Troy Brown

June 29, 2012

June 29, 2012 (Baltimore, Maryland) — Health systems that rely primarily on voluntary reporting of adverse drug events identify as few as 6% of events, according to information presented here at the American Society of Health-System Pharmacists Summer Meeting 2012 (ASHP SM12).

For this reason, several different methods should be used, explained James M. Hoffman, PharmD, MS, BCPS, medication outcomes and safety officer and associate member in pharmaceutical sciences at St. Jude Children's Research Hospital in Memphis, Tennessee.

Dr. Hoffman was part of a panel that presented strategies for improving detection of adverse drug events. The other speakers were Mike Cohen, RPh, MS, from the Institute for Safe Medication Practices in Hershey, Pennsylvania; David Classen, MD, MS, from Pascal Metrics, Washington, DC; and David Stockwell, MD, MBA, from Children's National Medical Center in Washington, DC.

Dr. Hoffman answered questions about the best methods for detecting adverse drug events in an e-mail interview with Medscape Medical News.

Medscape: Why are current traditional methods of adverse drug event detection inadequate?

Dr. Hoffman: Each approach for detecting adverse events has strengths and weaknesses. Different event detection methods find different opportunities for improvement. For example, some methods might find a particular phase of the medication use process better than others (eg, one method may find prescribing errors very well but not do a good job finding dispensing or administration errors; some methods are better at finding adverse drug events instead of medication errors). Therefore, multiple event detection methods are needed to get a more complete understanding of opportunities to improve the medication use process.

Voluntary incident (or event) reporting systems are common, and an important method of event detection. Many hospitals may rely primarily on voluntary event reporting systems, and when planning the conference session, one of our overarching objectives was to encourage session participants to go beyond voluntary reporting systems. Underreporting is a key limitation of error reporting systems. The estimates vary, but this is very well documented.

In 1 study, only 6% of adverse drug events had a corresponding event report. [Jt Comm J Qual Improv. 1995;21:541-548]

In another study, only 7.5% of the events detected by 4 different methods were also reported to the hospital's voluntary reporting system.

Medscape: What are some of the methods and data sources that can be used to detect medication errors and adverse drug events?

Dr. Hoffman: Four of the most common methods are:

  • incident reporting (individual reporting is the data source);

  • direct observation (trained observer collects data);

  • medical record review (trained reviewer looks at medical record); and

  • trigger tool (medical record review prompted by a specific action in the record document, such as a specific drug used).

As electronic medical records (EMRs) and other technology mature, more opportunities for novel event detection methods are emerging.

Medscape: Why is it important to use more than one method for detection of adverse drug events?

Dr. Hoffman: Multiple event detection methods are complementary, since each method finds different events. This is nicely illustrated by a study in Pharmacotherapy [2010;30:529-538] that used 4 event detection methods on a general medicine ward over an 8-week period. Only 19.8% of events were detected by multiple methods.

Medscape: How can electronic health records be used to improve detection?

Dr. Hoffman: Electronic health records are a key resource for event detection. For one thing, efficiency is improved with electronic records. Using the trigger tool approach with paper records, time must be spent acquiring and reviewing the record, but with an electronic record the information can be easily accessed. Or even better queries can be written to return the data needed.

An area of ongoing work with trigger tools is identifying and testing triggers that have the best positive predictive value (ie, that the trigger is identifying an actual adverse event). Two of the speakers in the session are experts in the trigger tool. Dr. David Stockwell from Children's National Medical Center in Washington, DC, leads a consortium of children's hospitals called the Automated Adverse Event Detection Collaborative, which is working to identifying pediatric triggers. St. Jude is a member of this consortium. Dr. David Classen at Pascal Metrics in Washington, DC, is a leading researcher in the use of the trigger tool. He has done some of the original studies of the approach.

Further, electronic records present many new opportunities to find events. As EMRs become more common, more methods for event detection can follow. It will really come down to asking what data in the electronic record will help you find events and improve medication safety, and then devising a plan to access the data.

Medscape: What can pharmacists do to improve event detection?

Dr. Hoffman: As pharmacists lead medication safety, it is important for pharmacists to embrace multiple event detection methods in order to improve the medication use process. Further, keep in mind that it is not just about finding events but ultimately about using the data to make improvements in the medication use process. While this is a rather basic point, at times it can be easy to get caught up in designing systems and collecting data, and one can lose sight of actually making improvements.

It is important to also use data from events that occur in other hospitals to make improvements. Reviewing what happens in another hospital and making improvements at your hospital is a great learning and improvement strategy.

Through their newsletter and other resources, the Institute for Safe Medication Practices (ISMP) is a great resource for descriptions of events and improvement strategies.

Medscape: Is there any research being done on this issue currently?

Dr. Hoffman: The trigger tool approach and leveraging EMR and other technology data are active areas of work.

Dr. Hoffman has disclosed no relevant financial relationships.