- Setting the Scene
Emerging technologies like Electronic Health Records (EHRs) and AI are set to revolutionize clinical research. Federated EHR platforms, supported by regulatory frameworks, enable collaboration with Healthcare Organizations (HCOs), enhancing Real-World Evidence (RWE). Advances in AI and Large Language Models (LLMs), fueled by computational power and digital data, drive innovation across industries, including healthcare. Integrating Real-World Data (RWD) and AI links healthcare delivery with pharmaceutical research, improving our understanding of drug performance in varied settings. High-quality, accurate data is crucial for reliable conclusions, impacting drug development and patient care.
- Impact Areas of RWD & AI Services in the Pharma Industry
Data science and associated technologies significantly impact pharmaceutical R&D, particularly in clinical trials, which account for 60% of the cost and 70% of the time required for a new drug to reach patients. The high cost of drug development, around $2.5 billion, is partly due to the need for large clinical trials that provide definitive evidence of efficacy and safety. These trials also fulfill regulatory requirements and generate value-based evidence, which may require real-world studies to evaluate comparative effectiveness, safety, and cost-effectiveness.
Data science technologies can reshape the industry by addressing bottlenecks such as difficulties in evaluating patient populations, identifying suitable patients for enrollment, optimizing protocols, and reducing manual and redundant data entry. For example, up to 70% of data in Randomized Clinical Trials (RCTs) is currently duplicated from health records, and verifying source data accounts for 20% of total study costs.
- Federated EHR Platforms: A Synergy of Services
Federated EHR research platform technology (FED EHR) can unlock trustworthy reuse of EHR data on a large scale. Developed within the European Innovative Medicine Initiative (IMI) EHR4CR project, FED EHR provides real-time services such as study design, feasibility, and patient recruitment across a growing network of hospitals.
Local hospitals communicate with a central platform using de-identified aggregated data derived from their EHRs. Data processing is done locally at the hospital level. Local ETL servers connect various data sources and regularly update structured EHR data (ICD-10, SNOMED CT, HL7, LOINC) within the hospital. A key GDPR principle is that no patient-level data leaves the hospital, ensuring data remains fully controlled by the hospitals.
- Impact on the Pharmaceutical Industry
For the pharmaceutical industry, key impact areas include real-time data access, enhanced collaboration with healthcare providers, and cost-effective data exchange. FED EHR technology offers a win-win opportunity: hospitals retain control of their data while benefiting from faster clinical setup, quicker recruitment, reduced site burden, improved quality, and consolidated EHR access. FED EHR platforms like Deep6, Flatiron, NextGen, Health Verity, and TriNetX can revolutionize clinical trials, offering collaborative services that yield regulator-worthy data, save money, and expedite decisions and time-to-market. They also support AI-enabled analytics and distributed machine learning. Major vendors connect over 200 million EHRs across 1,800 hospital sites to central platform services.
- Two Use Cases
Enhancing Study Design and Patient Recruitment. FED EHRsignificantly enhance clinical trial design by leveraging real-time data from a network of hospitals. Pharmaceutical companies can efficiently identify and recruit suitable patients, reducing recruitment time and cost while ensuring diverse patient populations for reliable trial outcomes. Sponsors can access the central platform to translate study protocols into EHR data concepts and run real-time feasibility queries across millions of EHRs. What once took 3-6 months can now be achieved in 1-2 days, including multiple protocol iterations. This technology promises to radically boost recruitment through hospitals connected to the FED EHR network using local EHR-enabled recruitment.
EHR2EDC or eSource Supported Trials involves moving EHR data to an EDC system for RCTs. This process includes quality assurance and adherence to regulatory eSource guidelines, with patient consent and without replacing manual data entry or causing data redundancy. The EHR data utilized typically comprises structured data or coding standards such as laboratory results, vital signs, medications, diagnoses, and patient demographics. Clinical trial data collection has become increasingly complex, with extensive data duplication across systems like EDC and hospital EHRs.
Manual data entry and verification processes are time-consuming and resource-intensive. Introducing EHR2EDC technology can reduce the total time per data point by +50%, offering significant value generation for investigational sites. The eSource approach improves clinical trial conduct by automating data transfer from existing EHR to EDC systems, reducing investigator site burden, and enhancing data quality. This automation minimizes transcription errors, improves data accuracy, and streamlines source data verification (SDV), leading to cost savings and faster trial timelines.
- Data First, AI Later
The success of digital health initiatives relies on data integrity and integration. High-quality, structured health data remains paramount as a prerequisite for AI advancements. Emphasizing the importance of the “3i’s of data”—Integrity, Integration, and Intelligence (AI)—is crucial for establishing a foundation for effective AI development. Limited data literacy among policymakers poses a barrier to effective AI use. Better education and awareness are needed to facilitate responsible AI deployment. Privacy and security concerns must be addressed early to prevent risks associated with implementing AI prematurely without robust data frameworks. Harmonizing data privacy laws between regions is essential for the effective management of health data. Future convergence of regulatory standards holds promise for streamlined data sharing and collaboration across borders.
- Conclusion & Outlook
The proliferation of sophisticated EHR systems is enriching the pool of structured health data, bolstered by national and EU/US investments in shared research infrastructures, enhancing learning health systems. FED EHR technology is reshaping data management, offering real-time access and revolutionizing traditional data ownership and storage models. Utilizing FED EHR services can streamline clinical trial designs, amplify recruitment, and expedite trial execution with eSource, leading to marked reductions in cost and administrative load. High-quality, structured health data remains paramount for AI advancements, emphasizing a ‘Data First, AI Later’ strategy to establish a foundation for effective AI development. Enhanced EHRs, coupled with robust AI applications, promise a transformative impact on clinical trials and overall healthcare outcomes.
- References
Kalankesh L.R, et al. (2024) Utilization of EHRs for clinical trials: a systematic review. BMC Medical Research Methodology BMC Medical Research Methodology Vol 24:70.
Monira S. (2024), “Possibilities and Challenges for Digital Medicine in Oncology Clinical Trials” Clinical Researcher, Vol 37: 4.
Sundgren M (2024) Keynote presentation on Revolutionizing Clinical Trials through RWD & AI: Unleashing Innovation. Conference: 14th Annual Outsourcing in Clinical Trials Europe, Barcelona May 7-8th, 2024. DOI: 10.13140/RG.2.2.32465.34405
Parab A, et al. (2020) Accelerating the Adoption of eSource in Clinical Research: A Transcelerate Point of View. Ther Innov Regul Sci.
Sundgren M, et al. (2023) eSource Interoperability Between EHR and EDC Applied Clinical Trials, July/August
Beresniak, A., et al. (2015) Cost-benefit assessment of using electronic health records data for clinical research versus current practices: Contribution of the Electronic Health Records for Clinical Research (EHR4CR) European Project. Contemp Clin Trials.
- Biographical note
Mats Sundgren, PhD, MSc, is a distinguished expert in Health Data Strategy with over 37 years in the pharmaceutical industry. His diverse contributions span Discovery, Development, Manufacturing, IT R&D, Patents, Clinical Science, and Data Science & AI. After a 12-year tenure as the global lead for EHR services at AstraZeneca, Mats now serves in several key roles: Senior Industry Scientific Director and co-founder of i~HD, Executive Strategic Advisor for IgniteData, Chairman of the Board at IMIT, and Board Member of the Center for Health Governance at Gothenburg University. In July 2024, he became the Scientific Adviser for IOMED. Mats has over 70 publications, books, and patents, focusing on Health Data Science & AI, Clinical Science, Clinical Trials Management, and Innovation.
