5. Amil Khanzada Anja Zehfuss Ayesha Fraser Alex Kalanda Mimi Dunne, MD
Speciality
Artificial Intelligence
Epidemiology
Infectious Disease
Global Health
Regulatory Affairs
and Quality
Operations
Artificial Intelligence
Emergency Medicine
Palliative Medicine
Role
Founder / Chief
Technical Officer
Chief Operating
Officer
Chief Clinical Officer
Head of Outreach:
Africa
Chief Medical Officer
Value to Team One Young World
ambassador
AI & software
engineering
Mobilizing global
teams
Stanford BS, Human
Biology
Epidemiology
GSB TA
Clinical Research
Regulatory Affairs
and Quality
Assurance
Operations
Entrepreneurship and
Innovation
Fellow, Stanford DCI
NY Covid-19 support
6.
7. Post H4R knowledge gained
• Nonprofit with a for-profit arm
• Need to be targeted (not “everyone”)
• Clarified beneficiaries and channels
• Solidified our value proposition
• “Pre-check” is our niche
• Contact list grew significantly
• Grassroots & industry before FDA
• Interviews reflect user enthusiasm
• Learn process to be entrepreneur / built a team
• The world is waiting for our product!
8. Pre-H4R thoughts
• Goals:
• Targeting a wide range national and international partners is the
best approach
• Build an FDA-approved diagnostic medical app
• Mission model canvas encapsulates our vision
• Public health officials and medical professionals are our primary
contacts/partners
9. KEY PARTNERS
- Public health system
- Stanford
- One Young World
KEY RESOURCES
- clinical data
- outreach team
- AI / app developers
- network & partnerships
- legal team
VALUE PROPOSITIONS
provide app to:
- easily, instant,
non-invasive
diagnose COVID
- limits virus spread
- restore mobility
- reopen business
- reassurance
- return to normal
- offline access
KEY ACTIVITIES
- Data collection / outreach
- AI & app development
- Regulatory compliance
- Grants / fundraising
- Legal entity / coverage
- Recruitment / HR
MISSION ACHIEVEMENT/IMPACT FACTORS
- AI performs at precision levels acceptable for FDA approval as
medical device
- Technology is easy to adopt and use
Mission/Problem Description: make
everyone’s smartphone an instant
and easy COVID testing kit
Designed by: The H4Di Team Date: 6/27/2020
Day 1
DEPLOYMENT
- website app
- mobile app store
- social media
- institutional rollouts
- integrate w/ telehealth
BUY-IN & SUPPORT
- Public Health System
-Healthcare professionals
- Hospitals and long-term
care facilities
- Community members
MISSION BUDGET/COST
- Legal fees (entity registration, regulatory compliance, data security)
- Data collection costs (PPE), IT (servers, GPUs, licenses)
- Salaries / stipends
- Budget in: donations, grants, licensing fee to hospitals/companies, app cost,
subscription cost (free)
BENEFICIARIES
- government
- public health
departments
everyone else
- people living in COVID
hotspots, limited
medical facilities,
dense population
- colleges,
corporations, hospitals
10. KEY PARTNERS
- Massive Global
Foundation
- African Public Health
Authority
- Stanford H4R SMEs
- One Young World
- 12,000+ ambassadors
- 190+ countries
KEY RESOURCES
- Clinical data
- AI / app developers
- Partners
VALUE PROPOSITIONS
- COVID ”pre-check”
- Free, fast, non-invasive
COVID self-testing
- PCR level confidence
for FDA emergency use
authorization
KEY ACTIVITIES
- Data collection
- clinical
- crowdsourcing
- AI & app development
MISSION ACHIEVEMENT/IMPACT FACTORS
- Deployed for testing in shortage areas
- Deployed as “pre-check” for industry and individuals
- AI refines to PCR precision level
Mission/Problem Description: make
everyone’s smartphone an instant
and easy COVID testing kit
Designed by: The H4Di Team Date: 7/3/2020
DEPLOYMENT
- Website app
- Mobile app
- Social media
BUY-IN & SUPPORT
- African PHS / NGOs
- US HCP Management
- Donors
MISSION BUDGET/COST
- Patent, trademark, FDA approval, legal fees (pro bono), IT (servers,
licenses), Salaries / stipends
Budget in:
- donations, grants, subscription / licensing fee for organizations
BENEFICIARIES
- LMIC (low/middle
income countries)
- Health system
- Employers
- Military
- Schools
- US (after FDA approval)
- Research community
11. We improved our AI technology
• Improved cough ML models
• Promising results on clinical data
• 80%+ accuracy
• 70%+ sensitivity / specificity
• 70/30 train/test split
• 20% false positive rate
• 36% false negative rate
12. We built our MVP to test our hypothesis and collect data
13. Our specificity has global applications
BUT...
● Employer and personal tests may have lower standards
● Far better than temperature checks
● FDA scrutiny highly likely
● Focus outside US until FDA-ready
FDA PCR
requires 95%
specificity
14. We collected initial data sets
300+
Clinical samples taken at PCR test site
25%
COVID-19 positive clinical samples
500
Crowdsourced data (largely COVID-19
negative or not tested)
15. What We Do Next
• Gather Diverse Data
• Crowdsourcing coughs globally
• Clinical research studies
• Refine Technology
• Show better metrics on existing clinical data
• Develop mobile app and backend server
• Solidify Partnerships
• County health departments
• International CDC task forces
• Seeking funding, partnerships, and mobile/AI engineers
16. Social good DNA
My story… never give up!
CS230 with Andrew Ng One Young World