2016 Scope david cocker
Download as PPTX, PDF0 likes29 views
The document discusses advanced feasibility modelling for predicting the success of clinical trials. It describes using data-driven algorithms and semantic analysis of large datasets to answer business questions. New tools are presented for semantically filtering trials by disease and inclusion/exclusion criteria. Case studies analyze trends in new clinical trial sites opened from 2006-2015 for oncology and cardiovascular indications.
2016 Scope david cocker
- 1. Advanced Feasibility Modelling Session: 23rd Feb. Data-Driven Country Feasibility and Site Performance Prediction Presenter: David J Cocker Chief Scientific Officer MDCPartners david.cocker.@mdcpartners.be
- 2. My talk today Useful visuals ‘Less bling more info’ Easy to understand Understanding the data you see Data sources Get to the value fasterDoing more NewOld Meta statistics Algorithms based on looking back Very Old Volume
- 5. Network of event horizons Planning v v v v v v v v v v vv v v Commonality of process vs “I’ll do it my way”
- 6. People think lots of data sources is good
- 7. BUSINESS QUESTIONS Translating silos of data into quantifiable information Semantic Transformation Creating ontologies to rank and classify medical concepts
- 8. Introduction Let’s start by agreeing that feasibility is a confidence prediction in your downstream research network.
- 11. A range of robust algorithms to get to your data fast All reports are exportable via URL and kept active for one month
- 12. A range of robust algorithms to get your data fast
- 13. New tools to semantically filter on trials and segment in/exclusion
- 14. New tools to semantically filter on trials and segment in/exclusionDiseases Excl-drug term indication Inclusion term drug
- 16. BUSINESS QUESTIONS Epidemiology Patient group discovery tools and options After 8 years of development, our technology stack is sufficiently robust to expand to public and private data assessment. From this private data repository, new dashboards and insights can be rendered.
- 17. BUSINESS QUESTIONS Epidemiology New tools to semantically filter on trials and segment in/exclusion 600,000 words grouped by protocol concepts
- 18. Looking back at past results Can we learn from our mistakes?
- 19. Opening of new sites: case study 1 2014 (57) 2015 (128) • 4 companies (Merck, Lilly, JNJ, Pfizer) • Period 2006-----2015 • Sites opened in cardiovascular indications (hypertension, hyperlipidemia, atherosclerosis, diabetes) • Sites opened in immunological factors in oncology
- 21. Immuno-Oncology new sites/trial 4.3 New sites/total sites opened 0.185 Sites/trial 23 In this new therapeutic area, on average, 4 out of 23 sites are new sites which were not used by the company before (18.5%) 1685/5,956 US sites 27%
- 23. Cardiovascular trial analysis 4 major companies in the space In these CV therapeutic areas, on average there are 3 new and 22 historic sites per trial. (11.6% new) Trials 1,549 Average new site openings Migrating from metabolic to oncology Representing 24% site load of this therapeutic area Diabetes Lipid lowering 10,000/22,000 US sites 47%
- 25. To concepts From text he illusion of just counting words