Optimizing Patient Matching, Recruitment And Retention In Clinical Trials 1 Azim Olatinwo & Andres Dumas There is a growing demand for healthcare services Image by rawpixel.com on Freepik 2 Canadian Healthcare Demands 1950 2100 1980 2000 2020 2040 2060 2080 0 10 20 30 40 20% 25% 2030 3 Median Age Optimizing Clinical Trials to Improve Healthcare Image by rawpixel.com on Freepik 4 Fixed Clinical Trials Location Centered Trials In conventional clinical trials, all factors of the study design are set in stone before starting... Conventional clinical trials are c onducted without addressing the patient’s convenience or needs Sunnybrook Research Institute 5 Patient Matching and Recruitment One in every five clinical trials does not complete its enrollment of participants Patient Retention Pre - screen Consent Screening Dropouts 31 Pre - screen qualified 13 Consented 9 Randomized 7 Completed trial 100 Patients Clinical Performance Partners, Inc. 6 Targeting Rare Diseases There are rare diseases Affect people <200,000 >7,000 NORD Rare Disease Database 7 Creating a smart solution in a smart world Image by rawpixel.com on Freepik 8 Patient Matching Patient Retention Patient Recruitment A Patient - Centric Solution 9 A Patient - Centric Solution 10 Patient Matching • Uses Natural Language Processing & Optical Character Recognition • Data from Electronic Medical Records and other sources A Patient - Centric Solution 11 Patient Recruitment • Identifies most effective communication • Tailored recruitment messages A Patient - Centric Solution 12 Patient Retention • Identifies factors affecting patient engagement • Demographics • Wealth/financial status • Treatment Adherence • Closer monitoring and intervention Patient Recruitment and Selection Traditional Approach Manual Screening Time - consuming Subjectivity and Bias Inclusion/Exclusion Criteria Inflexible Incomplete patient information AI - based Approach Electronic Health Records Real - time Data Comprehensive patient information Machine Learning Algorithms Data Analysis Identify patterns Classify patients Predictive Modelling Estimate Patient Response Improve Protocol Design 13 Extract Analyze and Connect Optical Character Recognition (OCR) Natural Language Processing (NLP) Machine Learning (ML) Patient matching Patient matching Unlock your unstructured clinical data 14 Recruitment module 100% PRECISION IN PHI PRIVACY Improved Data Quality Reduced Costs Faster Results Improved Patient Outcomes Enhanced Patient Satisfaction Benefits 15 Challenges ▪ Changes to data privacy and security regulations (PIPEDA AND PHIPA) ▪ Health Canada Regulatory Approval ▪ Resistance to Change ▪ Need for Additional Resources Limitations ▪ Reliance on Technology ▪ AI Technology Limitations Exubera’s Plan ▪ Work with regulatory bodies & HCPs ▪ Invest in AI research 16 Challenges and Limitations Testing our Solution Image by rawpixel.com on Freepik 17 Objectives ✓ Accuracy ✓ Reliability ✓ Safety Test Scenarios 1. Matching 2. Recruitment 3. Retainment Test Data Testing Approach Success Criteria ▪ Matching ▪ Exceed Recruitment target ▪ Dropout rates ▪ Privacy compliance 18 >95% 50% >99% h ttps://exubera.webflow.io 19 Learn more 20 h ttps://exubera.webflow.io