sought to identify “diagnostic pitfalls,” defined as clinical situations vulnerable to errors which may lead to diagnostic errors. The authors identified 21 generic diagnostic pitfall categories involving six different aspects of the clinical interaction – diagnosis and assessment, history and physical, testing, communication, follow-up, and other pitfalls (e.g., problems with inappropriate referral, urgency of the clinical situation not appreciated). The authors suggest that these findings can inform education and quality improvement efforts to anticipate and prevent future errors.

One element of conservative prescribing is minimizing the number of medications prescribed. This study compared the number of unique, newly prescribed medications (personal formularies) of primary care physicians across four health systems. Results indicated wide variability in the number of unique medications at the physician and institution levels. Further exploration of personal formularies and core drugs may illuminate opportunities for safer and more appropriate prescribing.

Clinical decision support (CDS) systems are intended to support diagnosis and therapeutic processes of care. This position paper defines adaptive CDS as “systems that can learn and change performance over time, incorporate new clinical evidence, data types, data sources, and methods for interpreting data.” Recommendations for the effective management and monitoring of adaptive CDS are outlined.

Quality measurement is increasingly being incorporated into policies outlining healthcare provider reimbursement. This study compared quality measure calculations between an individual electronic health record (EHR) source and the same EHR source combined with health information exchange (HIE) data. The results show that adding HIE data changed 15% of quality measure calculations. The authors suggest that incorporating HIE data into reimbursement programs could promote more accurate and representative quality measurement.

Patient misidentification can lead to serious patient safety risks. In this large academic medical center, displaying patient photographs in the electronic health record (EHR) resulted in fewer wrong-patient order entry errors. The authors suggest this may be a simple and cost-effective strategy for reducing wrong-patient errors.  

Medication errors occur frequently in the outpatient setting and can lead to patient harm. A common scenario is one in which a patient is prescribed multiple medications, does not know what each one is for, and takes them incorrectly. Medication safety experts have advocated that prescribers include indications on prescription labels to enable patients and pharmacists to check the bottle in order to remember a medication's purpose. Investigators examined more than 4 million outpatient prescriptions from a single institution and found that only 7.4% of prescriptions included an indication. Medications for symptoms like pain, nausea, and anxiety were much more likely to have indications than medications for chronic diseases. Internal medicine physicians, whose patients are more likely to take multiple medications, wrote indications 6% of the time. A PSNet perspective explored how community pharmacists can use medication indications and other tools to ameliorate medication-related harm.

Having multiple patient records open in the electronic health record increases the potential risk of wrong-patient actions. This randomized trial tested two different electronic health record configurations: one allowed up to four patient records to be open at a time, and the other allowed only one to be open. Among the 3356 clinicians with nearly 4.5 million order sessions, there were no significant differences in wrong-patient orders. However, the investigators noted that clinicians in the multiple records group placed most orders with just one record open. A post hoc analysis determined that the rate of errors increased when orders were placed with multiple records open. A related editorial highlights the tradeoffs between safety and efficiency and argues for examining the context of the two configurations, including throughput and clinician satisfaction. A previous PSNet perspective discussed assessing and improving the safety of electronic health records.

… associated with computerized decision support. … Wright A; McEvoy DS; Aaron S; McCoy AB; Amato MG; Kim H; Ai …

This simulation study compared computerized medication order entry between two commercial electronic health records and a prototype designed for safe prescribing. Physicians using the prototype had fewer errors compared to either commercial platform, highlighting the need to improve electronic health record usability in order to enhance medication safety.

This direct observation and interview study examined safety activities associated with electronic health records. The study team identified decision making, organizational learning, and frontline user engagement (including clinical decision support and training) as the three critical areas for safe use of electronic health records.

Although clinical decision support is intended to improve safety, decision support alerts often result in alert fatigue and overrides. This prospective observational study examined overrides for exceeding the maximum dose of a medication in the intensive care unit. Researchers determined that insulin was the most frequent medication for which a maximum dosage alert was overridden. In almost 90% of cases, the overrides were deemed clinically appropriate. The authors conclude that more intelligent clinical decision support for medication dosing is needed to balance safety with alert fatigue in the intensive care unit. A past PSNet perspective discussed the challenges of implementing effective medication decision support systems.

Although clinical decision support is an important safety tool, unintended consequences include frequent alerts with resultant alert fatigue and overrides. This observational study investigated whether override comments for clinical decision rules could be used to determine if a rule was broken. Investigators discovered that malfunctions were prevalent in clinical decision support. While the frequency of comments did not predict the presence of a rule malfunction, strongly negative comments were associated with erroneous decision support. The authors recommend routine monitoring of alert override comments to identify errors related to clinical decision support. A past WebM&M commentary recommended employing human factors engineering to make clinical decision support more effective.

Health information technology (IT) implementation is a complex endeavor that requires a sociotechnical orientation to succeed. This article outlines nine key challenges to safety that must be addressed across the three stages of health IT: design and development; implementation and use; and monitoring, evaluation, and optimization.

Although electronic prescribing has been shown to reduce prescribing errors, the impact on adverse drug events remains less certain. Overriding of drug–drug interaction alerts and inclusion of free-text notes that contain inaccurate information within electronic prescriptions suggests that safe prescribing requires more than an electronic system. To improve the safety of electronic prescribing, the authors make several broad recommendations. They suggest including drug indications on prescriptions, ensuring a readily available and accurate medication list, notifying pharmacies when previously prescribed medications are canceled, using standard prescription instructions, improving decision support, and promoting consideration of nondrug options. A previous WebM&M commentary discussed an incident involving an electronic prescribing error. The Moore Foundation provides free access to this article.

This study examined ambulatory follow-up of test results by aggregating multiple types of data—national surveys on safety culture and patient satisfaction; patient complaints; safety reports; and electronic health record audits of provider response times. Researchers found an association between quicker response time for test results and higher patient satisfaction. They conclude that merging these disparate data sources can uncover new levers to improve patient safety.

Clinical decision support systems are widely utilized to improve patient safety by alerting providers to potential medication errors and other safety concerns. However, these alerts are frequently overridden by clinicians and contribute to alert fatigue. Researchers sought to assess the impact of a transition from a legacy electronic health record system to a commercial system on provider reactions to drug–drug interaction alerts in the ambulatory setting. There was a sixfold increase in the burden of interruptive alerts with adoption of the commercial system, and clinician acceptance for the most severe interaction alerts decreased from 100% to 8.4%. The authors suggest that the timing of alerts in the workflow process, user-interface design factors, and the inability to distinguish between more and less severe drug interactions all contributed to the decline in provider acceptance after adoption of the new system. A previous WebM&M commentary discussed a case in which a patient experienced an adverse drug event after a clinician overrode a prescribing alert.