Reduction of Labeling Errors in Surgical Pathology by Barcoding.
Geoffrey A Talmon, Travis Rinehart, Milissa Gerken, Jason Gerriets, David Muirhead, Scott Smith, Audrey J Lazenby. University of Nebraska Medical Center, Omaha
Background: Most quality control (QC) literature in surgical pathology focuses on diagnostic error and interobserver reproducibility. Few have examined operational issues such as specimen mislabeling and of these, most focus on errors prior to receipt in pathology. Intradepartmental labeling errors (paraffin blocks, glass slides) are potential sources of specimen misidentification. Barcoding (BC) can reduce errors by switching from a human- to an electronic-driven system for labeling, reading and verification of specimen numbers. We examined the process changes in BC and compared labeling errors in the gross room and histology pre- and post-BC.
Design: Sequential cases for a 3 month period pre and post BC implementaion were used. Blocks and slides were examined for alteration of identification numbers (erasures, strikeouts, overwrites). Data collected included number of blocks, slides and cases with error. This reflected gross room or histology errors caught and corrected prior to delivery to pathologists. QC logs, incident reports and results of molecular identity testing were used to identify labeling errors that reach the pathologist. Work processes, including QC pre- and post-BC were also compared.
Results: Following BC, workflow changed from possible batch to obligate single part process. Also, manual/visual identification and QC were replaced by electronic labeling and verification. Material from 6,104 cases (21,067 blocks;38,264 slides) from the pre-BC period and 5,964 cases (23,531 blocks;38,946 slides) from the post-BC period were examined. Pre-BC period: 200 blocks (0.9%) and 809 slides (2.1%) were initially mislabeled, involving 81 (1.1%) and 331 cases (4.5%) respectively. Through multiple QC steps, each error was corrected before slide delivery. From QC and incident logs, 5 mislabeled cases were delivered to pathologists (<.001%). No cases were found to be discrepant by DNA testing. Post-BC period: no block labeling errors occurred (0%). QC and incident reports showed 3 cases with mislabeled slides reached pathologists (<.001%, due to histotech ignoring computer error message). No cases were discrepant by DNA testing due to block or slide mislabeling.
Conclusions: Labeling errors reaching pathologists were rare. Pre-BC, errors made in block and slide labeling were substantial (0.9%, 2.1%) but multiple QC steps corrected them prior to delivery to a pathologist. Post-BC, there was a dramatic reduction in block and slide labeling (0%, <.001%). BC technology switches from a manual/visual system with human QC designed to catch and correct errors to an electronic system that prevents errors.
Category: Quality Assurance
Tuesday, March 1, 2011 11:45 AM
Platform Session: Section H 2, Tuesday Morning