Dataset Information

De-identification of electronic health record using neural network.

ABSTRACT: According to a recent study, around 99% of hospitals across the US now use electronic health record systems (EHRs). One of the most common types of EHR is the unstructured textual data, and unlocking hidden details from this data is critical for improving current medical practices and research endeavors. However, these textual data contain sensitive information, which could compromise our privacy. Therefore, medical textual data cannot be released publicly without undergoing any privacy-protective measures. De-identification is a process of detecting and removing all sensitive information present in EHRs, and it is a necessary step towards privacy-preserving EHR data sharing. Over the last decade, there have been several proposals to de-identify textual data using manual, rule-based, and machine learning methods. In this article, we propose new methods to de-identify textual data based on the self-attention mechanism and stacked Recurrent Neural Network. To the best of our knowledge, we are the first to employ these techniques. Experimental results on three different datasets show that our model performs better than all state-of-the-art mechanism irrespective of the dataset. Additionally, our proposed method is significantly faster than the existing techniques. Finally, we introduced three utility metrics to judge the quality of the de-identified data.

SUBMITTER: Ahmed T

PROVIDER: S-EPMC7596089 | biostudies-literature | 2020 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

De-identification of electronic health record using neural network.

Ahmed Tanbir T Aziz Md Momin Al MMA Mohammed Noman N

Scientific reports 20201029 1

According to a recent study, around 99% of hospitals across the US now use electronic health record systems (EHRs). One of the most common types of EHR is the unstructured textual data, and unlocking hidden details from this data is critical for improving current medical practices and research endeavors. However, these textual data contain sensitive information, which could compromise our privacy. Therefore, medical textual data cannot be released publicly without undergoing any privacy-protecti ...[more]

PMID: 33122735

Similar Datasets

Project description:ObjectiveTo demonstrate the use of aggregated and de-identified electronic health record (EHR) data for multivariate post-marketing pharmacosurveillance in a case study of azathioprine (AZA).MethodsUsing aggregated, standardized, normalized, and de-identified, population-level data from the Explore platform (Explorys, Inc.) we searched over 10 million individuals, of which 14,580 were prescribed AZA based on RxNorm drug orders. Based on logical observation identifiers names and codes (LOINC) and vital sign data, we examined the following side effects: anemia, cell lysis, fever, hepatotoxicity, hypertension, nephrotoxicity, neutropenia, and neutrophilia. Patients prescribed AZA were compared to patients prescribed one of 11 other anti-rheumatologic drugs to determine the relative risk of side effect pairs.ResultsCompared to AZA case report trends, hepatotoxicity (marked by elevated transaminases or elevated bilirubin) did not occur as an isolated event more frequently in patients prescribed AZA than other anti-rheumatic agents. While neutropenia occurred in 24% of patients (RR 1.15, 95% CI 1.07-1.23), neutrophilia was also frequent (45%) and increased in patients prescribed AZA (RR 1.28, 95% CI 1.22-1.34). After constructing a pairwise side effect network, neutropenia had no dependencies. A reduced risk of neutropenia was found in patients with co-existing elevations in total bilirubin or liver transaminases, supporting classic clinical knowledge that agranulocytosis is a largely unpredictable phenomenon. Rounding errors propagated in the statistically de-identified datasets for cohorts as small as 40 patients only contributed marginally to the calculated risk.ConclusionOur work demonstrates that aggregated, standardized, normalized and de-identified population level EHR data can provide both sufficient insight and statistical power to detect potential patterns of medication side effect associations, serving as a multivariate and generalizable approach to post-marketing drug surveillance.

Project description:BackgroundCollaboration is vital within health care institutions, and it allows for the effective use of collective health care worker (HCW) expertise. Human-computer interactions involving electronic health records (EHRs) have become pervasive and act as an avenue for quantifying these collaborations using statistical and network analysis methods.ObjectiveWe aimed to measure HCW collaboration and its characteristics by analyzing concurrent EHR usage.MethodsBy extracting concurrent EHR usage events from audit log data, we defined concurrent sessions. For each HCW, we established a metric called concurrent intensity, which was the proportion of EHR activities in concurrent sessions over all EHR activities. Statistical models were used to test the differences in the concurrent intensity between HCWs. For each patient visit, starting from admission to discharge, we measured concurrent EHR usage across all HCWs, which we called temporal patterns. Again, we applied statistical models to test the differences in temporal patterns of the admission, discharge, and intermediate days of hospital stay between weekdays and weekends. Network analysis was leveraged to measure collaborative relationships among HCWs. We surveyed experts to determine if they could distinguish collaborative relationships between high and low likelihood categories derived from concurrent EHR usage. Clustering was used to aggregate concurrent activities to describe concurrent sessions. We gathered 4 months of EHR audit log data from a large academic medical center's neonatal intensive care unit (NICU) to validate the effectiveness of our framework.ResultsThere was a significant difference (P<.001) in the concurrent intensity (proportion of concurrent activities: ranging from mean 0.07, 95% CI 0.06-0.08, to mean 0.36, 95% CI 0.18-0.54; proportion of time spent on concurrent activities: ranging from mean 0.32, 95% CI 0.20-0.44, to mean 0.76, 95% CI 0.51-1.00) between the top 13 HCW specialties who had the largest amount of time spent in EHRs. Temporal patterns between weekday and weekend periods were significantly different on admission (number of concurrent intervals per hour: 11.60 vs 0.54; P<.001) and discharge days (4.72 vs 1.54; P<.001), but not during intermediate days of hospital stay. Neonatal nurses, fellows, frontline providers, neonatologists, consultants, respiratory therapists, and ancillary and support staff had collaborative relationships. NICU professionals could distinguish high likelihood collaborative relationships from low ones at significant rates (3.54, 95% CI 3.31-4.37 vs 2.64, 95% CI 2.46-3.29; P<.001). We identified 50 clusters of concurrent activities. Over 87% of concurrent sessions could be described by a single cluster, with the remaining 13% of sessions comprising multiple clusters.ConclusionsLeveraging concurrent EHR usage workflow through audit logs to analyze HCW collaboration may improve our understanding of collaborative patient care. HCW collaboration using EHRs could potentially influence the quality of patient care, discharge timeliness, and clinician workload, stress, or burnout.

Project description:AimsUnderstanding atypical forms of diabetes (AD) may advance precision medicine, but methods to identify such patients are needed. We propose an electronic health record (EHR)-based algorithmic approach to identify patients who may have AD, specifically those with insulin-sufficient, non-metabolic diabetes, in order to improve feasibility of identifying these patients through detailed chart review.MethodsPatients with likely T2D were selected using a validated machine-learning (ML) algorithm applied to EHR data. "Typical" T2D cases were removed by excluding individuals with obesity, evidence of dyslipidemia, antibody-positive diabetes, or cystic fibrosis. To filter out likely type 1 diabetes (T1D) cases, we applied six additional "branch algorithms," relying on various clinical characteristics, which resulted in six overlapping cohorts. Diabetes type was classified by manual chart review as atypical, not atypical, or indeterminate due to missing information.ResultsOf 114,975 biobank participants, the algorithms collectively identified 119 (0.1%) potential AD cases, of which 16 (0.014%) were confirmed after expert review. The branch algorithm that excluded T1D based on outpatient insulin use had the highest percentage yield of AD (13 of 27; 48.2% yield). Together, the 16 AD cases had significantly lower BMI and higher HDL than either unselected T1D or T2D cases identified by ML algorithms (P<0.05). Compared to the ML T1D group, the AD group had a significantly higher T2D polygenic score (P<0.01) and lower hemoglobin A1c (P<0.01).ConclusionOur EHR-based algorithms followed by manual chart review identified collectively 16 individuals with AD, representing 0.22% of biobank enrollees with T2D. With a maximum yield of 48% cases after manual chart review, our algorithms have the potential to drastically improve efficiency of AD identification. Recognizing patients with AD may inform on the heterogeneity of T2D and facilitate enrollment in studies like the Rare and Atypical Diabetes Network (RADIANT).

Dataset Information

De-identification of electronic health record using neural network.

Publications

De-identification of electronic health record using neural network.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets