The error rates in administering anesthesia drugs are “markedly higher” than retrospective, self-reported surveys indicate, and many of them could be prevented with simple solutions, according to investigators from several prominent Boston institutions.
In a lecture entitled “Drug Errors: Challenges and Opportunities in Anesthesia,” presented at the 2017 meeting of the Society for Pediatric Anesthesia/American Academy of Pediatrics Section on Anesthesiology and Pain Medicine, Karen Nanji, MD, MPH, assistant professor of anesthesia at Harvard Medical School, described the methodology and findings of an observational study (Anesthesiology 2016;124:25) designed to identify solutions to prevent perioperative medication errors (MEs) and/or adverse drug events (ADEs).
Perioperative medication administration may be more prone to error than other settings where medication is administered, such as hospital wards, outpatient clinics and the critical care unit, according to Dr. Nanji, because of its uniquely high-stress, high-stakes environment, which moves at a faster pace and may involve bypassing standard safety checks that exist elsewhere in the hospital. Those safety checks include electronic clinical decision support and two-provider verification of medications (Table).
| Setting | Medication Error Rate, range | Supporting References |
| Hospital wards | 5%-15% | JAMA. 2001;285:2114-2120. JAMA. 1998;279:1200-1205. J Am Med Inform Assoc. 2010;17:575-583. |
| Outpatient clinics | 7%-12% | J Am Med Inform Assoc. 2010;17:78-84. J Gen Intern Med.2005;20:837-841. J Am Med Inform Assoc. 2011;18:767-773. |
| Critical care | 9%-20% | Crit Care Med. 2006;34:415-425. Pediatr Crit Care Med. 2007;8: 145-152. |
| Simulation | 0.5%-26.5% | J Am Med Inform Assoc. 2010;17:575-583. West J Emerg Med.2014;15:486-490. |
Dr. Nanji added that existing studies of self-reported errors in the perioperative setting have several problems, including low response rate, several biases that lead to underreporting of errors and the use of inconsistent denominators, rendering the task of comparing studies difficult. As an example of the problems with self-reported data, Dr. Nanji cited a study by Flynn et al (Am J Health Syst Pharm 2002;59:436) that found, out of 2,557 medication doses, 456 errors flagged by direct observation, versus 34 by chart review and only one by self-report.
In the present study, a team of trained observers—all of whom were anesthesiologists except for a nurse anesthetist—observed randomly selected operations at a 1,046-bed tertiary care academic medical center over an eight-month period to identify potential MEs and ADEs. The investigators also conducted a retrospective chart review to find potential events missed by observation. All identified potential events were reviewed by two independent reviewers, who were members of an adjudication committee composed of board-certified anesthesiologists and/or ME experts.
Dr. Nanji and her colleagues observed 277 operations, during which 3,671 medication administrations took place. Of these, 193 involved MEs or ADEs (5.3%). Of this subgroup, 153 (79.3%) were MEs and 91 (47.2%) were ADEs. The area where these two categories overlap, which the investigators defined as preventable ADEs (PADEs), involved 51 events (26.4%). Dr. Nanji noted that the Hawthorne effect—in which individuals being watched alter their behavior in response to observation—may have skewed the results, but that the observation protocol employed in the study was designed to minimize this effect.
No correlations were found between error rates and patient age, sex, American Society of Anesthesiologists physical status score, body mass index, procedure type or provider type, Dr. Nanji said. However, higher error rates were correlated with procedures lasting longer than six hours (P<0.0001) and those that had more than 13 medication administrations (P=0.002 for higher ADE rates). The four most common types of errors were:
- labeling errors (37; 24.2%),
- incorrect dosing (35; 22.9%),
- medication omission or failure to act (27; 17.6%), and
- documentation errors (26; 17.0%).
Dr. Nanji and her colleagues identified several specific solutions, both technological and process based, for reducing the rate of errors. Technological solutions included:
- barcode-assisted documentation, which they estimated would eliminate 17.0% of MEs and 25.5% of PADEs;
- computerized specific drug decision support (28.8% of MEs, 13.7% of PADEs and 58.8% of ADEs); and
- medication alerts (52.9% of MEs, 32.4% of PADEs and 94.1% of ADEs).
Process-based solutions included:
- changing the timing of documentation (35.3% of MEs, 21.6% of PADEs and 62.8% of ADEs);
- reducing the opportunity for workarounds (24.2% of MEs and 36.3% of PADEs); and
- connecting infusions to the most proximal IV port (1.3% of MEs and 2.0% of PADEs).
Dr. Nanji stressed that some of those numbers, while suggestive, should be “taken with a grain of salt,” as they assume perfect implementation, executionand use of the strategies involved. “But this can be a useful tool for prioritizing different interventions for future research,” she said.
Thomas Chokwe, MD, an anesthesiologist in the Department of Surgery at the University of Nairobi, in Kenya, said he found Dr. Nanji’s lecture valuable and hoped to see the reported findings put into practice.
“You have to do a lot of resource mobilization and spread this philosophy to other institutions,” Dr. Chokwe said, “so that people can see that you can do very simple things to improve outcomes elsewhere.”