Medication Errors Involving Patient-controlled Analgesia

Rodney W. Hicks; Vanja Sikirica; Winnie Nelson; Jeff R. Schein; Diane D. Cousins


Am J Health Syst Pharm. 2008;65(5):429-440. 

In This Article


USP performed a free-text search of all Medmarx records submitted between July 1, 2000, and June 30, 2005, using key PCA-related search terms (e.g., "PCA," "patient controlled") from the error-description field (a free-text, narrative field that may contain up to 3000 characters). Wild-card searches (both incorporating mixed-case lettering and ignoring case and spacing) allowed for maximum inclusivity. Since the study used anonymous data voluntarily submitted to Medmarx, it was exempt from full human-subject review. All the textual error descriptions were read to determine the context of the use of "PCA." USP staff (three nurses, two pharmacists, and one analyst) participated in reading each unique record. These personnel identified 99 records in which "PCA" was felt to imply something other than patient-controlled analgesia. The investigators then rereviewed these 99 records to determine if they should be excluded. All remaining records were divided into one-year groups; a year was defined as July 1 through June 30.

Quantitative analysis of the records included the severity of each error (classified according to the National Coordinating Council for Medication Error Reporting and Prevention [NCCMERP] index for categorizing medication errors), the type of error, the node (the phase of the medication-use process where the error originated), the principal cause, the contributing factors, the actions taken following the error, the level of care rendered as a result of the error, the drugs involved in the error, the physical location of the error, the staff involved in the initial error, and (when known) the sex and age of the patient involved. Facility characteristics evaluated included bed capacity and the owner or operator classification.

To examine the causes of the errors, a derivative of NCCMERP's taxonomy and the cause-of-error selection list from Medmarx was used. The NCCMERP taxonomy has been endorsed by more than 25 constituent organizations and has been used in error-reporting research by more than 400 organizations. Causes were classified as communication issues; drug name confusion; storage, packaging, and labeling matters; human factors; system errors; contraindications; equipment related; or reconciliation related or were reverted to one of the default values in Medmarx (computerized prescriber order entry, wrong diluent, drug shortage, and nonformulary drug).

To examine the level of care rendered as a result of the errors, Medmarx's 19-item selection list was reduced to eight groups of related items. The selections were further examined through a cross-tabulation analysis of nonharmful and harmful errors. Product information was explored by generic name and intended route of administration. Staff members identified as having made the error were examined by general groupings, namely anesthesia providers, prescribing personnel, dispensing personnel, administering personnel, and other. The physical location of an error was reported as perioperative, maternity, critical care, pharmacy, nursing unit, or other. Patient demographics were explored through a cross-tabulation analysis that examined patients' ages (grouped by decade) and error severity.

A qualitative review of each PCA error's description by the USP staff was performed to obtain an in-depth understanding of the details surrounding each error. To develop themes from the data, the records were grouped by medication-use node and then sorted by error severity and type.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.