Data-driven approach for assessing utility of medical tests using electronic medical records.

OBJECTIVE: To precisely define the utility of tests in a clinical pathway through data-driven analysis of the electronic medical record (EMR). MATERIALS AND METHODS: The information content was defined in terms of the entropy of the expected value of the test related to a given outcome. A kernel density classifier was used to estimate the necessary distributions. To validate the method, we used data from the EMR of the gastrointestinal department at a university hospital. Blood tests from patients undergoing surgery for gastrointestinal surgery were analyzed with respect to second surgery within 30 days of the index surgery. RESULTS: The information content is clearly reflected in the patient pathway for certain combinations of tests and outcomes. C-reactive protein tests coupled to anastomosis leakage, a severe complication show a clear pattern of information gain through the patient trajectory, where the greatest gain from the test is 3-4 days post index surgery. DISCUSSION: We have defined the information content in a data-driven and information theoretic way such that the utility of a test can be precisely defined. The results reflect clinical knowledge. In the case we used the tests carry little negative impact. The general approach can be expanded to cases that carry a substantial negative impact, such as in certain radiological techniques.

Prevalence of heart failure signs and symptoms in a large primary care population identified through the use of text and data mining of the electronic health record.

BACKGROUND: The electronic health record (EHR) contains a tremendous amount of data that if appropriately detected can lead to earlier identification of disease states such as heart failure (HF). Using a novel text and data analytic tool we explored the longitudinal EHR of over 50,000 primary care patients to identify the documentation of the signs and symptoms of HF in the years preceding its diagnosis. METHODS AND RESULTS: Retrospective analysis consisted of 4,644 incident HF cases and 45,981 group-matched control subjects. Documentation of Framingham HF signs and symptoms within encounter notes were carried out with the use of a previously validated natural language processing procedure. A total of 892,805 affirmed criteria were documented over an average observation period of 3.4 years. Among eventual HF cases, 85% had ≥1 criterion within 1 year before their HF diagnosis, as did 55% of control subjects. Substantial variability in the prevalence of individual signs and symptoms were found in both case and control subjects. CONCLUSIONS: HF signs and symptoms are frequently documented in a primary care population as identified through automated text and data mining of EHRs. Their frequent identification demonstrates the rich data available within EHRs that will allow for future work on automated criterion identification to help develop predictive models for HF.

Early detection of heart failure with varying prediction windows by structured and unstructured data in electronic health records.

Heart failure (HF) prevalence is increasing and is among the most costly diseases to society. Early detection of HF would provide the means to test lifestyle and pharmacologic interventions that may slow disease progression and improve patient outcomes. This study used structured and unstructured data from electronic health records (EHR) to predict onset of HF with a particular focus on how prediction accuracy varied in relation to time before diagnosis. EHR data were extracted from a single health care system and used to identify incident HF among primary care patients who received care between 2001 and 2010. A total of 1,684 incident HF cases were identified and 13,525 controls were selected from the same primary care practices. Models were compared by varying the beginning of the prediction window from 60 to 720 days before HF diagnosis. As the prediction window decreased, the performance [AUC (95% CIs)] of the predictive HF models increased from 65% (63%-66%) to 74% (73%-75%) for the unstructured, from 73% (72%-75%) to 81% (80%-83%) for the structured, and from 76% (74%-77%) to 83% (77%-85%) for the combined data.

Prescription Extraction from Clinical Notes: Towards Automating EMR Medication Reconciliation.

Medication in for ma lion is one of [he most important clinical data types in electronic medical records (EMR) This study developed an NLP application (PredMED) to extract full prescriptions and their relevant components from a large corpus of unstructured ambulatory office visit clinical notes and the corresponding structured medication reconciliation (MED REC) data in the EMR. PredMED achieved an 84.4% F-score on office visit encounter notes and 95.0% on MED"REC data, outperforming two available medication extraction systems. To assess the potential for using automatically extracted prescriptions in the medication reconciliation task, we manually analyzed discrepancies between prescriptions found in clinical encounter notes and in matching MED_REC data for sample patient encounters.

Acquisition, Analysis, and Sharing of Data in 2015 and Beyond: A Survey of the Landscape: A Conference Report From the American Heart Association Data Summit 2015.

BACKGROUND: A 1.5-day interactive forum was convened to discuss critical issues in the acquisition, analysis, and sharing of data in the field of cardiovascular and stroke science. The discussion will serve as the foundation for the American Heart Association's (AHA's) near-term and future strategies in the Big Data area. The concepts evolving from this forum may also inform other fields of medicine and science. METHODS AND RESULTS: A total of 47 participants representing stakeholders from 7 domains (patients, basic scientists, clinical investigators, population researchers, clinicians and healthcare system administrators, industry, and regulatory authorities) participated in the conference. Presentation topics included updates on data as viewed from conventional medical and nonmedical sources, building and using Big Data repositories, articulation of the goals of data sharing, and principles of responsible data sharing. Facilitated breakout sessions were conducted to examine what each of the 7 stakeholder domains wants from Big Data under ideal circumstances and the possible roles that the AHA might play in meeting their needs. Important areas that are high priorities for further study regarding Big Data include a description of the methodology of how to acquire and analyze findings, validation of the veracity of discoveries from such research, and integration into investigative and clinical care aspects of future cardiovascular and stroke medicine. Potential roles that the AHA might consider include facilitating a standards discussion (eg, tools, methodology, and appropriate data use), providing education (eg, healthcare providers, patients, investigators), and helping build an interoperable digital ecosystem in cardiovascular and stroke science. CONCLUSION: There was a consensus across stakeholder domains that Big Data holds great promise for revolutionizing the way cardiovascular and stroke research is conducted and clinical care is delivered; however, there is a clear need for the creation of a vision of how to use it to achieve the desired goals. Potential roles for the AHA center around facilitating a discussion of standards, providing education, and helping establish a cardiovascular digital ecosystem. This ecosystem should be interoperable and needs to interface with the rapidly growing digital object environment of the modern-day healthcare system.

Automatic identification of heart failure diagnostic criteria, using text analysis of clinical notes from electronic health records.

OBJECTIVE: Early detection of Heart Failure (HF) could mitigate the enormous individual and societal burden from this disease. Clinical detection is based, in part, on recognition of the multiple signs and symptoms comprising the Framingham HF diagnostic criteria that are typically documented, but not necessarily synthesized, by primary care physicians well before more specific diagnostic studies are done. We developed a natural language processing (NLP) procedure to identify Framingham HF signs and symptoms among primary care patients, using electronic health record (EHR) clinical notes, as a prelude to pattern analysis and clinical decision support for early detection of HF. DESIGN: We developed a hybrid NLP pipeline that performs two levels of analysis: (1) At the criteria mention level, a rule-based NLP system is constructed to annotate all affirmative and negative mentions of Framingham criteria. (2) At the encounter level, we construct a system to label encounters according to whether any Framingham criterion is asserted, denied, or unknown. MEASUREMENTS: Precision, recall, and F-score are used as performance metrics for criteria mention extraction and for encounter labeling. RESULTS: Our criteria mention extractions achieve a precision of 0.925, a recall of 0.896, and an F-score of 0.910. Encounter labeling achieves an F-score of 0.932. CONCLUSION: Our system accurately identifies and labels affirmations and denials of Framingham diagnostic criteria in primary care clinical notes and may help in the attempt to improve the early detection of HF. With adaptation and tooling, our development methodology can be repeated in new problem settings.

A framework for mining signatures from event sequences and its applications in healthcare data.

This paper proposes a novel temporal knowledge representation and learning framework to perform large-scale temporal signature mining of longitudinal heterogeneous event data. The framework enables the representation, extraction, and mining of high-order latent event structure and relationships within single and multiple event sequences. The proposed knowledge representation maps the heterogeneous event sequences to a geometric image by encoding events as a structured spatial-temporal shape process. We present a doubly constrained convolutional sparse coding framework that learns interpretable and shift-invariant latent temporal event signatures. We show how to cope with the sparsity in the data as well as in the latent factor model by inducing a double sparsity constraint on the ß-divergence to learn an overcomplete sparse latent factor model. A novel stochastic optimization scheme performs large-scale incremental learning of group-specific temporal event signatures. We validate the framework on synthetic data and on an electronic health record dataset.

A healthcare utilization analysis framework for hot spotting and contextual anomaly detection.

Patient medical records today contain vast amount of information regarding patient conditions along with treatment and procedure records. Systematic healthcare resource utilization analysis leveraging such observational data can provide critical insights to guide resource planning and improve the quality of care delivery while reducing cost. Of particular interest to providers are hot spotting: the ability to identify in a timely manner heavy users of the systems and their patterns of utilization so that targeted intervention programs can be instituted, and anomaly detection: the ability to identify anomalous utilization cases where the patients incurred levels of utilization that are unexpected given their clinical characteristics which may require corrective actions. Past work on medical utilization pattern analysis has focused on disease specific studies. We present a framework for utilization analysis that can be easily applied to any patient population. The framework includes two main components: utilization profiling and hot spotting, where we use a vector space model to represent patient utilization profiles, and apply clustering techniques to identify utilization groups within a given population and isolate high utilizers of different types; and contextual anomaly detection for utilization, where models that map patient's clinical characteristics to the utilization level are built in order to quantify the deviation between the expected and actual utilization levels and identify anomalies. We demonstrate the effectiveness of the framework using claims data collected from a population of 7667 diabetes patients. Our analysis demonstrates the usefulness of the proposed approaches in identifying clinically meaningful instances for both hot spotting and anomaly detection. In future work we plan to incorporate additional sources of observational data including EMRs and disease registries, and develop analytics models to leverage temporal relationships among medical encounters to provide more in-depth insights.

Visual cluster analysis in support of clinical decision intelligence.

Electronic health records (EHRs) contain a wealth of information about patients. In addition to providing efficient and accurate records for individual patients, large databases of EHRs contain valuable information about overall patient populations. While statistical insights describing an overall population are beneficial, they are often not specific enough to use as the basis for individualized patient-centric decisions. To address this challenge, we describe an approach based on patient similarity which analyzes an EHR database to extract a cohort of patient records most similar to a specific target patient. Clusters of similar patients are then visualized to allow interactive visual refinement by human experts. Statistics are then extracted from the refined patient clusters and displayed to users. The statistical insights taken from these refined clusters provide personalized guidance for complex decisions. This paper focuses on the cluster refinement stage where an expert user must interactively (a) judge the quality and contents of automatically generated similar patient clusters, and (b) refine the clusters based on his/her expertise. We describe the DICON visualization tool which allows users to interactively view and refine multidimensional similar patient clusters. We also present results from a preliminary evaluation where two medical doctors provided feedback on our approach.

Real-time prognosis of ICU physiological data streams.

This paper presents a system capable of predicting in real-time the evolution of Intensive Care Unit (ICU) physiological patient data streams. It leverages a state of the art stream computing platform to host analytics capable of making such prognosis in real time. The focus is on online algorithms that do not require a training phase. We use Fading-Memory Polynomial filters [8] on the frequency domain to predict windows of ICU data streams. We report on both the system and the performance of this approach when applied to traces of more than 1500 ICU patients obtained from the MIMIC-II database [1].

Predicting Patient's Trajectory of Physiological Data using Temporal Trends in Similar Patients: A System for Near-Term Prognostics.

Providing near-term prognostic insight to clinicians helps them to better assess the near-term impact of their decisions and potential impending events affecting the patient. In this work, we present a novel system, which leverages inter-patient similarity for retrieving patients who display similar trends in their physiological time-series data. Data from the retrieved patient cohort is then used to project patient data into the future to provide insights for the query patient. The proposed approach and system were tested using the MIMIC II database, which consists of physiological waveforms, and accompanying clinical data obtained for ICU patients. In the experiments we report the effectiveness of the inter-patient similarity measure and the accuracy of the projection of patients' data. We also discuss the visual interface that conveys the near-term prognostic decision support to the user.