Pharmacists at MedStar Health could have spent decades manually implementing dose-range checking (DRC) to enhance patient safety across the organization's entire medication catalog.1 Instead, leaders at the not-for-profit health system serving Maryland, Virginia, and Washington, D.C., worked with Cerner's Continuous Improvement team to apply DRC parameters to 98% of MedStar's medications in just three months.
Their work spurred providers to collectively accept nearly 150 more alerts per day, while minimizing additional alerts received per provider.2,3
DRC helps tackle a common health care challenge — medication-related errors. Each year, these errors cost $42 billion globally and cause at least 7,000 deaths in the U.S. They occur at every phase of the medication process but are most common at the ordering or prescribing stage.4
"DRC assists with preventing one of the most common forms of adverse drug events, serving as a guardrail against potential drug-dose or drug-frequency prescribing errors", explained Joel McAlduff, MD, vice president and chief medical information officer. "It provides objective insights based on established dosing parameters designed to balance prescriber preference against potential human error. We have seen the benefits firsthand."
In one case, an emergency physician selected an order for vancomycin in an elderly septic patient with unknown kidney disease. After receiving an alert recommending a reduced dose due to renal dysfunction, the provider modified the order to align with DRC guidelines.
Many health care leaders would agree that if DRC prevents even a single adverse drug event (ADE), it's worth implementing. Even so, practical challenges can prevent organizations from fully utilizing DRC. Leaders often want to localize generic dosing guidelines to reduce perceived over-alerting, but traditional customization methods can be tedious to implement and challenging to get right.
"If dose-range guidelines are too conservative, they can contribute to provider alert fatigue," said Bonnie Levin, PharmD, AVP, pharmacy services. "Customized DRC guidelines help establish a middle ground between under- or over-alerting, but manual customization for thousands of potential medication prescribing scenarios is time-consuming."
Taking a different approach, MedStar and Cerner Continuous Improvement teams used simulated alert data to tackle implementation barriers and make DRC coverage attainable.
The Continuous Improvement team analyzed a month of MedStar's prescribing data against Multum®, Cerner's drug, herbal and nutraceutical database that assists with safe medication use. The team determined how many alerts each medication would generate if MedStar implemented Multum-based DRC guidelines.
The analysis revealed that Multum-based parameters wouldn't significantly increase alerts for 97% of medications and that another 2% of medications didn't have real dosing risks. That freed MedStar leaders to focus manual DRC reviews on the 1% of medications driving 73% of their medication alerts.
"We suddenly had an efficient, data-driven path to implementing comprehensive DRC sooner," said McAlduff. "There were compelling statistics based on our own prescribing data, backed by a solid mitigation plan to monitor alerts and intervene quickly if necessary."
After expanding DRC coverage, MedStar providers accepted an average of 124% more alerts per day.2 They also accepted alerts from 653% more distinct medications.5
The project strengthened patient safeguards without over-burdening providers, and post-implementation alert volumes aligned closely with projections. On average, physicians saw less than one additional alert per two days worked. Pharmacists received just over three additional alerts per day worked.3
"We received very few provider complaints," said McAlduff. "And we were able to address them within just a few days, thanks to the rich data. We continue making incremental DRC adjustments to reach the most effective balance between safe prescribing and provider preference."
1 Using traditional DRC implementation methods, manual review of 25 medications requires approximately three months and typically accounts for less than 1% of all medications. Three months (1%) * 100 = 300 months or 25 years to achieve 100% DRC coverage.
2 Comparing 120 average accepted alerts per day from May 10, 2018-Aug. 5, 2018 to 269 from Nov. 1, 2018-Feb. 1, 2019
3 Comparing May 10, 2018-Aug. 5, 2018 to Nov. 1, 2018-Feb. 1, 2019; average alerts received per day increased from 0.28 to 0.71 for physicians and increased from 2.20 to 5.51 for pharmacists
4 Tariq RA, Scherbak Y, "Medication Errors," StatPearls Publishing, Jan. 11, 2019, https://www.ncbi.nlm.nih.gov/books/NBK519065/
5 Comparing 118 distinct medications with accepted alert from May 10, 2018-Aug. 5, 2018 to 889 from Nov. 1, 2018-Feb. 1, 2019