Building bridges across electronic health record systems through inferred phenotypic topics.

OBJECTIVE: Data in electronic health records (EHRs) is being increasingly leveraged for secondary uses, ranging from biomedical association studies to comparative effectiveness. To perform studies at scale and transfer knowledge from one institution to another in a meaningful way, we need to harmonize the phenotypes in such systems. Traditionally, this has been accomplished through expert specification of phenotypes via standardized terminologies, such as billing codes. However, this approach may be biased by the experience and expectations of the experts, as well as the vocabulary used to describe such patients. The goal of this work is to develop a data-driven strategy to (1) infer phenotypic topics within patient populations and (2) assess the degree to which such topics facilitate a mapping across populations in disparate healthcare systems. METHODS: We adapt a generative topic modeling strategy, based on latent Dirichlet allocation, to infer phenotypic topics. We utilize a variance analysis to assess the projection of a patient population from one healthcare system onto the topics learned from another system. The consistency of learned phenotypic topics was evaluated using (1) the similarity of topics, (2) the stability of a patient population across topics, and (3) the transferability of a topic across sites. We evaluated our approaches using four months of inpatient data from two geographically distinct healthcare systems: (1) Northwestern Memorial Hospital (NMH) and (2) Vanderbilt University Medical Center (VUMC). RESULTS: The method learned 25 phenotypic topics from each healthcare system. The average cosine similarity between matched topics across the two sites was 0.39, a remarkably high value given the very high dimensionality of the feature space. The average stability of VUMC and NMH patients across the topics of two sites was 0.988 and 0.812, respectively, as measured by the Pearson correlation coefficient. Also the VUMC and NMH topics have smaller variance of characterizing patient population of two sites than standard clinical terminologies (e.g., ICD9), suggesting they may be more reliably transferred across hospital systems. CONCLUSIONS: Phenotypic topics learned from EHR data can be more stable and transferable than billing codes for characterizing the general status of a patient population. This suggests that EHR-based research may be able to leverage such phenotypic topics as variables when pooling patient populations in predictive models.

Overcome the Limitation of Phenome-Wide Association Studies (PheWAS): Extension of PheWAS to Efficient and Robust Large-Scale ICD Codes Analysis.

The Phenome-wide association studies (PheWAS) have become widely used for efficient, high-throughput evaluation of relationship between a genetic factor and a large number of disease phenotypes, typically extracted from a DNA biobank linked with electronic medical records (EMR). Phecodes, billing code-derived disease case-control status, are usually used as outcome variables in PheWAS and logistic regression has been the standard choice of analysis method. Since the clinical diagnoses in EMR are often inaccurate with errors which can lead to biases in the odds ratio estimates, much effort has been put to accurately define the cases and controls to ensure an accurate analysis. Specifically in order to correctly classify controls in the population, an exclusion criteria list for each Phecode was manually compiled to obtain unbiased odds ratios. However, the accuracy of the list cannot be guaranteed without extensive data curation process. The costly curation process limits the efficiency of large-scale analyses that take full advantage of all structured phenotypic information available in EMR. Here, we proposed to estimate relative risks (RR) instead. We first demonstrated the desired nature of RR that overcomes the inaccuracy in the controls via theoretical formula. With simulation and real data application, we further confirmed that RR is unbiased without compiling exclusion criteria lists. With RR as estimates, we are able to efficiently extend PheWAS to a larger-scale, phenome construction agnostic analysis of phenotypes, using ICD 9/10 codes, which preserve much more disease-related clinical information than Phecodes.

Assertion modeling and its role in clinical phenotype identification.

This paper describes an approach to assertion classification and an empirical study on the impact this task has on phenotype identification, a real world application in the clinical domain. The task of assertion classification is to assign to each medical concept mentioned in a clinical report (e.g., pneumonia, chest pain) a specific assertion category (e.g., present, absent, and possible). To improve the classification of medical assertions, we propose several new features that capture the semantic properties of special cue words highly indicative of a specific assertion category. The results obtained outperform the current state-of-the-art results for this task. Furthermore, we confirm the intuition that assertion classification contributes in significantly improving the results of phenotype identification from free-text clinical records.

Patient-specific comorbidities as prognostic variables for survival in myelofibrosis.

Treatment decisions in primary myelofibrosis (PMF) are guided by numerous prognostic systems. Patient-specific comorbidities have influence on treatment-related survival and are considered in clinical contexts but have not been routinely incorporated into current prognostic models. We hypothesized that patient-specific comorbidities would inform prognosis and could be incorporated into a quantitative score. All patients with PMF or secondary myelofibrosis with available DNA and comprehensive electronic health record (EHR) data treated at Vanderbilt University Medical Center between 1995 and 2016 were identified within Vanderbilt's Synthetic Derivative and BioVU Biobank. We recapitulated established PMF risk scores (eg, Dynamic International Prognostic Scoring System [DIPSS], DIPSS plus, Genetics-Based Prognostic Scoring System, Mutation-Enhanced International Prognostic Scoring System 70+) and comorbidities through EHR chart extraction and next-generation sequencing on biobanked peripheral blood DNA. The impact of comorbidities was assessed via DIPSS-adjusted overall survival using Bonferroni correction. Comorbidities associated with inferior survival include renal failure/dysfunction (hazard ratio [HR], 4.3; 95% confidence interval [95% CI], 2.1-8.9; P = .0001), intracranial hemorrhage (HR, 28.7; 95% CI, 7.0-116.8; P = 2.83e-06), invasive fungal infection (HR, 41.2; 95% CI, 7.2-235.2; P = 2.90e-05), and chronic encephalopathy (HR, 15.1; 95% CI, 3.8-59.4; P = .0001). The extended DIPSS model including all 4 significant comorbidities showed a significantly higher discriminating power (C-index 0.81; 95% CI, 0.78-0.84) than the original DIPSS model (C-index 0.73; 95% CI, 0.70-0.77). In summary, we repurposed an institutional biobank to identify and risk-classify an uncommon hematologic malignancy by established (eg, DIPSS) and other clinical and pathologic factors (eg, comorbidities) in an unbiased fashion. The inclusion of comorbidities into risk evaluation may augment prognostic capability of future genetics-based scoring systems.

Assessing the role of a medication-indication resource in the treatment relation extraction from clinical text.

OBJECTIVE: To evaluate the contribution of the MEDication Indication (MEDI) resource and SemRep for identifying treatment relations in clinical text. MATERIALS AND METHODS: We first processed clinical documents with SemRep to extract the Unified Medical Language System (UMLS) concepts and the treatment relations between them. Then, we incorporated MEDI into a simple algorithm that identifies treatment relations between two concepts if they match a medication-indication pair in this resource. For a better coverage, we expanded MEDI using ontology relationships from RxNorm and UMLS Metathesaurus. We also developed two ensemble methods, which combined the predictions of SemRep and the MEDI algorithm. We evaluated our selected methods on two datasets, a Vanderbilt corpus of 6864 discharge summaries and the 2010 Informatics for Integrating Biology and the Bedside (i2b2)/Veteran's Affairs (VA) challenge dataset. RESULTS: The Vanderbilt dataset included 958 manually annotated treatment relations. A double annotation was performed on 25% of relations with high agreement (Cohen's κ = 0.86). The evaluation consisted of comparing the manual annotated relations with the relations identified by SemRep, the MEDI algorithm, and the two ensemble methods. On the first dataset, the best F1-measure results achieved by the MEDI algorithm and the union of the two resources (78.7 and 80, respectively) were significantly higher than the SemRep results (72.3). On the second dataset, the MEDI algorithm achieved better precision and significantly lower recall values than the best system in the i2b2 challenge. The two systems obtained comparable F1-measure values on the subset of i2b2 relations with both arguments in MEDI. CONCLUSIONS: Both SemRep and MEDI can be used to extract treatment relations from clinical text. Knowledge-based extraction with MEDI outperformed use of SemRep alone, but superior performance was achieved by integrating both systems. The integration of knowledge-based resources such as MEDI into information extraction systems such as SemRep and the i2b2 relation extractors may improve treatment relation extraction from clinical text.

Pneumonia identification using statistical feature selection.

OBJECTIVE: This paper describes a natural language processing system for the task of pneumonia identification. Based on the information extracted from the narrative reports associated with a patient, the task is to identify whether or not the patient is positive for pneumonia. DESIGN: A binary classifier was employed to identify pneumonia from a dataset of multiple types of clinical notes created for 426 patients during their stay in the intensive care unit. For this purpose, three types of features were considered: (1) word n-grams, (2) Unified Medical Language System (UMLS) concepts, and (3) assertion values associated with pneumonia expressions. System performance was greatly increased by a feature selection approach which uses statistical significance testing to rank features based on their association with the two categories of pneumonia identification. RESULTS: Besides testing our system on the entire cohort of 426 patients (unrestricted dataset), we also used a smaller subset of 236 patients (restricted dataset). The performance of the system was compared with the results of a baseline previously proposed for these two datasets. The best results achieved by the system (85.71 and 81.67 F1-measure) are significantly better than the baseline results (50.70 and 49.10 F1-measure) on the restricted and unrestricted datasets, respectively. CONCLUSION: Using a statistical feature selection approach that allows the feature extractor to consider only the most informative features from the feature space significantly improves the performance over a baseline that uses all the features from the same feature space. Extracting the assertion value for pneumonia expressions further improves the system performance.

A knowledge extraction framework for biomedical pathways.

In this paper we present a novel knowledge extraction framework that is based on semantic parsing. The semantic information originates in a variety of resources, but one in particular, namely BioFrameNet, is central to the characterization of complex events and processes that form biomedical pathways. The paper discusses the promising results of semantic parsing and explains how these results can be used for capturing complex medical knowledge.