@article{6800, author = {Raymond H. Thornton and Jeremy Miransky and Aileen Killen and Stephen B. Solomon and Lynn A. Brody}, title = {Analysis and prioritization of near-miss adverse events in a radiology department.}, abstract = {

OBJECTIVE: The purpose of this study is to describe a method for the evaluation and prioritization of near-miss events in a radiology department.

MATERIALS AND METHODS: Sixty-two consecutive near-miss events occurring between 2007 and 2009 were retrospectively evaluated, classified by error type, and scored for five elements associated with risk. The worst outcome potentially associated with each event was predicted by consensus and scored on a standardized 5-point complications grading scale. Scores were then assigned for event frequency, method of detection, barrier number, and quality. The product of individual scores, ranging from 1 to 180, was termed the hazard score. Events were analyzed by error type, element scores, and hazard score.

RESULTS: Electronic order entry errors were the most common error type, and 90% of these errors originated outside the radiology department. More than half (65%) of the events were assigned maximal severity scores, and 68% of the errors had been encountered three or more times previously. Twenty-five events (40%) were detected by good fortune rather than by plan. No barrier to the projected worst outcome was identified in nearly half (47%) of cases. In most instances (73%), strong barriers were absent. Nine events (15%) had maximal hazard scores of 180, whereas 21 events (34%) had hazard scores of 30 or less.

CONCLUSION: This method was constructed from standardized definitions of outcome severity, the ability of current systems to detect or mitigate an adverse event or outcome, and event frequency and offers a tool for systematic evaluation and stratification of near-miss adverse events.

}, year = {2011}, journal = {AJR Am J Roentgenol}, volume = {196}, pages = {1120-4}, month = {05/2011}, issn = {1546-3141}, doi = {10.2214/AJR.10.5373}, language = {eng}, }