Relational machine learning for electronic health record-driven phenotyping

Abstract

Objective Electronic health records (EHR) offer medical and pharmacogenomics research unprecedented opportunities to identify and classify patients at risk. EHRs are collections of highly inter-dependent records that include biological, anatomical, physiological, and behavioral observations. They comprise a patient’s clinical phenome, where each patient has thousands of date-stamped records distributed across many relational tables. Development of EHR computer-based phenotyping algorithms require time and medical insight from clinical experts, who most often can only review a small patient subset representative of the total EHR records, to identify phenotype features. In this research we evaluate whether relational machine learning (ML) using Inductive Logic Programming (ILP) can contribute to addressing these issues as a viable approach for EHR-based phenotyping. Methods Two relational learning ILP approaches and three well-known WEKA (Waikato Environment for Knowledge Analysis) implementations of non-relational approaches (PART, J48, and JRIP) were used to develop models for nine phenotypes. International Classification of Diseases, Ninth Revision (ICD-9) coded EHR data were used to select training cohorts for the development of each phenotypic model. Accuracy, precision, recall, F-Measure, and Area Under the Receiver Operating Characteristic (AUROC) curve statistics were measured for each phenotypic model based on independent manually verified test cohorts. A two-sided binomial distribution test (sign test) compared the five ML approaches across phenotypes for statistical significance. Results We developed an approach to automatically label training examples using ICD-9 diagnosis codes for the ML approaches being evaluated. Nine phenotypic models for each MLapproach were evaluated, resulting in better overall model performance in AUROC using ILP when compared to PART (p=0.039), J48 (p=0.003) and JRIP (p=0.003). Discussion ILP has the potential to improve phenotyping by independently delivering clinically expert interpretable rules for phenotype definitions, or intuitive phenotypes to assist experts. Conclusion Relational learning using ILP offers a viable approach to EHR-driven phenotyping.

Big Data Science: Opportunities and Challenges to Address Minority Health and Health Disparities in the 21st Century

“…For example, machine learning applied to clinical data has been used to predict acute care use and cost of treatment for asthmatic patients, diagnose diabetes among adults, predict in-hospital mortality and drug response, improve disease classification, and identify disease subsets. [44][45][46][47] Taylor et al suggest that a machine learning algorithm using Big Data conforms to actual real time clinical practice, allows incorporation of far more clinical variables, and may assist in discovering unexpected predictors. 48 Big Data analytic tools such as natural language processing, machine-learning, or electronic casefinding algorithms applied to EHR data have produced a number of insights into genomics of disease and drug response.…”

National Public Health Data Systems in the United States: Applications to Child Agricultural Injury Surveillance

“…A system that includes both coded fields and a structured narrative has much broader potential . This is especially true given recent developments in electronic text mining and machine learning that have reduced the need for manual review when using data from narrative text to help address questions in medical research …”

“…17 This is especially true given recent developments in electronic text mining and machine learning that have reduced the need for manual review when using data from narrative text to help address questions in medical research. [18][19][20][21] One exemplary North American model of an agricultural injury surveillance system that employs such an approach is the Canadian Agricultural Injury Reporting (CAIR), 22 formerly the Canadian Agricultural Injury Surveillance Program. In this program, trained coders in each participating province abstract information from coroners' and medical examiners' files (deaths) and hospital discharge records (hospitalized cases) in a standardized manner.…”

Clinical Data Reuse or Secondary Use: Current Status and Potential Future Progress

“…Chen and colleagues described the application of relational data mining to detect anomalies in the accesses to communities information systems [179]. The study by Peissig and colleagues used Inductive Logic Programming (ILP) -a method that infers an hypothesis from the analysis of the background knowledge and examples -to derive phenotypes from EHR data [180].…”