De-identification of patient notes with recurrent neural networks

By Franck Dernoncourt, Ji Young Lee, Ozlem Uzuner, Peter Szolovits. 2016

433 citations recorded by the source corpus.

Objective Patient notes in electronic health records (EHRs) may contain critical information for medical investigations. However, the vast majority of medical investigators can only access de-identified notes, in order to protect the confidentiality of patients. In the United States, the Health Insurance Portability and Accountability Act (HIPAA) defines 18 types of protected health information that needs to be removed to de-identify patient notes. Manual de-identification is impractical given the size of electronic health record databases, the limited number of researchers with access to non-de-identified notes, and the frequent mistakes of human annotators. A reliable automated de-identification system would consequently be of high value. Materials and Methods We introduce the first de-identification system based on artificial neural networks (ANNs), which requires no handcrafted features or rules, unlike existing systems. We compare the performance of the system with state-of-the-art systems on two datasets: the i2b2 2014 de-identification challenge dataset, which is the largest publicly available de-identification dataset, and the MIMIC de-identification dataset, which we assembled and is twice as large as the i2b2 2014 dataset. Results Our ANN model outperforms the state-of-the-art systems. It yields an F1-score of 97.85 on the i2b2 2014 dataset, with a recall of 97.38 and a precision of 98.32, and an F1-score of 99.23 on the MIMIC de-identification dataset, with a recall of 99.25 and a precision of 99.21. Conclusion Our findings support the use of ANNs for de-identification of patient notes, as they show better performance than previously published systems while requiring no manual feature engineering.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Health Information Management

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Health Information Management

Citations in the corpus

Recurrent Neural Networks for Classifying Relations in Clinical Notes (2017, positive): “Recurrent Neural Networks (RNN) are designed to capture sequential patterns present in data and have been applied to longitudinal data (temporal sequence) [39], image data (spatial sequence) [49], and text data [44] in medical domain.”
What Does it Mean for a Language Model to Preserve Privacy? (2022, background): “Natural language processing algorithms that aim to respect privacy either remove private information from the data (through text sanitization [3, 19, 57, 73]), or design learning algorithms that mitigate the risks of information leakage by not memorizing private information (through differentially p”
Comparing deep learning and concept extraction based methods for patient phenotyping from clinical narratives (2018, background): “Applications of deep learning for other tasks in healthcare have shown promising results; examples include mortality prediction [22], patient note de-identification [23], skin cancer detection [24], and diabetic retinopathy detection [25].”
Deep learning for healthcare: review, opportunities and challenges (2018, background): “Consequently, the authors used the LSTM RNN model and identified three interesting subtypes for Parkinson’s disease, wherein each subtype demonstrates common disease progression trends.”
Knowledge Unlearning for Mitigating Privacy Risks in Language Models (2022, methodology): “This is effective at identifying well-formatted private information such as social security numbers or special forms of medical notes (Aura et al., 2006; Dernoncourt et al., 2017; Lison et al., 2021; Kandpal et al., 2022).”
Deep EHR: A Survey of Recent Advances in Deep Learning Techniques for Electronic Health Record (EHR) Analysis (2017, background): “[43], [45] [38], [44], [54] [52], [53], [60], [61] [13], [14]”