Guido Zuccon1, Daniel Kotzur2, Anthony Nguyen2, Anton Bergheim3. 1. The Australian e-Health Research Centre (Commonwealth Scientific and Industrial Research Organisation), Level 5 - UQ Health Sciences Building 901/16, Royal Brisbane and Women's Hospital, Herston, QLD 4029, Australia; School of Information Systems, Queensland University of Technology, Y Block Level 6, Gardens Point Campus, Brisbane, QLD, Australia. Electronic address: g.zuccon@qut.edu.au. 2. The Australian e-Health Research Centre (Commonwealth Scientific and Industrial Research Organisation), Level 5 - UQ Health Sciences Building 901/16, Royal Brisbane and Women's Hospital, Herston, QLD 4029, Australia. 3. Cancer Institute NSW, Australian Technology Park, Level 9, 8 Central Avenue, Eveleigh, NSW 2015, Australia.
Abstract
OBJECTIVE: Evaluate the effectiveness and robustness of Anonym, a tool for de-identifying free-text health records based on conditional random fields classifiers informed by linguistic and lexical features, as well as features extracted by pattern matching techniques. De-identification of personal health information in electronic health records is essential for the sharing and secondary usage of clinical data. De-identification tools that adapt to different sources of clinical data are attractive as they would require minimal intervention to guarantee high effectiveness. METHODS AND MATERIALS: The effectiveness and robustness of Anonym are evaluated across multiple datasets, including the widely adopted Integrating Biology and the Bedside (i2b2) dataset, used for evaluation in a de-identification challenge. The datasets used here vary in type of health records, source of data, and their quality, with one of the datasets containing optical character recognition errors. RESULTS: Anonym identifies and removes up to 96.6% of personal health identifiers (recall) with a precision of up to 98.2% on the i2b2 dataset, outperforming the best system proposed in the i2b2 challenge. The effectiveness of Anonym across datasets is found to depend on the amount of information available for training. CONCLUSION: Findings show that Anonym compares to the best approach from the 2006 i2b2 shared task. It is easy to retrain Anonym with new datasets; if retrained, the system is robust to variations of training size, data type and quality in presence of sufficient training data. Crown
OBJECTIVE: Evaluate the effectiveness and robustness of Anonym, a tool for de-identifying free-text health records based on conditional random fields classifiers informed by linguistic and lexical features, as well as features extracted by pattern matching techniques. De-identification of personal health information in electronic health records is essential for the sharing and secondary usage of clinical data. De-identification tools that adapt to different sources of clinical data are attractive as they would require minimal intervention to guarantee high effectiveness. METHODS AND MATERIALS: The effectiveness and robustness of Anonym are evaluated across multiple datasets, including the widely adopted Integrating Biology and the Bedside (i2b2) dataset, used for evaluation in a de-identification challenge. The datasets used here vary in type of health records, source of data, and their quality, with one of the datasets containing optical character recognition errors. RESULTS: Anonym identifies and removes up to 96.6% of personal health identifiers (recall) with a precision of up to 98.2% on the i2b2 dataset, outperforming the best system proposed in the i2b2 challenge. The effectiveness of Anonym across datasets is found to depend on the amount of information available for training. CONCLUSION: Findings show that Anonym compares to the best approach from the 2006 i2b2 shared task. It is easy to retrain Anonym with new datasets; if retrained, the system is robust to variations of training size, data type and quality in presence of sufficient training data. Crown
Authors: Anthony Nguyen; John O'Dwyer; Thanh Vu; Penelope M Webb; Sharon E Johnatty; Amanda B Spurdle Journal: BMJ Open Date: 2020-06-11 Impact factor: 2.692