An audit of a hospital pharmacy’s drug accountability system uncovered a significant discrepancy rate with the anesthesia department of opioid and sedative use and emphasized the need for an electronic dispensing and auditing system.
Opioid and sedative use in the operating room at the University of Oklahoma Medical Center (UOMC), Oklahoma City, is reconciled by the Pharmacy Department, which historically has randomly selected one-tenth of anesthesia cases for review. The department’s recent quarterly reports showed a 2.95% incidence of cases in which drug discrepancies were noted.
At UOMC, anesthesiologists are given sealed envelopes from the Pharmacy Department with prescribed opioids and sedatives for their patients. In addition to entering drug information in the electronic medical record (EMR), the anesthesiologist is required to document on the envelope itself the amount of drug that has been received, administered, wasted or returned. Co-author Alberto de Armendi, MD, PhD, an anesthesiologist at the UOMC Children’s Hospital who is professor of anesthesiology and Robert W. and Elise B. Lykins Chair in Anesthesiology, University of Oklahoma Health Sciences Center, in Oklahoma City, told Anesthesiology News that this low-tech system has been in place since 2000.
To check the validity of the Pharmacy Department’s sampling method, a full audit was performed. Every EMR entry from UOMC’s Anesthesia Department and the associated notations on the envelopes from the Pharmacy Department were reviewed over a three-month period.
In all, there were 2,870 audits, with a discrepancy rate between pharmacy and provider records of 28.3%—significantly greater than the 2.95% rate found earlier through the pharmacy’s sampling method. Propofol proved to have the highest discrepancy rate (Table).
Table. Full-Audit Discrepancy Rates
Controlled Substance Discrepancy Rate, %
Dr. de Armendi noted that the full audit looked for all discrepancies. “Some of the discrepancies were inconsequential—names missing, placing two patient names in one yellow slip, unreadable drugs and dosages, etc.,” Dr. de Armendi told Anesthesiology News. Overall, however, these errors did not represent a large percentage. “I would guess less than 5%. Most of the discrepancies were differences in medications taken out and medications not accounted for in the EMR as having been administered. Because all the drugs studied were controlled substances, they were considered serious errors or discrepancies. Therefore, all of the other 95% [of errors] were serious.”
There was no attempt at qualitative analysis, Dr. de Armendi said. Still, about two-thirds of discrepancies were found to have occurred when more drug than the amount indicated in the record was administered, and most of these cases involved propofol.
Why propofol? The researchers discussed this among themselves, with varying theories. Administration of propofol in a pump might increase the discrepancy rate, they surmised, but other drugs also were administered by pump and did not show as high an error rate. They also realized that Anesthesia Department personnel were not generally taking out more propofol from anesthesia systems—they were ordering it correctly, from the Pharmacy Department. “We finally concluded that the only reasonable answer is that propofol is administered to practically every patient, thereby influencing the higher chances of error,” Dr. de Armendi said, adding, “not all patients get the same opioid or any of the other medications. Some patients get morphine, others are administered Dilaudid [hydroxymorphone], while others respond better to fentanyl.”
The audit “found a worrisome lack of accuracy between the patient record and the controlled substance pharmacy envelope.” The investigators concluded, “Alternative dispensing and auditing systems ideally should include only electronic systems to better reconcile daily drug administration and distribution of opioid and sedation medication.” Dr. de Armendi added, “We will be using Pyxis [CareFusion] anesthesia systems—one in each operating room—starting some time in the first half of 2015.”