Rainer Winnenburg

A method for systematic discovery of adverse drug events from clinical notes

OBJECTIVE Adverse drug events (ADEs) are undesired harmful effects resulting from use of a medication, and occur in 30% of hospitalized patients. The authors have developed a data-mining method for systematic, automated detection of ADEs from electronic medical records. MATERIALS AND METHODS This method uses the text from 9.5 million clinical notes, along with prior knowledge of drug usages and known ADEs, as inputs. These inputs are further processed into statistics used by a discriminative classifier which outputs the probability that a given drug-disorder pair represents a valid ADE association. Putative ADEs identified by the classifier are further filtered for positive support in 2 independent, complementary data sources. The authors evaluate this method by assessing support for the predictions in other curated data sources, including a manually curated, time-indexed reference standard of label change events. RESULTS This method uses a classifier that achieves an area under the curve of 0.94 on a held out test set. The classifier is used on 2,362,950 possible drug-disorder pairs comprised of 1602 unique drugs and 1475 unique disorders for which we had data, resulting in 240 high-confidence, well-supported drug-AE associations. Eighty-seven of them (36%) are supported in at least one of the resources that have information that was not available to the classifier. CONCLUSION This method demonstrates the feasibility of systematic post-marketing surveillance for ADEs using electronic medical records, a key component of the learning healthcare system.

A time-indexed reference standard of adverse drug reactions

Undetected adverse drug reactions (ADRs) pose a major burden on the health system. Data mining methodologies designed to identify signals of novel ADRs are of deep importance for drug safety surveillance. The development and evaluation of these methodologies requires proper reference benchmarks. While progress has recently been made in developing such benchmarks, our understanding of the performance characteristics of the data mining methodologies is limited because existing benchmarks do not support prospective performance evaluations. We address this shortcoming by providing a reference standard to support prospective performance evaluations. The reference standard was systematically curated from drug labeling revisions, such as new warnings, which were issued and communicated by the US Food and Drug Administration in 2013. The reference standard includes 62 positive test cases and 75 negative controls, and covers 44 drugs and 38 events. We provide usage guidance and empirical support for the reference standard by applying it to analyze two data sources commonly mined for drug safety surveillance.

The digital revolution in phenotyping

Phenotypes have gained increased notoriety in the clinical and biological domain owing to their application in numerous areas such as the discovery of disease genes and drug targets, phylogenetics and pharmacogenomics. Phenotypes, defined as observable characteristics of organisms, can be seen as one of the bridges that lead to a translation of experimental findings into clinical applications and thereby support ‘bench to bedside’ efforts. However, to build this translational bridge, a common and universal understanding of phenotypes is required that goes beyond domain-specific definitions. To achieve this ambitious goal, a digital revolution is ongoing that enables the encoding of data in computer-readable formats and the data storage in specialized repositories, ready for integration, enabling translational research. While phenome research is an ongoing endeavor, the true potential hidden in the currently available data still needs to be unlocked, offering exciting opportunities for the forthcoming years. Here, we provide insights into the state-of-the-art in digital phenotyping, by means of representing, acquiring and analyzing phenotype data. In addition, we provide visions of this field for future research work that could enable better applications of phenotype data.

Exploring adverse drug events at the class level

BackgroundWhile the association between a drug and an adverse event (ADE) is generally detected at the level of individual drugs, ADEs are often discussed at the class level, i.e., at the level of pharmacologic classes (e.g., in drug labels). We propose two approaches, one visual and one computational, to exploring the contribution of individual drugs to the class signal.MethodsHaving established a dataset of ADEs from MEDLINE, we aggregate drugs into ATC classes and ADEs into high-level MeSH terms. We compute statistical associations between drugs and ADEs at the drug level and at the class level. Finally, we visualize the signals at increasing levels of resolution using heat maps. We also automate the exploration of drug-ADE associations at the class level using clustering techniques.ResultsUsing our visual approach, we were able to uncover known associations, e.g., between fluoroquinolones and tendon injuries, and between statins and rhabdomyolysis. Using our computational approach, we systematically analyzed 488 associations between a drug class and an ADE.ConclusionsThe findings gained from our exploratory techniques should be of interest to the curators of ADE repositories and drug safety professionals. Our approach can be applied to different drug-ADE datasets, using different drug classification systems and different signal detection algorithms.

Generalized enrichment analysis improves the detection of adverse drug events from the biomedical literature

BackgroundIdentification of associations between marketed drugs and adverse events from the biomedical literature assists drug safety monitoring efforts. Assessing the significance of such literature-derived associations and determining the granularity at which they should be captured remains a challenge. Here, we assess how defining a selection of adverse event terms from MeSH, based on information content, can improve the detection of adverse events for drugs and drug classes.ResultsWe analyze a set of 105,354 candidate drug adverse event pairs extracted from article indexes in MEDLINE. First, we harmonize extracted adverse event terms by aggregating them into higher-level MeSH terms based on the terms’ information content. Then, we determine statistical enrichment of adverse events associated with drug and drug classes using a conditional hypergeometric test that adjusts for dependencies among associated terms. We compare our results with methods based on disproportionality analysis (proportional reporting ratio, PRR) and quantify the improvement in signal detection with our generalized enrichment analysis (GEA) approach using a gold standard of drug-adverse event associations spanning 174 drugs and four events. For single drugs, the best GEA method (Precision: .92/Recall: .71/F1-measure: .80) outperforms the best PRR based method (.69/.69/.69) on all four adverse event outcomes in our gold standard. For drug classes, our GEA performs similarly (.85/.69/.74) when increasing the level of abstraction for adverse event terms. Finally, on examining the 1609 individual drugs in our MEDLINE set, which map to chemical substances in ATC, we find signals for 1379 drugs (10,122 unique adverse event associations) on applying GEA with p < 0.005.ConclusionsWe present an approach based on generalized enrichment analysis that can be used to detect associations between drugs, drug classes and adverse events at a given level of granularity, at the same time correcting for known dependencies among events. Our study demonstrates the use of GEA, and the importance of choosing appropriate abstraction levels to complement current drug safety methods. We provide an R package for exploration of alternative abstraction levels of adverse event terms based on information content.

Toward multimodal signal detection of adverse drug reactions

OBJECTIVE Improving mechanisms to detect adverse drug reactions (ADRs) is key to strengthening post-marketing drug safety surveillance. Signal detection is presently unimodal, relying on a single information source. Multimodal signal detection is based on jointly analyzing multiple information sources. Building on, and expanding the work done in prior studies, the aim of the article is to further research on multimodal signal detection, explore its potential benefits, and propose methods for its construction and evaluation. MATERIAL AND METHODS Four data sources are investigated; FDAs adverse event reporting system, insurance claims, the MEDLINE citation database, and the logs of major Web search engines. Published methods are used to generate and combine signals from each data source. Two distinct reference benchmarks corresponding to well-established and recently labeled ADRs respectively are used to evaluate the performance of multimodal signal detection in terms of area under the ROC curve (AUC) and lead-time-to-detection, with the latter relative to labeling revision dates. RESULTS Limited to our reference benchmarks, multimodal signal detection provides AUC improvements ranging from 0.04 to 0.09 based on a widely used evaluation benchmark, and a comparative added lead-time of 7-22 months relative to labeling revision dates from a time-indexed benchmark. CONCLUSIONS The results support the notion that utilizing and jointly analyzing multiple data sources may lead to improved signal detection. Given certain data and benchmark limitations, the early stage of development, and the complexity of ADRs, it is currently not possible to make definitive statements about the ultimate utility of the concept. Continued development of multimodal signal detection requires a deeper understanding the data sources used, additional benchmarks, and further research on methods to generate and synthesize signals.

Using description logics to evaluate the consistency of drug-class membership relations in NDF-RT.

BackgroundThe NDF-RT (National Drug File Reference Terminology) is an ontology, which describes drugs and their properties and supports computerized physician order entry systems. NDF-RT’s classes are mostly specified using only necessary conditions and lack sufficient conditions, making its use limited until recently, when asserted drug-class relations were added. The addition of these asserted drug-class relations presents an opportunity to compare them with drug-class relations that can be inferred using the properties of drugs and drug classes in NDF-RT.MethodsWe enriched NDF-RT’s drug-classes with sufficient conditions, added property equivalences, and then used an OWL reasoner to infer drug-class membership relations. We compared the inferred class relations to the recently added asserted relations derived from FDA Structured Product Labels.ResultsThe inferred and asserted relations only match in about 50% of the cases, due to incompleteness of the drug descriptions and quality issues in the class definitions.ConclusionsThis investigation quantifies and categorizes the disparities between asserted and inferred drug-class relations and illustrates issues with class definitions and drug descriptions. In addition, it serves as an example of the benefits DL can add to ontology development and evaluation.

U-Index, a dataset and an impact metric for informatics tools and databases

Measuring the usage of informatics resources such as software tools and databases is essential to quantifying their impact, value and return on investment. We have developed a publicly available dataset of informatics resource publications and their citation network, along with an associated metric (u-Index) to measure informatics resources’ impact over time. Our dataset differentiates the context in which citations occur to distinguish between ‘awareness’ and ‘usage’, and uses a citing universe of open access publications to derive citation counts for quantifying impact. Resources with a high ratio of usage citations to awareness citations are likely to be widely used by others and have a high u-Index score. We have pre-calculated the u-Index for nearly 100,000 informatics resources. We demonstrate how the u-Index can be used to track informatics resource impact over time. The method of calculating the u-Index metric, the pre-computed u-Index values, and the dataset we compiled to calculate the u-Index are publicly available.