Supporting a Collaborative R&D Organization with a Dynamic Big Data Solution
1 like200 views
The document discusses the conceptual data landscape and the flow of information in clinical development, highlighting various analytic solutions to optimize processes from discovery to post-approval. It emphasizes the challenges of real-time data ingestion, standardization, and the complexity of managing diverse data sources. The reflections suggest the need for user-friendly analytics and the importance of preparing information based on typical user inquiries.
Supporting a Collaborative R&D Organization with a Dynamic Big Data Solution
- 1. Copyright © 2016, Saama Technologies | Confidential Supporting a Collaborative R&D Organization with a Dynamic Big Data Solution Big Data & Analytics for Healthcare & Life Sciences Summit 10/18/2017 San Francisco #DataLifeSci
- 2. 2 Copyright © 2017, Saama Technologies NCSU B.S. Biochemistry 2010 Nanoparticle Vaccines Development 2013 Rutgers M.S. Biomedical Informatics 2017/18 Big Data, Informatics, Productized Solutions in LifeSci 2014-? A Little About Me
- 3. 3 Copyright © 2017, Saama Technologies What I’m Sharing Today Conceptual data landscape and flow of information (and challenges) Various analytic solutions and how they are built for various objectives Using the right solution(s) and the results (better patients, better business) Reflect on learnings and possible innovations Demo
- 4. 4 Copyright © 2017, Saama Technologies Clinical Development: Translational Continuum Discovery Phase 1 Phase 2 Phase 3 Approval Post- Approval Basic Science Discovery Proposed Human Application Proven Clinical Application Clinical Practice Public Health Impact Chasm 1 (T1) Chasm 2 (T2) Chasm 3 (T3) Chasm 4 (T4) Bench to Clinical Development Safety & Efficacy Research Implementation & Adoption Population Health Research Drolet, Brian C., and Nancy M. Lorenzi. "Translational research: understanding the continuum from bench to bedside." Translational Research 157.1 (2011): 1-5.
- 5. 5 Copyright © 2017, Saama Technologies Clinical Development: Data Flow Discovery Phase 1 Phase 2 Phase 3 Approval Post- Approval Pre-Clinical Non-Clinical Trans. Med. In Silico In Vitro In Vivo Animal Studies DMPK Clinical Science Clinical Operations Modeling Protocol/Study Design Feasibility Regulatory Clinical Science Clinical Operations Biostatistics Monitor/Tune Trial Management Study Report Epidemiology Medical Affairs Pharmacovigilance IST’s/IIT’s Safety Signals CTMS, EDC, IRT, TMF... Argus, Real WorldFiles/DBs...
- 6. 6 Copyright © 2017, Saama Technologies Clinical Development: Challenges Discovery Phase 1 Phase 2 Phase 3 Approval Post- Approval Intake Real-time data ingestion Connected data integration Standardized data collection Data variety, volume, etc. Process & Use Data Management, Quality, Queries Conform to Standards Coding Adhere to process Collate & Submit Initial - Final Submission Data Data Base Lock Clinical Study Report Monitor & Report Valid Safety Signals Health Economics @ Scale CTMS, EDC, IRT, TMF... Argus, Real WorldFiles/DBs...
- 7. 7 Copyright © 2017, Saama Technologies Unified Cloud Data as a Service with Analytic Applications Clinical Data Safety Data Syndicated & Large Data Other Data Data Sources Enabled Analytics Patient & Studies Analytics Clinical Study Data Mart Clinical Outcomes Analytics Drug Safety & Analytics Safety Outcome & Reporting Analytics Signal Detection Real World Analytics Risk Based Monitoring Electronic Data Capture Clinical Trials Management System CRO Data Labs / Biomarker Safety Data Warehouse Global Safety Data Warehouse ARGUS / ARISg Real World Claims EMR / EHR Omics Data Public Data (Kegg, NCBI, and,etc.) Trials Trove, CT.gov Social Media IOT/ Sensor/ Wearables Non- Clinical & Pre-Clinical Trial Management Analytics
- 8. 8 Copyright © 2017, Saama Technologies Analytic Solutions to Optimize Clinical Development Design for Indication Feasibility of Study Strategy & Monitoring Wisdom • Substantiate adaptive design of studies that are less-prone to failure and translate into characterized subject populations Knowledge • Protocol criteria that isolate ideal subjects that will respond and progress towards endpoints Information • Variants in subjects that influence the degree of their response • Validity of endpoints and diagnostic measures to track Data • Publications, Clinical, Biomarker, Tissue, Lab, Comparators Wisdom • Dynamically assess the feasibility of protocols and sites targeted for the study and translate into specific design (dosing, etc.) Knowledge • Protocol criteria that elicit amendments and complications • Investigators that are best across all parameters Information • Investigator & Site performance • Scoring of protocols and sites based on parameters Data • CT.gov, Knowledge Store, US Claim/EMR, International Administrative, Repositories Wisdom • Go/No-Go Decision making of programs and translation of successful completions to design of next phase. • Pro-active response and resolution to risk-prone entities Knowledge • Studies that contain subjects trending positively towards endpoints and avoiding adverse events (patient level) • Complete view of performance of study (site level) Information • Modeling of Study Success KPI’s, KRI’s Data • Historical Studies, Publications, Clinical Systems, CRO, CT.gov
- 9. 9 Copyright © 2017, Saama Technologies Reflections & Learnings Source data is varied, complex and dirty – High volumes of this will be consumed for training AI-driven bots that clean, standardize, transform and analyze data at scale Over-analysis is paralysis – Seek out to make solutions that are intuitive and don’t require a 200 page training manual and SOP – Least common denomination set of analytics that caters to most users is nirvana – customizations can always be made later Prepare information based on how users typically ask their questions – E.g. screen failure rate a METRIC, not the answer to a question. If the question is “What is causing a high SFR at this site?”, seek to augment SFR with SFR Reason as found in the EDC, IRT, or CRO extract.