When Context Matters for Credible Measurement of Drug-Drug Interactions Based on Real-World Data.

The frequency of potential drug-drug interactions (DDI) in published studies on real world data considerably varies due to the methodological framework. Contextualization of DDI has a proven effect in limiting false positives. In this paper, we experimented with the application of various DDIs contexts elements to see their impact on the frequency of potential DDIs measured on the same set of prescription data collected in EDSaN, the clinical data warehouse of Rouen University Hospital. Depending on the context applied, the frequency of daily prescriptions with potential DDI ranged from 0.89% to 3.90%. Substance-level analysis accounted for 48% of false positives because it did not account for some drug-related attributes. Consideration of the patient's context could eliminate up to an additional 29% of false positives.

Evaluation of Doc'EDS: a French semantic search tool to query health documents from a clinical data warehouse.

BACKGROUND: Unstructured data from electronic health records represent a wealth of information. Doc'EDS is a pre-screening tool based on textual and semantic analysis. The Doc'EDS system provides a graphic user interface to search documents in French. The aim of this study was to present the Doc'EDS tool and to provide a formal evaluation of its semantic features. METHODS: Doc'EDS is a search tool built on top of the clinical data warehouse developed at Rouen University Hospital. This tool is a multilevel search engine combining structured and unstructured data. It also provides basic analytical features and semantic utilities. A formal evaluation was conducted to measure the impact of Natural Language Processing algorithms. RESULTS: Approximately 18.1 million narrative documents are stored in Doc'EDS. The formal evaluation was conducted in 5000 clinical concepts that were manually collected. The F-measures of negative concepts and hypothetical concepts were respectively 0.89 and 0.57. CONCLUSION: In this formal evaluation, we have shown that Doc'EDS is able to deal with language subtleties to enhance an advanced full text search in French health documents. The Doc'EDS tool is currently used on a daily basis to help researchers to identify patient cohorts thanks to unstructured data.

Assisting Data Retrieval with a Drug Knowledge Graph.

The Normandy health data warehouse EDSaN integrates the medication orders from the University Hospital of Rouen (France). This study aims at describing the design and the evaluation of an information retrieval system founded on a complex and semantically augmented knowledge graph dedicated to EDSaN drugs' prescriptions. The system is intended to help the selection of drugs in the search process by health professionals. The manual evaluation of the relevance of the returned drugs showed encouraging results as expected. A deeper analysis in order to improve the ranking method is needed and will be performed in a future work.

Performance evaluation of three semantic expansions to query PubMed.

BACKGROUND: PubMed is one of the most important basic tools to access medical literature. Semantic query expansion using synonyms can improve retrieval efficacy. OBJECTIVE: The objective was to evaluate the performance of three semantic query expansion strategies. METHODS: Queries were built for forty MeSH descriptors using three semantic expansion strategies (MeSH synonyms, UMLS mappings, and mappings created by the CISMeF team), then sent to PubMed. To evaluate expansion performances for each query, the first twenty citations were selected, and their relevance were judged by three independent evaluators based on the title and abstract. RESULTS: Queries built with the UMLS expansion provided new citations with a slightly higher mean precision (74.19%) than with the CISMeF expansion (70.28%), although the difference was not significant. Inter-rater agreement was 0.28. Results varied greatly depending on the descriptor selected. DISCUSSION: The number of citations retrieved by the three strategies and their precision varied greatly according to the descriptor. This heterogeneity could be explained by the quality of the synonyms. Optimal use of these different expansions would be through various combinations of UMLS and CISMeF intersections or unions. CONCLUSION: Information retrieval tools should propose different semantic expansions depending on the descriptor and the search objectives.

Identification of the Best Semantic Expansion to Query PubMed Through Automatic Performance Assessment of Four Search Strategies on All Medical Subject Heading Descriptors: Comparative Study.

BACKGROUND: With the continuous expansion of available biomedical data, efficient and effective information retrieval has become of utmost importance. Semantic expansion of queries using synonyms may improve information retrieval. OBJECTIVE: The aim of this study was to automatically construct and evaluate expanded PubMed queries of the form "preferred term"[MH] OR "preferred term"[TIAB] OR "synonym 1"[TIAB] OR "synonym 2"[TIAB] OR …, for each of the 28,313 Medical Subject Heading (MeSH) descriptors, by using different semantic expansion strategies. We sought to propose an innovative method that could automatically evaluate these strategies, based on the three main metrics used in information science (precision, recall, and F-measure). METHODS: Three semantic expansion strategies were assessed. They differed by the synonyms used to build the queries as follows: MeSH synonyms, Unified Medical Language System (UMLS) mappings, and custom mappings (Catalogue et Index des Sites Médicaux de langue Française [CISMeF]). The precision, recall, and F-measure metrics were automatically computed for the three strategies and for the standard automatic term mapping (ATM) of PubMed. The method to automatically compute the metrics involved computing the number of all relevant citations (A), using National Library of Medicine indexing as the gold standard ("preferred term"[MH]), the number of citations retrieved by the added terms ("synonym 1"[TIAB] OR "synonym 2"[TIAB] OR …) (B), and the number of relevant citations retrieved by the added terms (combining the previous two queries with an "AND" operator) (C). It was possible to programmatically compute the metrics for each strategy using each of the 28,313 MeSH descriptors as a "preferred term," corresponding to 239,724 different queries built and sent to the PubMed application program interface. The four search strategies were ranked and compared for each metric. RESULTS: ATM had the worst performance for all three metrics among the four strategies. The MeSH strategy had the best mean precision (51%, SD 23%). The UMLS strategy had the best recall and F-measure (41%, SD 31% and 36%, SD 24%, respectively). CISMeF had the second best recall and F-measure (40%, SD 31% and 35%, SD 24%, respectively). However, considering a cutoff of 5%, CISMeF had better precision than UMLS for 1180 descriptors, better recall for 793 descriptors, and better F-measure for 678 descriptors. CONCLUSIONS: This study highlights the importance of using semantic expansion strategies to improve information retrieval. However, the performances of a given strategy, relatively to another, varied greatly depending on the MeSH descriptor. These results confirm there is no ideal search strategy for all descriptors. Different semantic expansions should be used depending on the descriptor and the user's objectives. Thus, we developed an interface that allows users to input a descriptor and then proposes the best semantic expansion to maximize the three main metrics (precision, recall, and F-measure).

Building a Semantic Health Data Warehouse in the Context of Clinical Trials: Development and Usability Study.

BACKGROUND: The huge amount of clinical, administrative, and demographic data recorded and maintained by hospitals can be consistently aggregated into health data warehouses with a uniform data model. In 2017, Rouen University Hospital (RUH) initiated the design of a semantic health data warehouse enabling both semantic description and retrieval of health information. OBJECTIVE: This study aimed to present a proof of concept of this semantic health data warehouse, based on the data of 250,000 patients from RUH, and to assess its ability to assist health professionals in prescreening eligible patients in a clinical trials context. METHODS: The semantic health data warehouse relies on 3 distinct semantic layers: (1) a terminology and ontology portal, (2) a semantic annotator, and (3) a semantic search engine and NoSQL (not only structured query language) layer to enhance data access performances. The system adopts an entity-centered vision that provides generic search capabilities able to express data requirements in terms of the whole set of interconnected conceptual entities that compose health information. RESULTS: We assessed the ability of the system to assist the search for 95 inclusion and exclusion criteria originating from 5 randomly chosen clinical trials from RUH. The system succeeded in fully automating 39% (29/74) of the criteria and was efficiently used as a prescreening tool for 73% (54/74) of them. Furthermore, the targeted sources of information and the search engine-related or data-related limitations that could explain the results for each criterion were also observed. CONCLUSIONS: The entity-centered vision contrasts with the usual patient-centered vision adopted by existing systems. It enables more genericity in the information retrieval process. It also allows to fully exploit the semantic description of health information. Despite their semantic annotation, searching within clinical narratives remained the major challenge of the system. A finer annotation of the clinical texts and the addition of specific functionalities would significantly improve the results. The semantic aspect of the system combined with its generic entity-centered vision enables the processing of a large range of clinical questions. However, an important part of health information remains in clinical narratives, and we are currently investigating novel approaches (deep learning) to enhance the semantic annotation of those unstructured data.

Word Embedding for French Natural Language in Healthcare: A Comparative Study.

Structuring raw medical documents with ontology mapping is now the next step for medical intelligence. Deep learning models take as input mathematically embedded information, such as encoded texts. To do so, word embedding methods can represent every word from a text as a fixed-length vector. A formal evaluation of three word embedding methods has been performed on raw medical documents. The data corresponds to more than 12M diverse documents produced in the Rouen hospital (drug prescriptions, discharge and surgery summaries, inter-services letters, etc.). Automatic and manual validation demonstrates that Word2Vec based on the skip-gram architecture had the best rate on three out of four accuracy tests. This model will now be used as the first layer of an AI-based semantic annotator.

Word Embedding for the French Natural Language in Health Care: Comparative Study.

BACKGROUND: Word embedding technologies, a set of language modeling and feature learning techniques in natural language processing (NLP), are now used in a wide range of applications. However, no formal evaluation and comparison have been made on the ability of each of the 3 current most famous unsupervised implementations (Word2Vec, GloVe, and FastText) to keep track of the semantic similarities existing between words, when trained on the same dataset. OBJECTIVE: The aim of this study was to compare embedding methods trained on a corpus of French health-related documents produced in a professional context. The best method will then help us develop a new semantic annotator. METHODS: Unsupervised embedding models have been trained on 641,279 documents originating from the Rouen University Hospital. These data are not structured and cover a wide range of documents produced in a clinical setting (discharge summary, procedure reports, and prescriptions). In total, 4 rated evaluation tasks were defined (cosine similarity, odd one, analogy-based operations, and human formal evaluation) and applied on each model, as well as embedding visualization. RESULTS: Word2Vec had the highest score on 3 out of 4 rated tasks (analogy-based operations, odd one similarity, and human validation), particularly regarding the skip-gram architecture. CONCLUSIONS: Although this implementation had the best rate for semantic properties conservation, each model has its own qualities and defects, such as the training time, which is very short for GloVe, or morphological similarity conservation observed with FastText. Models and test sets produced by this study will be the first to be publicly available through a graphical interface to help advance the French biomedical research.

Terminology Coverage from Semantic Annotated Health Documents.

BACKGROUND: Unstructured health documents (e.g. discharge summaries) represent an important and unavoidable source of information. METHODS: A semantic annotator identified all the concepts present in the health documents from the clinical data warehouse of the Rouen University Hospital. RESULTS: 2,087,784,055 annotations were generated from a corpus of about 11.9 million documents with an average of 175 annotations per document. SNOMED CT, NCIt and MeSH were the top 3 terminologies that reported the most annotation. DISCUSSION: As expected, the most general terminologies with the most translated concepts were those with the most concepts identified.

Querying EHRs with a Semantic and Entity-Oriented Query Language.

While the digitization of medical documents has greatly expanded during the past decade, health information retrieval has become a great challenge to address many issues in medical research. Information retrieval in electronic health records (EHR) should also reduce the difficult tasks of manual information retrieval from records in paper format or computer. The aim of this article was to present the features of a semantic search engine implemented in EHRs. A flexible, scalable and entity-oriented query language tool is proposed. The program is designed to retrieve and visualize data which can support any Conceptual Data Model. The search engine deals with structured and unstructured data, for a sole patient from a caregiver perspective, and for a number of patients (e.g. epidemiology). Several types of queries on a test database containing 2,000 anonymized patients EHRs (i.e. approximately 200,000 records) were tested. These queries were able to accurately treat symbolic, textual, numerical and chronological data.

InfoRoute: the CISMeF Context-specific Search Algorithm.

OBJECTIVE: The aim of this paper was to present a practical InfoRoute algorithm and applications developed by CISMeF to perform a contextual information retrieval across multiple medical websites in different health domains. METHODS: The algorithm was developed to treat multiple types of queries: natural, Boolean and advanced. The algorithm also generates multiple types of queries: Boolean query, PubMed query or Advanced query. Each query can be extended via an inter alignments relationship from UMLS and HeTOP portal. RESULTS: A web service and two web applications have been developed based on the InfoRoute algorithm to generate links-query across multiple websites, i.e.: "PubMed" or "ClinicalTrials.org". CONCLUSION: The InfoRoute algorithm is a useful tool to perform contextual information retrieval across multiple medical websites in both English and French.