Dataset Information

Learning to identify Protected Health Information by integrating knowledge- and data-driven algorithms: A case study on psychiatric evaluation notes.

ABSTRACT: De-identification of clinical narratives is one of the main obstacles to making healthcare free text available for research. In this paper we describe our experience in expanding and tailoring two existing tools as part of the 2016 CEGS N-GRID Shared Tasks Track 1, which evaluated de-identification methods on a set of psychiatric evaluation notes for up to 25 different types of Protected Health Information (PHI). The methods we used rely on machine learning on either a large or small feature space, with additional strategies, including two-pass tagging and multi-class models, which both proved to be beneficial. The results show that the integration of the proposed methods can identify Health Information Portability and Accountability Act (HIPAA) defined PHIs with overall F1-scores of ?90% and above. Yet, some classes (Profession, Organization) proved again to be challenging given the variability of expressions used to reference given information.

SUBMITTER: Dehghan A

PROVIDER: S-EPMC5705401 | biostudies-literature | 2017 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Learning to identify Protected Health Information by integrating knowledge- and data-driven algorithms: A case study on psychiatric evaluation notes.

Dehghan Azad A Kovacevic Aleksandar A Karystianis George G Keane John A JA Nenadic Goran G

Journal of biomedical informatics 20170607

De-identification of clinical narratives is one of the main obstacles to making healthcare free text available for research. In this paper we describe our experience in expanding and tailoring two existing tools as part of the 2016 CEGS N-GRID Shared Tasks Track 1, which evaluated de-identification methods on a set of psychiatric evaluation notes for up to 25 different types of Protected Health Information (PHI). The methods we used rely on machine learning on either a large or small feature spa ...[more]

PMID: 28602908

Similar Datasets

Project description:PurposeCardiovascular disease (CVD) is a major worldwide health burden. As the risk factors of CVD, hypertension, and hyperlipidemia are most mentioned. Early stage hypertension in the population with dyslipidemia is an important public health hazard. This study was the application of data-driven machine learning (ML), demonstrating complex relationships between risk factors and outcomes and promising predictive performance with vast amounts of medical data, aimed to investigate the association between dyslipidemia and the incidence of early stage hypertension in a large cohort with normal blood pressure at baseline.MethodsThis study analyzed annual health screening data for 71,108 people from 2005 to 2017, including data for 27 risk-related indicators, sourced from the MJ Group, a major health screening center in Taiwan. We used five machine learning (ML) methods-stochastic gradient boosting (SGB), multivariate adaptive regression splines (MARS), least absolute shrinkage and selection operator regression (Lasso), ridge regression (Ridge), and gradient boosting with categorical features support (CatBoost)-to develop a multi-stage ML algorithm-based prediction scheme and then evaluate important risk factors at the early stage of hypertension, especially for groups with high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) levels within or out of the reference range.ResultsAge, body mass index, waist circumference, waist-to-hip ratio, fasting plasma glucose, and C-reactive protein (CRP) were associated with hypertension. The hemoglobin level was also a positive contributor to blood pressure elevation and it appeared among the top three important risk factors in all LDL-C/HDL-C groups; therefore, these variables may be important in affecting blood pressure in the early stage of hypertension. A residual contribution to blood pressure elevation was found in groups with increased LDL-C. This suggests that LDL-C levels are associated with CPR levels, and that the LDL-C level may be an important factor for predicting the development of hypertension.ConclusionThe five prediction models provided similar classifications of risk factors. The results of this study show that an increase in LDL-C is more important than the start of a drop in HDL-C in health screening of sub-healthy adults. The findings of this study should be of value to health awareness raising about hypertension and further discussion and follow-up research.

Project description:BackgroundCluster randomized trials (CRTs) are becoming an increasingly important design. However, authors of CRTs do not always adhere to requirements to explicitly identify the design as cluster randomized in titles and abstracts, making retrieval from bibliographic databases difficult. Machine learning algorithms may improve their identification and retrieval. Therefore, we aimed to develop machine learning algorithms that accurately determine whether a bibliographic citation is a CRT report.MethodsWe trained, internally validated, and externally validated two convolutional neural networks and one support vector machine (SVM) algorithm to predict whether a citation is a CRT report or not. We exclusively used the information in an article citation, including the title, abstract, keywords, and subject headings. The algorithms' output was a probability from 0 to 1. We assessed algorithm performance using the area under the receiver operating characteristic (AUC) curves. Each algorithm's performance was evaluated individually and together as an ensemble. We randomly selected 5000 from 87,633 citations to train and internally validate our algorithms. Of the 5000 selected citations, 589 (12%) were confirmed CRT reports. We then externally validated our algorithms on an independent set of 1916 randomized trial citations, with 665 (35%) confirmed CRT reports.ResultsIn internal validation, the ensemble algorithm discriminated best for identifying CRT reports with an AUC of 98.6% (95% confidence interval: 97.8%, 99.4%), sensitivity of 97.7% (94.3%, 100%), and specificity of 85.0% (81.8%, 88.1%). In external validation, the ensemble algorithm had an AUC of 97.8% (97.0%, 98.5%), sensitivity of 97.6% (96.4%, 98.6%), and specificity of 78.2% (75.9%, 80.4%)). All three individual algorithms performed well, but less so than the ensemble.ConclusionsWe successfully developed high-performance algorithms that identified whether a citation was a CRT report with high sensitivity and moderately high specificity. We provide open-source software to facilitate the use of our algorithms in practice.

Dataset Information

Learning to identify Protected Health Information by integrating knowledge- and data-driven algorithms: A case study on psychiatric evaluation notes.

Publications

Learning to identify Protected Health Information by integrating knowledge- and data-driven algorithms: A case study on psychiatric evaluation notes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets