Joanne G. Kowiatek

A comparison of voluntarily reported medication errors in intensive care and general care units

Background Few institutions currently track intensive care unit (ICU)-specific medication safety data. A comparison of medication error data for intensive care and general care units may determine if ICU-specific surveillance is needed. Objective To compare the type, cause, contributing factors, level of staff initiating an error, medication use process node, drug classes and patient outcomes for voluntarily reported medication errors occurring in ICUs and general care units. Design Retrospective evaluation of voluntarily reported medication errors over 4.5 years at a 647-bed academic medical centre containing greater than 120 ICU beds. Adult patients with a reported medication error in intensive care and general care units were included. Medication error data were compared for ICUs with general care units. Main measures and results There were a total of 3252 medication errors reported with 541 and 2711 occurring in ICUs and general care units, respectively. Primary types of medication errors were prescribing in the ICUs and omission in the general care units. Leading causes of medication errors were procedure/protocol not followed and knowledge deficit in the ICU and general care units. More frequently there was no contributing factor identified for medication errors in the ICUs. The top three drugs associated with medication errors in the ICUs were opioid analgesics, β-lactam antimicrobials and blood coagulation modifiers compared with anti-asthma/bronchodilators, narcotic analgesics and vaccines in the general care units. The level of care provided after the error was observation increased/initiated in ICUs and no additional care in general care units. Prolonged hospitalisation was a result of medication errors in 1% of ICU cases and 0.4% of general care unit errors (p = 0.056). Medication errors were associated with harm in 12% and 6% of cases in the ICUs and general care units, respectively (p<0.001). Conclusion Type, contributing factors, drug classes and patient outcomes associated with voluntarily reported medication errors differ in intensive care and general care units so it is important to develop surveillance systems that analyse ICU-specific data allowing systematic changes for this patient population.

Medication errors resulting from computer entry by nonprescribers

PURPOSE The characteristics of medication errors associated with the use of computer order-entry systems by nonprescribers are discussed. METHODS A retrospective analysis of records submitted to MEDMARX was conducted for the period from July 1, 2001, to December 31, 2005, to identify all computer-related medication errors made by nonprescribers. Quantitative analysis of the records included the severity of each error, the origin within the medication-use process, the type of error, principal causes, the location within the facility where the error was made, and the therapeutic drug classes frequently involved. Similar data from the University of Pittsburgh Medical Center (UPMC) were also analyzed and compared with the national data set. RESULTS During the 4.5 years, 693 unique facilities submitted 90,001 medication error records that were the result of computer entry by nonprescribers. The national data set and the UPMC data had similar findings for error severity, error origin, and type of error but showed some differences in the rank ordering of error causes, location where the error occurred, and drug classes frequently associated with such errors. The percentage of harm associated with computer-entry errors was small for both the national data set and UPMC data (0.99% and 0.80%, respectively). Both data sets cited performance deficit as the leading cause of computer-entry errors, but large percentage differences were seen with other causes, including inaccurate or omitted transcription (30% versus 12.6%, respectively), documentation (19.5% versus 10.6%, respectively), and procedure or protocol not followed (21.7% versus 30.3%, respectively). Both data sets implicated the inpatient pharmacy department as the location where most computer-entry errors occurred (49.3% versus 69.0%, respectively). CONCLUSION Analysis of the characteristics of medication errors associated with the use of computer-entry systems by non-prescribers from both MEDMARX and an individual health system database demonstrated that computer systems create new opportunities for errors to occur. Working closely with information technology personnel dedicated to assisting pharmacy departments and vendors, adequate training of pharmacy staff, and development of national standards for drug information displays in computer order-entry systems may help minimize such errors.

Insulin Medication Error Reduction: A Quality Improvement Initiative

This study describes a multidisciplinary approach to reducing insulin medication errors. In 1998, a medication error analysis at our institution demonstrated that insulin was the top drug associated with medication errors. A collaborative continuous quality improvement initiative was undertaken to develop recommendations for reducing insulin medication errors. Multidisciplinary root cause analysis revealed that during order prescribing and transcribing, the abbreviation of “u” for units—combined with the lack of a formal education program for physicians on appropriate use of abbreviations—was a major contributing cause to insulin medication errors. An extensive multidisciplinary education program was undertaken to target insulin orders for medication error reduction. To date, compliance with prescribing insulin orders within guidelines has increased from 30% to 74%, and compliance with transcribing insulin orders within guidelines has increased from 36% to 83%. This collaborative project increased awareness of insulin medication error causation and will be used as a template to reduce other types of medication errors.

Assessing and Monitoring Override Medications in Automated Dispensing Devices

BACKGROUND Medication override is not without risk because the absence of pharmacist review may increase the potential for a medication error. The University of Pittsburgh Medical Centers Department of Pharmacy and Therapeutics assessed the safety of the automated dispensing device (ADD) override process to reduce the number of override medications stored in the ADD. METHODS A representative expert panel developed criteria for override access and revised the override medication list; an override monitoring tool was used to perform random audits and determine nursing compliance; and changes in override practices for use of opioids, a high-alert medication class, were measured. RESULTS The expert panel reduced the number of medications and dosage forms on the override list by 42%, from 119 different medications (in 244 different dosage forms) in 2001 to 92 different medications (in 163 different dosage forms) by December 2003. By May 2004 the opioid override rates were significantly decreased; although slight increases in rates occurred by December 2004, possibly reflecting the lack of formal override monitoring by nursing and pharmacy, the rates were still significantly lower than the baseline in October 2003. CONCLUSION Pharmacists, in collaboration with the medical and nursing staffs, developed a sustainable process for preventing unauthorized and inappropriate override medication dispensing from ADDs.

Equipment, Medications, and Supplies for a Medical Emergency Team Response

To mount an effective emergency response, medication and equipment resources must be available, reliable, and organized in a way to make them easily usable. Staff must be trained adequately so that they know what their resources are and how to manage them. Standardizing the equipment and medications contributes to a safe system by improving a number of logistic issues, including staff training, performance, error reduction, equipment maintenance and replacement after a crisis, and finally the institution’s ability to revise medication and equipment resources for crises. We believe that improving efficiency and reliability can reduce delays and errors, and contribute to the primary goal of improving patient outcomes following a crisis event.