Study Finds Computers Greatly Reduce Prescription Errors

11/04/02    Portland, Ore.

Computer prescriptions are three times less likely to contain errors than handwritten prescriptions

NOTE TO EDITORS: Kenneth Bizovi, M.D., is available today for interviews and to demonstrate the computer-assisted prescription program. A pharmacist who worked with the study team is also available to discuss how this new prescription method has impacted the pharmacy's ability to deliver medication. Electronic examples of a computer-assisted prescription, as well as a handwritten prescription are available at http://www.ohsu.edu/news/110402scriptImages.html.

Have you ever received a drug prescription from a physician that looked like chicken scratch? You're not alone. Pharmacists sometimes have a hard time reading prescriptions and in some cases they also are incomplete. To avoid errors, pharmacists have to spend precious time tracking down prescribers to clarify illegible or possibly inaccurate prescriptions. A new study by researchers at Oregon Health & Science University shows that prescriptions written on a computer are less likely to contain errors.

A study by the OHSU School of Medicine's Department of Emergency Medicine study found that prescriptions initially entered into a computer reduce errors in the prescription by one-third and are five times less likely to require pharmacist clarification than handwritten prescriptions. The study was published in the November 2002 issue of Academic Emergency Medicine (www.aemj.org), published by Hanley & Belfus.

"This new computer system is really the foundation for ongoing improvements in safe prescribing practice," said Kenneth E. Bizovi, M.D., assistant professor of emergency medicine in the OHSU School of Medicine, toxicologist and emergency physician.

There are only a few studies on prescription error rates, and none conducted an emergency department. OHSU's researchers thought that a new prescription computer program, implemented in the hospital's emergency department March 2000, provided the perfect opportunity to investigate a more effective method of prescribing medications. The team compared standard handwritten prescriptions created before the computer program was implemented with computer-assisted prescriptions created after implementation. The computer program allowed prescibers the ability to choose a drug from a list of available medications. Each prescription contained necessary information, such as dose, quantity, frequency and amount to dispense. This system greatly reduces the possibility of writing a prescription that is incomplete or that contains a formulation that does not exist.

For example, instead of prescribing 500 mg of ibuprofen, which is not an available dose, a physician can look at the doses that do exist and choose from those. This reduces the risk of dosing errors and reduces the need for pharmacists to clarify the prescription.

Patient information, such as name, medical record number and age, were also included in the computer-written prescription. This information, along with the prescription and prescriber's name, was then printed legibly for a pharmacist to read. All the information becomes part of the patient's computer medical record instantly. This is an improvement from the standard handwritten format, which uses a prescription pad, a stamp of the patient's information and the handwritten prescription, creating only one copy of the prescription and requiring a separate entry into the medical record.

The research team hoped the computer-assisted prescription system would create legible prescriptions that decreased errors related to dosing, missing information, incorrect information, legibility and ordering of drugs that weren't available. These types of errors require a pharmacist to track down the prescriber to clarify before the prescription can be filled. Although this occurs infrequently, it takes extra time from the pharmacist and could lead to increase time to get a prescription filled.

The study proved their theory. With the assistance of the OHSU Hospital Pharmacy, researchers were able to track the notations on the prescription made by pharmacists when making a clarification. Of the 2,326 handwritten prescriptions filled by OHSU Hospital's Pharmacy before implementation of the computer program, 2.3 percent of them contained errors. The OHSU Hospital Pharmacy received 1,594 computer-assisted prescriptions, only 0.8 percent of which contain errors that required clarification by a pharmacist.

Even some pharmacists commented on the improvement. "They said you could read the prescription, which was great. They are very concerned about legibility," said Bizovi.

"OHSU is at the forefront in using the power of computer technology to ensure patient safety," said Christine Cassel, M.D., dean of the OHSU School of Medicine and a co-author of the 1999 Institute of Medicine report "To Err is Human," which prompted a national dialogue on medical errors. "In the IOM report, experts emphasized that information science can help create systems which deliver patient care of higher quality and also keep costs down because they are more efficient."

OHSU's ED is one of the few around the country using this prescription program. As more medical clinics acquire computerized systems, Bizovi feels the computer-generated prescriptions will prove to be effective in reducing errors in many medical practices. Since the time of the study the, OHSU Department of Emergency Medicine has integrated patient allergies into the prescription, adding one more safety improvement to its prescribing practices.

###