Detection and prevention of medication errors using real-time bedside nurse charting
This study was performed at the LDS Hospital and evaluated the nursing compliance with medication administration processes utilizing the HELP (Health Evaluation through Logical Processing) hospital information system and its clinical decision support (CDS). The major nursing parameters studied included real time charting rate and the bedside charting rate.
In spite of a well developed clinical decision support structure within HELP, it was noticed that medication errors were still occurring that should have been prevented by the system. These included early and late administrations, wrong drugs given, missed doses, and doses given after drug discontinued by physician. Direct observation detected that the workflow processes of nursing had bypassed the established system of built in alerts and warnings. If the nursing processes had been adhered to with documentation of medication administrations done at the patient’s terminal in the room, the errors would have been prevented. A multidisciplinary team called the medication error team (MET) ascertained that in 2000, over a nine-month period between April and December that a total of 829 medication errors were reported. They found that 240 of these, or 29%, should have been easily prevented if the HELP and clinical decision support (CDS) alerts and warnings had been adhered to as developed. This included adherence to the existing nursing training given for medication administration.
A prospective study was designed with two 40 bed clinical surgical units with similar nursing staffs, medication order volumes, and physical setup. It was proposed to evaluate the effectiveness of performing a 12 week nursing education intervention and to see if it had an effect on the compliance of the medication administration, process and if it improved utilization of the HELP system and its CDS. One group would receive the educational intervention and the other group would be the control.
The baseline study was done and confirmed by additional direct observational studies that the nurses in both groups were routinely bypassing nursing procedures and documenting medication administrations at the nursing computer terminals and not within the patients’ rooms. The standard management reports would not have detected this process. In addition, the nurses were also using paper printouts of their medication administration records which allowed them to miss new or discontinued orders if the current computer records were not checked.
The two nursing clinical surgical units began to be evaluated and compared in 2002. There were a total of 118,612 medications charted. The study group had 55,080 and the control group had 63,532. During the 12 week interventional educational study, there was a statistical significant improvement in the group receiving the education versus the control group. Real time charting increased in the educated group from 59.5% to the post study of 75.8%. The control group 53.5% pre-study and 56.9% post study. The bedside charting rate also showed a statistically significant difference between the groups. The educated group changed from 40.6% pre-study to 66.6% post study. The control group had little change from 33.6% pre-study to 38.5% post-study. Equally impressive in this study were the gains reported and maintained one year at follow-up. The educated group’s real time charting was 75% versus 57% for the control group. The bedside charting rate also was maintained with the study group maintaining a rate of 58% and the control group was 43%.
Comment: This excellent study demonstrates that even with a superb long standing clinical information system with good clinical decision support that this is not sufficient by itself to prevent medication errors. Human compliance requires repeated training and education as well as a system of direct observation to prevent development and utilization of nursing work processes that bypass established safety processes.