Data integration of structured and unstructured sources for assigning clinical codes to patient stays

Scheurwegs, Elyne; Luyckx, Kim; Luyten, Léon; Daelemans, Walter; Bulcke, Tim Van den

doi:10.1093/jamia/ocv115

Cited by 73 publications

(53 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To tackle the integration problem, studies have been conducted on the knowledge level and the data level, respectively. Some rely on domain knowledge to extract a joint patient cohort by dening criteria from dierent data types [4], while others explore the possibility of integrating heterogeneous EHR data prior to or post modeling [5,6,7,8]. The focus of this study is on the latter: analyzing complex longitudinal data.…”

Section: Introductionmentioning

confidence: 99%

Learning from heterogeneous temporal data in electronic health records

Zhao

Papapetrou

Asker

et al. 2017

Journal of Biomedical Informatics

View full text Add to dashboard Cite

Electronic health records contain large amounts of longitudinal data that are valuable for biomedical informatics research. The application of machine learning is a promising alternative to manual analysis of such data. However, the complex structure of the data, which includes clinical events that are unevenly distributed over time, poses a challenge for standard learning algorithms. Some approaches to modeling temporal data rely on extracting single values from time series; however, this leads to the loss of potentially valuable sequential information. How to better account for the temporality of clinical data, hence, remains an important research question. In this study, novel representations of temporal data in electronic health records are explored. These representations retain the sequential information, and are directly compatible with standard machine learning algorithms. The explored methods are based on symbolic sequence representations of time series data, which are utilized in a number of different ways. An empirical investigation, using 19 datasets comprising clinical measurements observed over time from a real database of electronic health records, shows that using a distance measure to random subsequences leads to substantial improvements in predictive performance compared to using the original sequences or clustering the sequences. Evidence is moreover provided on the quality of the symbolic sequence representation by comparing it to sequences that are generated using domain knowledge by clinical experts. The proposed method creates representations that better account for the temporality of clinical events, which is often key to prediction tasks in the biomedical domain.

show abstract

Section: Introductionmentioning

confidence: 99%

Learning from heterogeneous temporal data in electronic health records

Zhao

Papapetrou

Asker

et al. 2017

Journal of Biomedical Informatics

View full text Add to dashboard Cite

show abstract

“…The judgement as to whether to implement a technology depends not only on how the technology fits with what that corporation already possesses, but also what is available on the market [15,[18][19][20]. Some of these technological aspects mentioned in the literature are technology integration, readiness, and security [4,17,21,22]. Technology integration is "the level of interconnectivity among back-office databases and information system inside corporations and these externally combined with databases and interdependent system" [17].…”

Section: Technological Dimensionmentioning

confidence: 99%

“…Technology integration is "the level of interconnectivity among back-office databases and information system inside corporations and these externally combined with databases and interdependent system" [17]. Technology integration is a critical factor for EHR adoption, given that EHR requires simplified data flowing throughout the value chain, and should automatically respond to alterations in downstream systems or procedures [21,22]. Technology readiness refers to "IT human resources and technology structure .…”

Section: Technological Dimensionmentioning

confidence: 99%

A Hybrid MCDM Model for Improving the Electronic Health Record to Better Serve Client Needs

Liou

et al. 2017

Sustainability

View full text Add to dashboard Cite

Abstract:Although the electronic health record (EHR) is a promising innovation in the healthcare industry, the implementation of EHR has been relatively slow. A theoretical structure for the exploration and improvement of this usage of EHR is proposed. Incorporating the theoretical structure of TOE (technology-organization-environment), we apply the DEMATEL (decision-making trial and evaluation laboratory) technique to illustrate the influence-matrix and to construct the INRM (influential network relationship map). Based on this DEMATEL influence matrix and the fundamental concepts of ANP (Analytic Hierarchy Process), we derive influential weights for the criteria. These influential weights are then combined with the modified VIKOR (VlseKriterijumska Optimizacija I Kompromisno Resenje) method to find ways to understand and enhance the usage of EHR technology. The outcome demonstrates that our model can not only be used for implementation of EHR technology, but can also be applied to analyze the gaps in performance between the aspiration level and present performance values in individual criterion/dimension.

show abstract

“…To assist decision making, relevant information in such reports needs to be extracted, for which machine learning methods are typically used. This extracted information can be used as input for several applications, such as the identification of diagnostic criteria for heart failure [1, 2] or the prediction of diagnosis and procedure codes [3]. Since the effectiveness of data-driven algorithms not only depends on data size, but also on the quality of data representation, robust information extraction techniques are required [4, 5].…”

Section: Introductionmentioning

confidence: 99%

Counting trees in Random Forests: Predicting symptom severity in psychiatric intake reports

Scheurwegs

Sushil

Tulkens

et al. 2017

Journal of Biomedical Informatics

Self Cite

View full text Add to dashboard Cite

The CEGS N-GRID 2016 Shared Task (Filannino, Stubbs, Uzuner (2017)) in Clinical Natural Language Processing introduces the assignment of a severity score to a psychiatric symptom, based on a psychiatric intake report. We present a method that employs the inherent interview-like structure of the report to extract relevant information from the report and generate a representation. The representation consists of a restricted set of psychiatric concepts (and the context they occur in), identified using medical concepts defined in UMLS that are directly related to the psychiatric diagnoses present in the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) ontology. Random Forests provides a generalization of the extracted, case-specific features in our representation. The best variant presented here scored an inverse mean absolute error (MAE) of 80.64%. A concise concept-based representation, paired with identification of concept certainty and scope (family, patient), shows a robust performance on the task.

show abstract

Data integration of structured and unstructured sources for assigning clinical codes to patient stays

Abstract: We demonstrated that models using multiple electronic health record data sources systematically outperform models using data sources in isolation in the task of predicting ICD-9-CM codes over a broad range of medical specialties.

Cited by 73 publications

References 24 publications

Learning from heterogeneous temporal data in electronic health records

Learning from heterogeneous temporal data in electronic health records

A Hybrid MCDM Model for Improving the Electronic Health Record to Better Serve Client Needs

Counting trees in Random Forests: Predicting symptom severity in psychiatric intake reports

Contact Info

Product

Resources

About