Regulatory Issues on Digital Medicine

1. Regulatory Issueson Digital Medicine Luca Pani,MD Professor of Clinical Psychiatry, University of Miami,USA Professor of Pharmacology, University of Modena and Reggio Emilia VP for Regulatory Strategy and Market Access Innovation, VeraSci,USA Former Director General Italian Medicines Agency (AIFA),Rome Former CHMP and SAWP Member, European Medicines Agency (EMA),London

2. Disclaimer andDisclosure The opinions expressed in this presentation are my personal views and may not be understood or quoted as being made on behalf of or reflecting the position of any of the Institutions or Companies for which I have worked or I collaborate with. The mention of commercial products, their sources, or their use in connection with material reported herein is not to be constructed as either an actual or implied endorsement of such products of any Public Department or Health and/or PayerServices. Apart from my Academic roles, I am the Chief Scientific Officer of EDRA-LSWR Publishing Company and of Inpeco SA Total Lab Automation Company. In the last year I have been a scientific consultant to Acadia USA, Ferrer Spain, Johnson & Johnson USA, VeraSci USA, Otsuka USA, Pfizer Global USA, PharmaMarSpain. I do not bear any direct or indirect financial interest in products quoted in thistalk. These slides are both original or have been modified from presentations/videos at other meetings. Acknowledging: S.Chassang,V.Mantua,E.Snowberg,R.Triunfo. This presentation is updated to March 10th2019.

3. Regulators have Written Extensively on theIssue… https://www.ema.europa.eu/en/news/first-guidance-new-rules-certain-medical-devices https://www.ema.europa.eu/en/news/role-big-data-evaluation-supervision-medicines-eu https://www.ema.europa.eu/en/documents/regulatory-procedural-guideline/esource-direct-data-capture-ddc-qualification-opinion_en.pdf https://www.fda.gov/downloads/MedicalDevices/DigitalHealth/UCM568735.pdf https://www.fda.gov/MedicalDevices/DigitalHealth/ucm562577.htm https://www.fda.gov/downloads/RegulatoryInformation/Guidances/UCM630458.pdf

4. Background RegulatoryConsiderations • Some digital medicine considerations with regards to validation and potential for productlabeling? • Like any measure, basic metrics and testing considerations suchas: • Reliability • Validity • Practicality • Data analytics, extraction, and integrityprocesses • Is the biomarker/device measuring the construct that you think it ismeasuring? • Consider translational (does it fit with animal models?) • Validation and Certification of measures will be necessary for Regulators /Payers • FDA/EMA could request to issue a qualification opinion and then use your biomarker to enrich apopulation • These surrogate and secondary endpoint measures should be linked with primary orco-primary • Product LabelingConsideration • Potential to replace Phase IV studies with patient centered outcomes (see regulatory / payer considerations above) • Example-Improves social / functional activity in individuals withdepression

5. Digital Medicine Regulatory Comments (1 of3) • Selecting the right device for the rightendpoint • Start and end with theScience • DeviceFeatures • Form factor/Durability/Battery • Connectivity • Wifi/Bluetooth • Data andMetrics • Sensitive, Reliable, Clinical Relevance,Accuracy • Scalability • Deployment ease across large populations and differentcountries • PatientExperience • Usefulness/Burden/Adherence and technologybias

6. Digital Medicine Regulatory Comments (2 of3) • Regulators and Sponsors Looking for: • Better datacapture • Enhanced patient experience(automation) • Efficient-Data structured forsubmission • Real time transparency • Data Challenges • Datavolume-Sponsorsnotequippedtoreceivemoredatafromonepatient actigraphy than previous entirestudies • Continuous data-hard to determine a defined variable, or definiteendpoint

7. Digital Medicine Regulatory Comments (3 of3) • Unique and Unequivocal (100%) Biometrical subjectauthentication • Management of exponential increase in volume of datacollected • Lack of common datastandards • Consumer and even FDA approved devices could be plagued with many data reliability issues / missing data • Retention/adherence issues, subjects neglecting towear/charge • Subject Device training and correct use monitoring without overburdening the subject • Support:Need to make sure solutions don’t require the site personnel to be “tech people” • Variability in homeenvironments • Rate of change in consumertechnology • Privacy,Legal

8. Reconciling Big Data and Privacy inEurope • How good faith legal safeguards could de facto make Europe noncompetitive • The new EU General Data Protection Regulation has entered into full force on May 25th,2018 • Driven by the principle of dataminimisation • Privacy by design anddefault • Right to optout • Informedconsent • Data ownership – personal / public /private • Need access to a sufficient amount of “good qualitydata”

9. Reconciling Big Data and Privacy inEurope • How good faith legal safeguards could de facto make Europe non competitive • The new EU General Data Protection Regulation has entered into full force on May 25th,2018 • Driven by the principle of data minimisation (it will preclude machinelearning) • Privacy by design and default (it will preclude predictive analytics) • Right to opt out (but how? ) • Informed (really informed?) consent (impossible to predict to what I am giving consent to) • Data ownership – personal / public / private (issue is not on ownership but on access) • Need access to a sufficient amount of “good quality data” (indeed impossible with limits above) • EU Regulators / Payers will have these additionalproblems: • Not enough (sometime none) competence with in-house and hands-onskills • Education of new types of assessors with very broad data science and life scienceknowledge. • Inability to certify and validate different data sources to be integrated amongthem. • Rule the emerging strong engagement by patients as datagenerators.

10. How to Capture and Elaborate Data from aDevice? • Use this kind of”stuff” • Adoption ofA.I. • Machine Learning • Integration of structured and unstructureddata • Applying an iterative approach • Use real life certified database • Potential pitfalls • Junk in / Junkout • Missing data • Unrealistic values • Not preparing the data for further elaboration /submission • What we have vs. what is coming • Sensors for “activity”monitor • Pattern recognition andanalysis • Human to machineinterface

11. High-PerformanceMedicine Where Human and Artificial IntelligenceConverge Topol E., Nature Medicine, Vol 25, Jan 2019,44-56

12. Will A.I. Ever Replace Human Health CareProviders? Topol E., Nature Medicine, Vol 25, Jan 2019,44-56

13. Conclusions • The most important innovation disrupter: i.e. the ICT impact on an expanding set of end-user devices is here now • This will challenge and force all of us to rethink how data-information-knowledge transitions are being created andused • Business models will evolve (fast) thanks to smart machinetechnologies • We must however augment perimeter defense for privacy and rule-based security detection with user and entity behavioranalytics • Datagathering,elaborationandanalysismethodswillevolveandthedifference,ifany, • between patients and consumers will fadeaway. • The bigger the Data the bigger the Biology as we will move from inferential to representativeknowledge.However,Data ≇Information≇Knowledge • Now, you have to ask yourself a very fundamental question: are we all reallyready • for this? • ≇Neither approximately nor actually equal to