Deep Learning Framework for Advanced De-Identification of Protected Health Information
Electronic health records (EHRs) are widely used in healthcare institutions worldwide, containing vast amounts of unstructured textual data. However, the sensitive nature of Protected Health Information (PHI) embedded within these records presents significant privacy challenges, necessitating robust...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Future Internet |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-5903/17/1/47 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832588442792361984 |
|---|---|
| author | Ahmad Aloqaily Emad E. Abdallah Rahaf Al-Zyoud Esraa Abu Elsoud Malak Al-Hassan Alaa E. Abdallah |
| author_facet | Ahmad Aloqaily Emad E. Abdallah Rahaf Al-Zyoud Esraa Abu Elsoud Malak Al-Hassan Alaa E. Abdallah |
| author_sort | Ahmad Aloqaily |
| collection | DOAJ |
| description | Electronic health records (EHRs) are widely used in healthcare institutions worldwide, containing vast amounts of unstructured textual data. However, the sensitive nature of Protected Health Information (PHI) embedded within these records presents significant privacy challenges, necessitating robust de-identification techniques. This paper introduces a novel approach, leveraging a Bi-LSTM-CRF model to achieve accurate and reliable PHI de-identification, using the i2b2 dataset sourced from Harvard University. Unlike prior studies that often unify Bi-LSTM and CRF layers, our approach focuses on the individual design, optimization, and hyperparameter tuning of both the Bi-LSTM and CRF components, allowing for precise model performance improvements. This rigorous approach to architectural design and hyperparameter tuning, often underexplored in the existing literature, significantly enhances the model’s capacity for accurate PHI tag detection while preserving the essential clinical context. Comprehensive evaluations are conducted across 23 PHI categories, as defined by HIPAA, ensuring thorough security across critical domains. The optimized model achieves exceptional performance metrics, with a precision of 99%, recall of 98%, and F1-score of 98%, underscoring its effectiveness in balancing recall and precision. By enabling the de-identification of medical records, this research strengthens patient confidentiality, promotes compliance with privacy regulations, and facilitates safe data sharing for research and analysis. |
| format | Article |
| id | doaj-art-c4827a27d0f1424285d6d4f7f9e18450 |
| institution | Kabale University |
| issn | 1999-5903 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Future Internet |
| spelling | doaj-art-c4827a27d0f1424285d6d4f7f9e184502025-01-24T13:33:41ZengMDPI AGFuture Internet1999-59032025-01-011714710.3390/fi17010047Deep Learning Framework for Advanced De-Identification of Protected Health InformationAhmad Aloqaily0Emad E. Abdallah1Rahaf Al-Zyoud2Esraa Abu Elsoud3Malak Al-Hassan4Alaa E. Abdallah5Department of Information Technology, Faculty of Prince Al-Hussein Bin Abdullah II for Information Technology, The Hashemite University, P.O. Box 330127, Zarqa 13133, JordanDepartment of Information Technology, Faculty of Prince Al-Hussein Bin Abdullah II for Information Technology, The Hashemite University, P.O. Box 330127, Zarqa 13133, JordanDepartment of Information Technology, Faculty of Prince Al-Hussein Bin Abdullah II for Information Technology, The Hashemite University, P.O. Box 330127, Zarqa 13133, JordanDepartment of Computer Science, Faculty of Information Technology, Zarqa University, P.O. Box 330127, Zarqa 13133, JordanKing Abdullah II School of Information Technology, The University of Jordan, Amman 11942, JordanDepartment of Computer Science, Faculty of Prince Al-Hussein Bin Abdullah II for Information Technology, The Hashemite University, P.O. Box 330127, Zarqa 13133, JordanElectronic health records (EHRs) are widely used in healthcare institutions worldwide, containing vast amounts of unstructured textual data. However, the sensitive nature of Protected Health Information (PHI) embedded within these records presents significant privacy challenges, necessitating robust de-identification techniques. This paper introduces a novel approach, leveraging a Bi-LSTM-CRF model to achieve accurate and reliable PHI de-identification, using the i2b2 dataset sourced from Harvard University. Unlike prior studies that often unify Bi-LSTM and CRF layers, our approach focuses on the individual design, optimization, and hyperparameter tuning of both the Bi-LSTM and CRF components, allowing for precise model performance improvements. This rigorous approach to architectural design and hyperparameter tuning, often underexplored in the existing literature, significantly enhances the model’s capacity for accurate PHI tag detection while preserving the essential clinical context. Comprehensive evaluations are conducted across 23 PHI categories, as defined by HIPAA, ensuring thorough security across critical domains. The optimized model achieves exceptional performance metrics, with a precision of 99%, recall of 98%, and F1-score of 98%, underscoring its effectiveness in balancing recall and precision. By enabling the de-identification of medical records, this research strengthens patient confidentiality, promotes compliance with privacy regulations, and facilitates safe data sharing for research and analysis.https://www.mdpi.com/1999-5903/17/1/47protected health informationelectronic health recorddeep learningde-identificationBi-LSTM-CRF |
| spellingShingle | Ahmad Aloqaily Emad E. Abdallah Rahaf Al-Zyoud Esraa Abu Elsoud Malak Al-Hassan Alaa E. Abdallah Deep Learning Framework for Advanced De-Identification of Protected Health Information Future Internet protected health information electronic health record deep learning de-identification Bi-LSTM-CRF |
| title | Deep Learning Framework for Advanced De-Identification of Protected Health Information |
| title_full | Deep Learning Framework for Advanced De-Identification of Protected Health Information |
| title_fullStr | Deep Learning Framework for Advanced De-Identification of Protected Health Information |
| title_full_unstemmed | Deep Learning Framework for Advanced De-Identification of Protected Health Information |
| title_short | Deep Learning Framework for Advanced De-Identification of Protected Health Information |
| title_sort | deep learning framework for advanced de identification of protected health information |
| topic | protected health information electronic health record deep learning de-identification Bi-LSTM-CRF |
| url | https://www.mdpi.com/1999-5903/17/1/47 |
| work_keys_str_mv | AT ahmadaloqaily deeplearningframeworkforadvanceddeidentificationofprotectedhealthinformation AT emadeabdallah deeplearningframeworkforadvanceddeidentificationofprotectedhealthinformation AT rahafalzyoud deeplearningframeworkforadvanceddeidentificationofprotectedhealthinformation AT esraaabuelsoud deeplearningframeworkforadvanceddeidentificationofprotectedhealthinformation AT malakalhassan deeplearningframeworkforadvanceddeidentificationofprotectedhealthinformation AT alaaeabdallah deeplearningframeworkforadvanceddeidentificationofprotectedhealthinformation |