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

Abstract

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.

Patient and Graph Embeddings for Predictive Diagnosis of Drug Iatrogenesis

“…The IA.TROMED project (Fig. 1) relies on several hypotheses: (i) new approaches (embeddings and transformer-based) are needed to be explored and adapted to real life biomedical data: several billion concepts have to be linked from EDSaN [2] to the SNDS; (ii) the new models need to be launched and trained on several complementary and heterogeneous data (scientific literature in English and French [3]); (iii) Enriching semantically the deep learning models will allow to benefit from advanced rule-based processes and semantic web technologies; (iv) Mashing-up all the extracted knowledge from those heterogeneous resources will facilitate a better collaboration between clinicians and pharmacists, and researchers that will (v) adapt and optimize the algorithms, including technological trial.…”

Functional Requirements for Medical Data Integration into Knowledge Management Environments: Requirements Elicitation Approach Based on Systematic Literature Analysis

“…A complete list of requirements can be found in Multimedia Appendix 1 requirements. The most cited requirement is DA-01 with 5 articles [11,[33][34][35][36] [42][43][44]) extends DA-01 to include different data from electronic medical records or EHRs containing standard data sets about patients' demographics and more specific data sets about vital signs, laboratory tests, medication data, diagnoses, and procedures. DA-03 was derived from 3 publications [36,45,46], each of which describes the integration of various multimedia data.…”

“…DA-05 points out the need for an abstraction layer for acquiring both data and metadata from heterogeneous sources such as Patient Management Systems. [11], Lelong et al [33], Tahar et al [34], Tsiknakis et al [35], and Gupta et al [36] Possible heterogeneous source systems are LIMS a , HIS b , and CTMS c .…”

“…The first one, DA-01, states that data are integrated from different heterogeneous source systems. In this sense, the 5 studies[11,[33][34][35][36][37] mention different source systems, which include Health Care Information Systems in general, Laboratory Information Management Systems, and Clinical Trial Management Systems. These integration projects span various medical domains, including clinical trials in general, systems biology, and OMICS.…”

Assisting Data Retrieval with a Drug Knowledge Graph

“…Existing medicinal drug databases such as Wikidata [5], Drug Bank 2 , or GoodRx 3 contain valuable information but lack of comprehensiveness when taken separately and/or store some of this information as unstructured data [6]. In this study, the design of a system enabling the retrieval of prescription orders contained in the Normandy's Health Data Warehouse (EDSaN) [7] at the Rouen University Hospital (Normandy, France) is described. A conceptual graph of drug knowledge data was designed and used in the information retrieval process to retrieve the French Common Dispensing Unit (UCD) codes that are used to encode and bill administrated drugs in France.…”