The Effects of an Experimental Medication System on Medication Errors and Costs Part I: Introduction and Errors Study

This paper is the first in a series summarizing the development and evaluation of an experimental (centralized unit-dose dispensing type) medication system at the University of Arkansas Medical Center, Little Rock. This series is abstracted from the two-volume, 848-page report to the U.S.P.H.S. on this three-year project. The project was conducted by a multidisciplinary research group and the investigators having primary responsibility for the various parts of the project are so identified in the report. The objectives of the system itself were to reduce the frequency of medication errors, improve the utilization of pharmacists and nurses, increase the control over storage and distribution of drugs throughout the hospital, and provide an easily accessible source of drug usage data for medical audit and research purposes. The experimental system differed radically from conventional systems. Differences included delivery of all doses needed in ready-to-administer form from the pharmacy on a dose-by-dose basis at the time of administration, editing of all drug orders by a pharmacist and use of low cost data processing and teletype equipment to maintain all drug records throughout the hospital automatically by remote control from the pharmacy. The primary objective of the evaluation was to determine whether the switch to the new system was accompanied by significant and desirable changes in selected parameters expected to be affected—errors, utilization, job satisfaction and costs. The acceptance and legal implications of the new system were also studied. Serious delays in physical alterations disrupted the original study plan; although the existing system was studied for six months the experimental system study had to be squeezed into a two-month period with the system serving about three-fourths of the target medication workload (medicine, surgery and pediatrics—eight divisions) and incomplete in some respects. The findings of the evaluation studies other than errors and cost are summarized briefly and evidence showing that base conditions were not dissimilar during the control and experimental periods is presented. During the medication errors study, the data were collected by direct observation of nurses by pharmacist-observers during 192 eight-hour work shifts for the existing system and 64 for the experimental system. The errors detected were 3,558 for the control period and 443 for the experimental period. Following the installation of the experimental system, the error rate declined from 31.2 to 13.4 percent. If wrong time errors are discounted, the decline was from 18.3 to 3.3 percent. If both wrong time and wrong brand errors are excluded, the decline was from 13.0 to 1.9 percent. Clues to the causes of many of the errors are presented although concern for causation was outside the scope of the evaluation.