By Agatha L. Nolen, MS, DPh, FASHP and W. Dyer Rodes, II, MD

"I almost made a life-threatening medication error. Can we meet today to discuss how to prevent this from ever actually happening?"

A “Near Miss”
It was early morning in the spring of 2004, when I received this phone call from one of Centennial Medical Center’s anesthesiologists regarding a case in the OR. Our pharmacy department had a great relationship with our anesthesiology staff; in fact, they were the biggest supporters of our patient safety initiatives. Later that day, I received more information on the “near miss” in question.

Pharmacy supplied cefazolin 1 g and 2 g (mixed in 50 mL of fluid) in IV piggybacks, which our anesthesiologists typically administered following the induction of anesthesia. All other drugs for use by the anesthesiologist were provided in a “case tray,” which included common pre-filled syringes, as well as vials and ampoules of items that were used less often. We also mixed any IV infusions prior to the case, usually in 250 mL of fluid. Until recently, the pre-op antibiotics were the only drugs mixed in a 50-mL piggyback. Our cardiac anesthesiologists aggressively controlled blood glucose levels to reduce mediastinal infections. We started routinely sending an insulin drip (100 units in 100 mL NS) on selected OR cases, such as open-heart surgery. To a busy anesthesiologist, the insulin drip could look a lot like the pre-op antibiotic. Luckily, in the case of this particular “near miss,” the anesthesiologist read the label before opening the clamp on the IV infusion line, and a medication error was averted. However, it was too close a call.

Immediate Reactions vs. Long-Term Solutions
As a temporary solution, our pharmacy staff agreed to place a large, yellow label marked “INSULIN” in bold, black letters on all insulin drips. While color-coding (use of color to identify drug classes) has long been utilized as a safety feature in drug administration, there have been no large-scale research studies demonstrating the effectiveness or dangers of color differentiation (use of color to identify a specific product). Furthermore, following three infant deaths at a hospital in Indianapolis, the FDA warned of the potential for life-threatening medication errors resulting from confusion between two heparin products with similar blue labels.1 Among other reasons, the finite number of humanly distinguishable colors clearly demonstrates the limitations of color differentiation as a medication safety tool.

Bar Coded Medication Administration
Because we had recently completed the implementation of a bar coded medication administration (BCMA) system on all inpatient units, we felt we could use similar technology to increase the safety of our patients in the OR.

We outlined an OR anesthesiologist’s workflow for medication administration and compared it to our nurses’ bedside medication administration workflow. Obviously, the processes, cycle time, and people involved in each situation are very different. Each step of the inpatient nursing unit medication-use process is distinct, and does not overlap the next, and each step typically involves a different caregiver.

In the OR, each phase of prescribing, dispensing, administering, and monitoring is handled by the same person: the anesthesiologist. In addition, the timeframe is usually measured in seconds or minutes, rather than hours. Because of the urgency of the patient’s condition, documentation of drug administration often happens after the fact or at the end of the case, if at all. Reports from our OR pharmacy staff indicated that 20 to 40% of medications administered by anesthesiologists are not documented on a handwritten anesthesia record. With this in mind, we felt a BCMA system for the OR could address both safety and documentation issues.
With plans in place to expand our use of BCMA to outpatient areas, such as the emergency department, the cardiac catheterization laboratory, and pre-op holding areas, implementing a similar system in the OR seemed a natural next step in our BCMA patient safety plan.

Automated Clinical Anesthesia Information Systems
We reviewed a handful of commercially available automated clinical anesthesia information systems and selected two vendors as finalists. From our anesthesiologists’ viewpoint, these systems were ideal, because, in addition to capturing drug administration information, they also interfaced with patient monitoring equipment. Consequently, their output at the end of the case is a comprehensive, legible, and complete document that fully captures the patient status changes and medication administration throughout the case. We decided to conduct trials on two different clinical anesthesia information systems. However, after a short installation phase, the anesthesiologists felt that one system was too cumbersome and did not offer the same advantages — bar coded medication administration, experation-date verification, and the minimal amount of user training needed — as the other system. They voluntarily stopped using the less satisfactory system, and we began a pilot study with the other system, SAFERsleep, in September 2004.

System Features and Functions
Using the SAFERsleep system, bar coded medications are dispensed to the anesthesiologist in an anesthesia tray. Upon administration, the anesthesiologist scans the medication’s bar code and the system provides both a visual and auditory confirmation of the administration; the system actually “speaks” the name of the drug and displays it on the anesthesia workstation computer screen. The time of administration is captured electronically, precluding the need for written documentation. Since many drugs are typically titrated, the anesthesiologist can adjust each dose administered online, and re-scan the syringe or vial when a subsequent dose is administered. Built-in system alerts remind caregivers of events that need to be completed or documented. For instance, the system reminds the practitioner to administer the pre-op antibiotics following the induction of anesthesia. If the anesthesiologist chooses to not administer a pre-op antibiotic, he must acknowledge a system warning. SAFERsleep will also notify the practitioner of any medications that have exceeded their expiration dates.

Implementation
From the pharmacy standpoint, the SAFERsleep system was easy to set up, and we were able to use it to produce bar coded labels for the pre-filled syringes and vials in our anesthesia trays. In addition, we reviewed the system’s onboard generic drug dictionary and made some minor modifications to add products that were specific to our institution. And, in order to capture each drug administration, we included standard concentrations of IV infusions, as well as blank IV fluids in our database. We participated in a two-week crossover study in which both handwritten anesthesia records and the SAFERsleep system were utilized. We verified that there was a 100% match between the written records and the SAFERsleep system.

We implemented the system in three open-heart rooms and scanned over 800 patients in the first year. After one year, we extended the implementation to additional OR suites and asked the anesthesiologists to continue using the system on a voluntary basis or, if they preferred, to resume charting their drug administration on a written anesthesia record. Our anesthesiologists’ acceptance of the system was excellent, with an eventual 100% voluntary participation by the cardiac anesthesiologists within a year after the pilot study concluded.

Results
We observed a 20%-plus increase in documentation capture with the system compared to the written anesthesia record. We had no reports of medication errors in cases where the SAFERsleep system was utilized, and because we now have an electronic database of drug administrations on over 1,000 cases, we are beginning to analyze our drug utilization to reveal opportunities for practice improvement. We are also planning to build templates for each case type based upon our historical data.

Prior to the SAFERsleep implementation, the anesthesiologists’ handwritten records were often incomplete or difficult to read. Furthermore, CMS billing regulations require that a patient be billed only for medications documented as administered on the medical record. Using SAFERsleep’s computerized record, our OR pharmacy staff can more efficiently and accurately bill our patients for medications administered, and are able to match 100% of our administered and billed medications. In summary, we successfully implemented the SAFERsleep BCMA system in our OR, resulting in improved documentation, billing processes, and, of course, patient safety.

Agatha L. Nolen, MS, DPh, FASHP, is the director of OPPS (Outpatient Prospective Payment System) with HCA’s corporate offices in Nashville, Tennessee, and is also working toward a PhD in public administration at Tennessee State University. At the time of this study she was the director of pharmacy at Centennial Medical Center in Nashville, an HCA facility.
W. Dyer Rodes, II, MD, is the chairman of the board of trustees and a member of the department of anesthesiology at Centennial Medical Center in Nashville.

References:
1. FDA Medwatch: February 7, 2007. http://www.fda.gov/medwatch/ safety/2007/safety07.htm#Heparin. (accessed February 14, 2007)