Removing Patient Identifiers from Medical Notes and Data
The confidentiality of patient information is an essential part of the Electronic Health Record System. Patient information, if exposed, can cause serious damage to the privacy of individuals receiving healthcare.
This paper proposes a deep learning model—using a deconvolutional neural network, bi-directional LSTM-CNN, and regular expressions—to recognize and remove individually identifiable information from physician notes. This information is then removed from a medical practitioner’s data, further allowing the fair use of anonymous data among researchers and in clinical trials.