The Duke University Hospital in Durham, North Carolina, is a 924-bed academic medical center that serves a wide range of patient populations. As such, the pharmacy is tasked with compounding custom, highly complex therapies, resulting in the dispensing of over 2800 IV fluids and syringes each day. With this kind of volume, medication errors are always a consideration, and IV mistakes are both exceedingly dangerous and difficult to detect. Should a patient-labeled IV product leave the pharmacy with an error, it is almost certain to reach the patient undetected.

In 2006, Duke’s pharmacy department began exploring options for using automation to support the quality-control work of its IV room staff. Few commercial solutions were available at that time, and robotic compounding systems were just beginning to be tested in a few US hospitals. Given the limited alternatives, Duke’s pharmacy department developed its own bar code checking application dubbed ScanIV.

System Development
The Duke pharmacy dispenses over 900 premixed solutions daily, many of which have similar names and packages. The original version of ScanIV—designed to check only commercially available, premixed
solutions—simply verified that the product dispensed matched the order in the pharmacy information system (PIS). The development of ScanIV effectively eliminated dispensing errors for premixed products.
During development, we modified our PIS labels to include a 2D bar code—a format more compact and readable than linear bar codes—that uniquely identifies each dose. The program and its intended workflows were designed and developed by the pharmacy department, then reviewed by a group of physicians, nurses, IT staff, medication safety experts, and human factors engineers. The success of our initial program led to additional enhancements that expanded bar code checking to nearly every dose dispensed by the IV room. The second generation of ScanIV was developed, tested, and rolled out in about four weeks in 2008.

Using ScanIV
ScanIV connects in real time to the PIS database to retrieve information about each specific IV dose as it is prepared and checked. Once a patient label is printed by the PIS, the technician or pharmacist assembles the components necessary to prepare the ordered product. The PIS label is scanned into ScanIV, which then displays the expected product contents on the screen. The user then scans each component’s bar code until all ingredients have been accounted for. If an incorrect item is scanned, the user is notified with distinctive audio and visual feedback, and further scanning is disabled until the error is acknowledged.
When the dose has been prepared, the pharmacist checks it before dispensing, also using ScanIV. The “verify” screen displays extensive order information to the verifying pharmacist, such as order status, current patient location, patient demographics, and information about who entered the order and what clinical warnings they received as a result. This gives the IV room pharmacist a robust view of the order, allowing them to exercise better professional judgment than would be possible by simply reading the IV label’s list of ingredients.

Ingredients included in the order that don’t require scanning, such as empty bags or syringes, can be flagged to skip the ScanIV check. Items also can be defined to automatically require additional components not specifically ordered in the PIS, such as filter sets. In addition, ScanIV is pre-programmed to determine which manufacturer bar codes to accept when verifying a specific PIS-ordered ingredient. Each drug can have an unlimited number of acceptable manufacturer product matches, allowing for new products to be added to the system.

ScanIV records all activity, including a record of successfully completed scans, scanning errors, scans aborted due to improper item setup, and a history of every item scanned. Daily reports and worksheets are automatically e-mailed to managers to identify opportunities for improvement.

Conclusion
Today, Duke’s IV room uses ScanIV to check syringes, multi-component IVs, stock solutions, and bags pumped from its IV compounding systems. It requires minimal maintenance and has never gone down. User acceptance has been excellent, as our pharmacists and technicians recognize that a simple scanning process serves as a safety net for potential errors. Since the enhanced version of ScanIV was completed in 2008, no significant adverse events involving pharmacy IV preparation have been reported at Duke University Hospital.

Timothy Dotson, MBA, MS, is an informatics pharmacist at Duke University Hospital in Durham, NC. He earned a BS from West Virginia University, an MBA from Heriot-Watt University, and a MS in education from California State University, East Bay.

Stephen Arrowood, PharmD, is manager of the sterile preparations cleanroom at Duke University Hospital. He earned his PharmD from the University of North Carolina.

Michael Canale, RPh, MS, is coordinator of pharmacy informatics for Duke University
Hospital. He earned his BS in pharmacy from Albany College and his MS from Binghamton University.