Study points to methods for safe drug dispensing via computer

COLUMBUS, Ohio — Researchers have found that a new computer system that uses bar codes to safeguard patients’ medications will work successfully, but not without creating new, serious problems for nurses charged with patient care.

“In general, we viewed the system as successful. There are no magic bullet solutions to human error in any setting, and even the best systems will require constant maintenance and flexible redesign after implementation,” said Emily Patterson, a research specialist in Ohio State’s Institute for Ergonomics.

The Veterans Health Administration (VA) recently designed a drug dispensing system called Bar Code Medication Administration (BCMA), and asked Patterson to evaluate it.

At issue is whether bar codes could enable healthcare professionals to verify that a patient is receiving the right drug, at the right dose, at the right time.

Patterson conducted the research with the VA Midwest Patient Safety Center of Inquiry in Cincinnati and published the results in a recent issue of the Journal of the American Medical Informatics Association. Her coauthors include Marta Render, director of the center and adjunct associate professor of internal medicine at the University of Cincinnati, and Richard Cook, director of the Cognitive Technologies Laboratory at the University of Chicago. The Department of Veterans Affairs funded the study.

With BCMA, hospital pharmacies label medications with bar codes, and patients wear bar-coded wristbands. Nurses scan a patient’s wristband, and a laptop computer on the medication cart displays that patient’s prescriptions. Before giving the medicine, the nurse scans the medicine bottle or other container, and BCMA records the drug as delivered. If the nurse accidentally scans the wrong medicine or dosage, or tries to give medicine at the wrong time, a warning pops up on the computer screen.

Patterson followed the activities of 26 nurses at three VA hospitals as they dispensed medication with BCMA. She also watched as doctors entered new prescriptions into the electronic medical record, and pharmacists labeled prescriptions. Then she interviewed these people as well as hospital computer support personnel and nurse managers, to gauge everyone’s opinion of the system.

The study did not specifically examine errors caught or prevented by BCMA, but focused instead on the interaction of users with the system, in order to find ways to make the system work better.

Nearly all VA hospitals are now using BCMA software version 2.0, and Patterson and her colleagues are helping create version 3.0, which will address some of the problems found during the study. After that, Patterson and her colleagues will continue to help the system evolve over time.

Patterson cited a 1999 study at Brigham and Women’s Hospital and Harvard Medical School that found medication errors fell 86 percent when doctors began entering their prescriptions orders via computer. When computer systems are optimized for taking human factors into account, errors can decrease even further, she said.

The new Ohio State study found five unanticipated negative side effects of introducing BCMA to hospitals:

? Sometimes the computer automatically removed medications from a patient’s prescription list. For example, one patient could not receive his dose of a drug on time, because he had been away in another part of the hospital when he was supposed to receive it. When the patient returned to the ward and the nurse administered his medications, BCMA no longer displayed the medication because it was dropped when a time window had elapsed. In most cases, the nurses knew to administer the medication and so asked a pharmacist to add it back to the prescription list.

? There was less coordination between doctors and nurses, compared to a paper-based system. Doctors reviewed patient’s medication orders less often, because doing so through the BCMA computer was more difficult or time-consuming than the old procedure, which involved simply reading the nurses’ notes on a paper medical chart. That means doctors and nurses were less likely to know if a patient’s medication needed to be changed.

? During the busiest parts of the day, nurses had to ignore some of the required BCMA procedures to save time. For instance, bar codes didn’t always scan properly on the first try. To avoid re-scanning a patient during crunch periods, nurses would often enter the seven-digit bar code number manually.

? Nurses became anxious about delivering medications on time. The computer required the nurses to type an explanation when medications were given even a few minutes early or late, and nurses were concerned that the late administrations would reflect badly on their job performance. As a result, nurses tended to make just-on-time administration of medicines a high priority, compared to other duties.

? The computer didn’t easily accept unusual dosage orders. While the system streamlined the administration of consistent dosages of drugs, it wasn’t set up to accept dosages that increased or decreased over time. For example, pharmacists had to enter 14 separate daily doses for a patient whose medicine was supposed to taper off over a two-week period.

In the future, Patterson and her colleagues will examine how BCMA is used differently in acute care wards, nursing homes, and intensive care units. Roger Chapman, another research specialist in Ohio State’s Institute for Ergonomics, is going to investigate how nurses’ use of PDAs, or personal digital assistants, instead of laptop computers will affect the use of BCMA.