VIDAL partners with ARCASCIENCE

VIDAL creates and delivers health product databases that empower medical decisions.

Integrated into medical software vendor applications, VIDAL solutions help to improve safety, efficiency and outcomes. By a combination of medical knowledge and technology, VIDAL contributes to minimize preventable errors and adverse drug events. On a day to day basis, the VIDAL team works with healthcare professionals to meet their needs at the point of care.

Based on VIDAL’s knowledge base, VIDAL Drug Information Systems (DIS) are designed and used internationally.

To achieve the challenge of its international expansion, and quality on the French market, DDI is one of the key data to secure prescriptions. So VIDAL would like to provide references to users for each DDI described in his database, enriching the information beyond that usually provided for DDI.

Detecting DDI, completing scientific literature that goes with it and preventing adverse effects is complex. The huge number of drug-drug interactions requires the use of reference works or databases. Moreover, there is considerable variation in the DDIs included in the different international databases. In addition, unconfirmed interactions are often documented which is not clinically helpful. The aim of the project is to get a targeted and automated bibliographic monitoring service to be able to update DDI data and reduce drug-related medical risk.

The solution explored together for enhancing drug-drug interaction detection in scientific literature integrates multiple data sources, advanced AI models, and an expert-built scoring system to address the limitations of traditional static databases.

Based on articles published in Pubmed, Clinicaltrials.gov & Medline, AI models prioritize high-evidence publications (stratification model), classify DDI-related content (NLP), assess the safety implications of detected interactions (NLP), and relate these DDIs to various patient characteristics (NER -Named Entity Extraction).

A scoring system, developed through input from Vidal’s experts, ranks articles based on relevance, impact, and clinical importance, ensuring the highest clinical relevance and accuracy standards.

This expert-driven scoring process is validated through a gold standard and sensitivity analysis, ensuring robustness and reliability. Compared to existing systems, this solution offers greater flexibility by generating various types of evidence, significantly improving DDI analysis with comprehensive coverage and minimal missed signal, thanks to the expert insights from VIDAL.

As a result of this experimentation period: ArcaScience AI algorithm proved its accuracy for DDI that are well known and documented in France, and its ability to find new documented and consistent DDI on the international market. Three months after its launch, VIDAL is now ready to accelerate its international DDI management.

This solution combines ArcaScience’s expertise in AI model development and its application to scientific literature with VIDAL’s deep understanding of drug-drug interactions and their clinical implications, creating a powerful tool that enhances DDI screening.