AbbVie set the ambitious task to simultaneously integrate DHT processes into existing policies. Learn how AbbVie’s DHT Task Force developed unified processes to scale innovation across the enterprise.

AI technologies are advancing at an unprecedented pace. To fully harness their potential in clinical development, we are implementing a holistic digital strategy that seamlessly integrates Generative AI and Machine Learning into the end-to-end clinical trial data flow. This unified clinical platform enhances operational efficiency, reduces manual intervention, and speeds up data-driven decision-making, unlocking significant business value in the complex, highly regulated clinical development landscape.

This talk will share Genentech’s first end-to-end experience operationalizing a digital cough endpoint in an interventional clinical trial. We will share results from a phase 2a chronic cough trial (NCT05660850) that deployed a novel AI-enabled cough monitor alongside the current gold standard. We will present key scientific and operational learnings from continuous, 24/7 monitoring over weeks compared to the traditional single-day snapshots.

In this presentation, we will discuss advances in wearable monitoring for bowel movements and bowel habits. This area is relevant as digital endpoints for GI conditions such as inflammatory bowel disease.

This presentation will showcase the analytical validation of two DHTs: the GENEActiv wrist band with Philips sleep and scratch algorithms and the Emerald radio frequency touchless sensor to measure nocturnal scratching in adults with AD.

The integration of Digital Health Technologies (DHTs), including mobile and wearable sensors, enables the collection of rich, real-world data in clinical trials. However, the traditional approach of developing bespoke applications for each study, leads to significant redundancy, high development costs, and extended timelines. To address these operational inefficiencies, we have strategically pivoted to a unified, Generalizable Application Platform. This presentation details the development and operationalization of this platform, designed to streamline the acquisition of diverse datasets, such as Patient-Reported Outcomes (PROs), voice recordings, and inertial sensor data, within a single infrastructure. We will explore the platform’s architecture, which is built on core principles of modularity, configurability, and extensibility. This design allows the same core application to be rapidly adapted and deployed across the entire development lifecycle, from Phase 0 early clinical studies to Phase 4 market studies. The system comprises a patient-facing mobile app for data collection and a site-facing web application integrated with existing clinical trial management infrastructure, ensuring seamless participant management and the linking of clinical and DHT data. Furthermore, we will share insights from the deployment in a clinical study, outlining the standardized operational workflow, Configuration (Web and Mobile), End-to-End Testing, UAT, and deployment. This generalized strategy demonstrates a significant reduction in development costs and accelerates the deployment of digital tools in clinical research.

This talk examines the process of selecting endpoints for mobility and activity in clinical studies. It emphasizes balancing statistical sensitivity with patient relevance to ensure meaningful outcomes.

This work explores how digital health technologies and novel digital endpoints can objectively measure physical activity and functional changes in cancer patients and cancer patients with cachexia. The findings support the use of these digital measures in clinical trials, providing more patient-centric assessments of treatment effectiveness in cancer cachexia.

The Tufts CSDD team has compiled recently published empirical evidence on investigative site use and experience with decentralized clinical trial components including their impact on clinical trial cycle times and enrollment performance. Quantitative data has been gathered and aggregated across several primary areas including technology adoption, workforce changes, IT infrastructure, and patient and staff satisfaction. This session will review evidence across these domains and summarize key takeaways and their implications.

This study examines the value of decentralized clinical trial (DCT) and digital solutions across global investigative sites. Findings show sites are independently investing in DCTs, signaling a need for sponsors to adapt partnership strategies. The research assesses current and emerging solutions for their ability to deliver richer insights, faster data access, improved patient participation, system integration, and enhanced data quality and integrity.

This initiative team developed a comprehensive digital toolkit (DHT Tubestop) that streamlines the integration and implementation of digital health technologies in clinical trials, enabling optimized trial design, improved data collection modalities, enhanced patient engagement, and more efficient trial operations while ensuring compliance with regulatory requirements and data security standards.

Current methods for informed consent are inadequate, requiring substantial staff availability and using lengthy, non-informative documents, leading to poor understanding and engagement. Advances in technology, particularly generative AI, can drive the transformation of consent processes. We have developed and evaluated an AI agent driven consent process, for informed consent, which has potential to enhance transparency, understanding, and accessibility to clinical trials.

We present a systems-focused pipeline that converts population-scale biosensor-based actigraphy and PPG data (i.e., biosensor data) into reproducible, high-throughput training datasets for multimodal digital health foundation models. We implement a two-stage design: preprocessing and chunking the time series into binned window sizes using scalable EKS clusters, followed by distributed multi-GPU transformer training with a GPU-aware DataLoader and a time-series database to reduce I/O contention, sustain higher GPU utilization, and enable fast metadata-driven retrieval across multi-terabyte storage. In brief, this work emphasizes the infrastructure and data engineering trade-offs, as well as the operational patterns required to convert siloed time series into reproducible, high-throughput training pipelines—independent of model architecture or downstream performance metrics.