Elsevier

Academic Radiology

Original ivestigations

Investigation of misfiled cases in the PACS environment and a solution to prevent filing errors for chest radiographs1

Rationale and objective

The aim of the study was to survey misfiled cases in a picture archiving and communication system environment at two hospitals and to demonstrate the potential usefulness of an automated patient recognition method for posteroanterior chest radiographs based on a template-matching technique designed to prevent filing errors.

Materials and methods

We surveyed misfiled cases obtained from different modalities in one hospital for 25 months, and misfiled cases of chest radiographs in another hospital for 17 months. For investigating the usefulness of an automated patient recognition and identification method for chest radiographs, a prospective study has been completed in clinical settings at the latter hospital.

Results

The total numbers of misfiled cases for different modalities in one hospital and for chest radiographs in another hospital were 327 and 22, respectively. The misfiled cases in the two hospitals were mainly the result of human errors (eg, incorrect manual entries of patient information, incorrect usage of identification cards in which an identification card for the previous patient was used for the next patient's image acquisition). The prospective study indicated the usefulness of the computerized method for discovering misfiled cases with a high performance (ie, an 86.4% correct warning rate for different patients and 1.5% incorrect warning rate for the same patients).

Conclusion

We confirmed the occurrence of misfiled cases in the two hospitals. The automated patient recognition and identification method for chest radiographs would be useful in preventing wrong images from being stored in the picture archiving and communication system environment.

Section snippets

Survey of misfiled cases in image modalities in a hospital

Misfiled cases in various modalities such as radiography, computed tomography (CT), magnetic resonance imaging (MRI), fluoroscopy, angiography, ultrasound, scintigraphy, and radiography in operating rooms and in emergency rooms were surveyed at Osaka University Hospital. This hospital has digital image acquisition devices in most modalities except for mammography; these were connected with each other by intrahospital networks interfacing with the radiology information system (RIS; PC-RIS

Results

Table 1 indicates misfiled cases obtained from various modalities during 25 months at the Osaka University Hospital. The total number of misfiled cases found in this survey was 327 in 279,222 cases. The numbers of misfiled cases in the various modalities were 4–139. The majority of misfiled cases were observed in radiography, which constituted the largest number of examinations among the various modalities. The average rate of misfiled cases for radiography was 0.075% (139/186,521), and the

Discussion

We found that misfiled cases in the PACS environment occurred in the two hospitals. The rate of misfiled cases in the two hospitals was different because of the PACS system design and because of human mistakes. Our investigations were performed on a limited scale, but still indicated one of the important problems to be solved for practical use of the PACS in preventing medical accidents.

The main reason for misfiled cases in radiography at the Osaka University Hospital was due to incorrect entry

Acknowledgments

The authors are grateful to Mr. Seiichi Murakami and Mr. Keiji Fujimoto (Department of Radiology, University of Occupational and Environmental Health, Kita-Kyushyu, Japan) for their helpful discussions; Mr. Kondo (Komazawa University, Tokyo, Japan), Mr. Saki, and Mr. Yamamoto for their technical assistance; and Mrs. E. Lanzl for improving the manuscript.

Cited by (19)

  • Rating and Classification of Incident Reporting in Radiology in a Large Academic Medical Center

    2016, Current Problems in Diagnostic Radiology

    The Joint Commission reported on 1102 sentinel events because of wrong patient, wrong site, and wrong procedure from 2004-2014.23 Their analysis showed leadership, human factors, and communication as the most common root causes.23 Knowing how often these errors occur and what is the most probable root cause can help prevent future errors.

  • Automated patient identity recognition by analysis of chest radiograph features

    2013, Academic Radiology

    For some identities, however, Scoresimilarity values were low. Figure 9 shows the three major causes of Scoresimilarity values lower than 2.0: (1) insufficient inhalation, (2) abnormal gross opacity, and (3) incorrect identification of the top of the thoracic cage. A standard chest radiograph should be taken after a full inspiration (8) since insufficient inhalation results in inaccurately low measurements of lung length and thoracic cage width and can also produce a heart shadow below the diaphragm.

Arrow Up and Right View all citing articles on Scopus
View full text

Copyright © 2005 Association of University Radiologists. Published by Elsevier Inc. All rights reserved.