DataSifterText: Partially Synthetic Text Generation for Sensitive Clinical Notes
Authors:
- 16 November 2022
- journal article
- research article
- Published by Springer Nature in Journal of Medical Systems
- Vol. 46 (12) , 1-14
Abstract
Petabytes of health data are collected annually across the globe in electronic health records (EHR), including significant information stored as unstructured free text. However, the lack of effective mechanisms to securely share clinical text has inhibited its full utilization. We propose a new method, DataSifterText, to generate partially synthetic clinical free-text that can be safely shared between stakeholders (e.g., clinicians, STEM researchers, engineers, analysts, and healthcare providers), limiting the re-identification risk while providing significantly better utility preservation than suppressing or generalizing sensitive tokens. The method creates partially synthetic free-text data, which inherits the joint population distribution of the original data, and disguises the location of true and obfuscated words. Under certain obfuscation levels, the resulting synthetic text was sufficiently altered with different choices, orders, and frequencies of words compared to the original records. The differences were comparable to machine-generated (fully synthetic) text reported in previous studies. We applied DataSifterText to two medical case studies. In the CDC work injury application, using privacy protection, 60.9-86.5% of the synthetic descriptions belong to the same cluster as the original descriptions, demonstrating better utility preservation than the naïve content suppressing method (45.8-85.7%). In the MIMIC III application, the generated synthetic data maintained over 80% of the original information regarding patients’ overall health conditions. The reported DataSifterText statistical obfuscation results indicate that the technique provides sufficient privacy protection (low identification risk) while preserving population-level information (high utility).Keywords
Funding Information
- NIH (P20 NR015331, UL1TR002240, R01CA233487, R01MH12107, T32GM141746)
- NSF (1916425, 1734853, 1636840)
This publication has 17 references indexed in Scilit:
- Development and Validation of a Rubric to Evaluate Diabetes SOAP Note Writing in APPEAmerican Journal of Pharmaceutical Education, 2018
- Medical subdomain classification of clinical notes using a machine learning-based natural language processing approachBMC Medical Informatics and Decision Making, 2017
- De-identification of patient notes with recurrent neural networksJournal of the American Medical Informatics Association, 2016
- MIMIC-III, a freely accessible critical care databaseScientific Data, 2016
- A hybrid machine learning model for multi-document summarizationApplied Intelligence, 2013
- Updating and Validating the Charlson Comorbidity Index and Score for Risk Adjustment in Hospital Discharge Abstracts Using Data From 6 CountriesAmerican Journal of Epidemiology, 2011
- Privacy-preserving data publishingACM Computing Surveys, 2010
- Interpreting TF-IDF term weights as making relevance decisionsACM Transactions on Information Systems, 2008
- Relating patient characteristics at the time of admission to outcomes of hospitalizationJournal of Clinical Epidemiology, 1991
- Term-weighting approaches in automatic text retrievalInformation Processing & Management, 1988