Collaboration among academia, non-profits, and industry are essential for the creation of disease-specific pre-competitive informational resources for applying AI/ML models in identification of disease biomarkers, patient subpopulations, and creating disease progression and disease staging models. To improve the Findability, Accessibility, Interoperability, and Reuse (FAIR) of digital knowledge, i.e. introduction of precision research, the clinical research community shall agree to operate under similar operational guidelines, from regulatory/legal, to patient identification, to data standards.

Preclinical and clinical trial data form the foundation of biomedical and regulatory science.  Typically, invitro preclinical tests indicate the potential potency and metabolic stability of a given substance and potential drug-drug interactions that should be avoided in initial clinical trials. Over the past five years the FDA and NCATS/NIH developed the Global Substance Registration System (GSRS) which defines substances in compliance with ISO IDMP standard. The system links active substances to products, targets, preclinical data, and clinical trials.  The talk will focus on the role the GSRS can play in the digitization of regulatory information.

Although submission packages are mostly composed of documents, health authorities increasingly request access to structured data. We present here our “Meaning-based Computing” (MBC) initiative and related digital products that improve knowledge extraction and integration from multiple and heterogeneous sources. Fetching knowledge from regulatory and manufacturing artefacts, we identify, categorise, and expose key knowledge elements (GMP process deviations, expert skills, regulatory risks) and expose them to solve concrete problems and secondary uses. MBC uses state-of-the art AI/NLP technology and agile/scrum product development to FAIRify our digital environment to fasten and funnel knowledge flow where it is needed most.

The explosive growth of genomic-assisted clinical trials provides great opportunity to investigate the mechanism of drug response and for personalized medicine. However, challenges emerge due to the heterogeneity of patient population and complexity of patient data. The talk will discuss approaches on how we (1) dive into deep data mining to stratify patient population and (2) apply AI-powered research engine to build explainable prediction models, which leads to novel insights into drug response mechanisms.

We made use of Social Media Listening to identify patient subpopulations with dry eye disease in a disease-specific forum. We also extracted the reported symptoms and preferred treatments to amplify the voice of the patient, supporting the discovery of potential problems with current treatments, the detection of unmet medical needs, and the deeper understanding of the disease burden and its effects on everyday life.

Clinical study reports (CSRs) are long, detailed documents that combine study information contained in the clinical protocol with the data, results, and analyses from the study. In turn, CSRs serve as a basis for more concise manuscripts that eventually get submitted for publication in scientific journals. Both of these steps are time-consuming and expensive. We present two ongoing projects that leverage AI to guide and accelerate these steps.

Physicians and scientists are increasingly leveraging digital health capabilities to enhance disease understanding, prevention, diagnosis, and treatment and to improve patient outcome through more targeted healthcare delivery. Likewise, patients are increasingly using digital health capabilities to manage their health and to make informed decisions about their healthcare options. This talk will provide examples of how digital technologies are evolving and transforming clinical research and practice to deliver precision medicine.

There has been a substantial increase in the use of digital wearables and biomarkers in decentralized clinical trials (DCTs). DCTs may be more appealing to patients as they can decrease the need for travel to clinical sites, leading to improved recruitment and retention, and may also allow for clinical trials in indications previously thought to be unfeasible. Here we present a digitally-enabled DCT example along with technical and operational considerations.

Pharmaceutical sponsors are deploying digital health tools in clinical development more than ever. These tools are incredibly helpful in acquiring real-time quality patient data. However, sites face an additional burden due to these technologies. The session will talk about the approach to balance the site burden.

Atopic dermatitis is often accompanied by unrelenting nighttime pruritus/scratching and sleep disruptions. Quantitatively evaluating nighttime scratch and sleep via accelerometry using digital wearables to monitor patients passively and continuously would provide insight into the disease and effectiveness of treatments. Key aspects of providing value in these assessments are that the methods and devices are vetted and verified to detect clinically meaningful changes.