MRCT Center - Wellcome Trust Mee2ng on The Future of Clinical Trial - - PowerPoint PPT Presentation

MRCT Center - Wellcome Trust Mee2ng on The Future of Clinical Trial Data Sharing

Monday 21 March 2016

Please Note the Following:

• This mee2ng is being recorded for internal

purposes:

• If you choose to par2cipate in a discussion, you are presumed

to consent to the use of your comments in these recordings

• This mee2ng will be documented by

photography:

• These photographs may be used in MRCT Center marke2ng or

promo2onal materials, including the MRCT Center website and newsleJer

4/5/16 2

Disclaimer:

• The opinions contained therein are those of the

authors and are not intended to represent the posi2on of Brigham and Women's Hospital or Harvard University.

4/5/16 3

Welcome and Introduc2ons

Presenter: Nicola Perrin, Wellcome Trust

4/5/16 4

• 5 -

What might a future model look like?

REPOSITORY

• r

FEDERATED PORTAL DATA IN DATA OUT

• 6 -

What might a future model look like?

REPOSITORY

• r

FEDERATED PORTAL DATA IN DATA OUT What data, from where? What is the access model?

Open access / reviewed access / stricter controls

Can data be downloaded? Who operates this? What is the governance model? Who pays?

Is data deposited, or kept by the data generator until requested?

Is there some resource to help with curation and de-identification? How does an IRP operate?

Proposed Model PlaUorm - straw man

4/5/16 7

Central mul2- stakeholder governance

• rganiza2on

Central repository for academics (or

• thers) who do

not wish to “host” data Provides shared services for:

• Administer

researcher requests

• Review process
• De-iden2fica2on
• Se_ng policies
• Define standards

PORTAL: Central user interface portal with search engine building upon exis2ng search engines (e.g. ClinicalTrials.gov and ICTRP) to pull informa2on from registries / provide complete and robust “denominator” of exis2ng data PLATFORM: federated plaUorm model with op2onal central component enabling access to data, combining datasets and allowing downloading as appropriate

A

Sponsor A Data sets

B

Repository B (other data sets)

Researcher Researcher

Perform Feasibility checks

The Importance of Data Sharing and Transparancy (EMA)

Presenters: Barbara Bierer, MRCT Center Fergus Sweeney, European Medicines Agency

4/5/16 8

The Importance of Data Sharing and Transparency

MRCT CENTRE – Wellcome Trust meeting The Future of Clinical Trial Data Sharing

Fergus Sweeney, EMA Head of Division, Inspections and Human Medicines Pharmacovigilance

Our Mission

The mission of the European Medicines Agency is to foster scientific excellence in the evaluation and supervision of medicines, for the benefit of public and animal health.

10

11

EMA is designed to coordinate and mobilise the existing scientific resources of Member States:

• for the evaluation & supervision of medicines
• to coordinate inspections
• to advise on research & development programmes
• 28 Member States, 4,500 experts

European experts’ network underpins the work of the CHMP, CVMP, COMP, PDCO, CAT, PRAC, HMPC, working parties, SAGs

EMA: A Networking Agency

12

The Agency is responsible for:

• The evaluation of marketing authorisation for human and veterinary

applications submitted by pharmaceutical companies

• The coordination of European pharmacovigilance (supervision of the medicines
• n the market)
• The provision of scientific advice on the development of medicines
• The evaluation of applications for orphan designation in EU
• The evaluation of paediatric investigation plans (or waivers)
• The evaluation of arbitration and referral procedures
• The provision of good quality and independent information on the medicines it

evaluates to patients and healthcare providers

• The coordination of Member States’ inspections (GMP, GCP, GVP, GLP)

The various roles of the EMA

13 13

28 EU + 3 EEA Member States + 4,500 European experts EU institutions: Commission - Parliament Committee for Orphan Medicinal Products (COMP) Committee for Herbal Medicinal Products (HMPC)

EMA Secretariat

Committee for Veterinary Medicinal Products (CVMP)

Management Board

Committee for Human Medicinal Products (CHMP) Committee for Advanced Therapies (CAT) Paediatric Committee (PDCO)

EMA-EU Network

Pharmacovigilance Risk Assessment Committee (PRAC)

Transparency – some history:

• 20 years on, easy to forget how ‘radical’ early ones were:
• A long series of initiatives right from start of Agency

§ EPAR (1995); Same SPC and PIL in all languages (1995); Publication of all experts (1996); Summaries of opinions (1998, 2002); Rules on access to documents (1997, 2006); Orphan drug INN and orphan indication (2004); PIP opinions (2006)

14

Examples of key transparency initiatives

• HMA/EMA recommendations on transparency, PSURs, CCI, PPD
• EudraVigilance access policy, ADR website
• EudraCT and EU Clinical Trials Register websites
• EudraGMDP website
• CVs and declarations of interest of all committee members, experts and staff
• More scientific and regulatory articles in various journals

15

…and more examples

• More information on workshops and meetings at EMA, more meetings broadcast live with

recordings available

• Stakeholder workshops and opportunities to feedback
• Staff code of conduct updated and published
• PRAC: list of products under additional monitoring, information on signal management etc.
• Agendas and minutes of all scientific committees published
• Public hearings coming soon

16

Access to documents, requests for information

17

• Policy published November 2010
• ‘Outputs’ table to guide

implementation

• Change in access policy had big

impact on Agency operations

• Average 450-500 requests for

information each month

• But the information only goes to the

requester…..although they may share, it is less visible

Clinical trial transparency and EMA – 3 pillars

18

EU Clinical Trials Register Proactive publication of clinical study reports – Policy 70 Clinical Trial Regulation and EU Portal and Database – public information clinical trials authorized in EU

Objectives of clinical trial transparency

• Have all clinical trials been publicly registered?
• Is there a trial in which I could participate?
• What was the outcome of the trial I did participate in?
• What trials were the basis of the marketing authorisation, what were their results?
• What is known about the medicine I am taking/prescribing?
• Can we review the data used to support the marketing authorisation?
• Has the trial we are designing already been conducted? Were there problems with

similar trials?

• Strike the right balance to inform the public, protect public health and foster the

innovation capacity of European medical research.

19

20

EU Clinical Trials Register Proactive publication of clinical study reports – Policy 70 Clinical Trial Regulation and EU Portal and Database – public information clinical trials authorized in EU

EU Clinical Trials Register

• Launched in March 2011
• Contains protocol and results related data for interventional CT started after May 2004

– Phase II-III-IV trials conducted in adults in the EEA – Phase I-II-III-IV paediatric trials in the EEA – Only phase I trials conducted in adults & part of a PIP are made public (small %) – NCA decision positive and IEC opinion positive recorded in EudraCT for adult trials – for paediatric trials IEC opinion positive or negative – Paediatric CT outside of the EEA if they are part of an agreed Paediatric Investigation Plan (PIP) (including a small % of adult phase I trials if they are part of a PIP)

21

22

23

24

25

EU Clinical Trials Register Proactive publication of clinical study reports – Policy 70 Clinical Trial Regulation and EU Portal and Database – public information clinical trials authorized in EU

26

Chapter 1 Scope and definitions Chapter 2 Procedural guidance and processes on submission Chapter 3 Guidance on anonymization of clinical study reports for publication Chapter 4 Guidance on what is no considered to be CCI, and on redaction in the limited circumstances where information can be considered CCI.

27

• First phase: publication of clinical reports expected Sep 2016 once IT tool is ready

– Any new MAAs and Article 58 applications, submitted after January 2015 – Extension of indication applications and line extension applications for existing CAPs, submitted after 1 July 2015 – All other post-authorisation applications for existing CAPs submitted as of a date TBD in 2016

• Second phase: review of various aspects in relation to IPD will follow once first phase is underway
• Scope:

– Clinical data: clinical reports (i.e. clinical overviews (module 2.5), clinical summaries (module 2.7) and clinical study reports - CSRs (module 5), together with appendices 16.1.1, 16.1.2 and 16.1.9) and individual patient data (IPD)

• Not within scope legacy data for centrally authorised products (CAPs) and Clinical data for non-CAPs
• My colleague Frances Nuttall will provide a more detail on Policy 70 and its implementation in

this afternoon’s session.

28

Policy 70 on proactive publication of clinical data submitted under centralised procedure in MAAs to EMA

29

EU Clinical Trials Register Proactive publication of clinical study reports – Policy 70 Clinical Trial Regulation and EU Portal and Database

Scope of clinical trial Regulation (EU) No. 536/2014

• Interventional clinical trials on medicines conducted in the EU/EEA (i.e. with at least one

investigator site in EU/EEA)

• Clinical trials authorized under the new Regulation or still ongoing three years after it comes

into application

30

Member States Sponsors EMA

General public

Commission Applicant of a MA

Submit submission package (CTA & dossier) / Address request for information Submit notifications:

• Withdrawal
• Start of trial
• First visit first subject
• End of recruitment
• End of trial (in each MS, All MS,

Global)

• Temporary halt
• Restart of trial
• Early termination
• Serious Breaches
• Unexpected events which

affect risk/benefit Submission of clinical study result summary Submission of Inspection Reports

• f third country authorities

Update of Clinical Trial information re non substantial modifications Submission of CSR Submission of Union Control Reports Notification of willingness to be RMS(Part 1)/Decision on RMS Submission of requests for information Notification of the final validation (initial, additional MS or Substantial Modification) Submission final AR Part 1 and 2 Final single decision notification Communication disagreement to part 1 assessment Communication on implementation

• f corrective measures

Submission Inspection Information This slide depicts the processes each stakeholder will be able to complete in the new EU Portal and Database:

Search and view CT related information saved in the EU database (that is not confidential) Runs the system and generates reports on the data in the system

Processes for each stakeholders in the system

31

EU portal and database – Data standardisation

WHO

WHO ICTRP standard will be fully met, and data provided to the ICTRP by the EU database

NIH

(clinicaltrials.gov)

Collaboration and discussion on the anticipated changes to the data model (focusing on protocol / results) to ensure convergence and alignment where the same elements are used in both US and EU systems

CDISC

Collaboration on clinical trial registration including study design data model, and in due course on results model

32

Transparency legal requirements: Clinical Trials Regulation

Article 81(4) of Regulation (EU) No. 536/2014

• EU database publically accessible by default, with exceptions justified on any of the following

grounds:

– Protection of personal data; – Protection of commercially confidential information in particular taking into account the MA status of the medicinal product, unless there is an overriding public interest in disclosure; – Protecting confidential communication between MS in relation to the preparation of the assessment report; – Ensuring effective supervision of the conduct of a clinical trial MSs.

33

34

• Appendix, on disclosure rules, to the

“Functional specifications for the EU portal and EU database to be audited - EMA/42176/2014”

• Endorsed on 2 October 2015 by EMA

Management Board and published on 6 October 2015

http://www.ema.europa.eu/docs/en_GB/ document_library/Other/2015/10/ WC500195084.pdf

Requirements for operation of a feasible system

• To enable public access to the database, rules for the application of the exceptions, set out in

Article 81(4), are required.

• Rules, criteria and data to enable the system software to determine, automatically, when a

particular data element or document should be made public.

• Automatic rules are necessary because there will be 4-5000 clinical trial applications, dozens
• f documents and hundreds of data fields per clinical trial and multiple processes per trial

taking place in the system every year.

• Rules designed to produce a consistent and predictable outcome so that those submitting

data and documents and those viewing them know what will be made public and when.

• A manual override available to enable earlier publication in exceptional circumstances where

an overriding public interest applies, or to remediate a publication error.

35

Summary of rules

• The same rules apply whether the sponsor is a pharmaceutical company, academic research

group or other type of organisation.

• The rules depend on the IMPs used in a trial and how they are used.
• Trials defined as belonging to one of three categories, at the time of initial assessment of the

clinical trial application:

– Category One: Pharmaceutical development trials– essentially Phase I trials in healthy or patient volunteers, bio-equivalence and bio-similarity trials. – Category Two: Therapeutic exploratory and confirmatory trials - essentially Phase II and III trials of novel products or new indications or formulations of existing products – Category Three: Therapeutic use trials – essentially Phase IV and low-intervention trials

• Depending on the category of trial the sponsor will have the possibility to defer publication of

certain data and documents up to a maximum time limit, if needed, to protect commercially confidential information.

• The use of deferrals will be monitored and should not exceed what is really needed.

36

Summary of Rules

• Public registration of trials at their start including all information needed for patients who may

wish to participate in trials with therapeutic, diagnostic or preventive objectives.

• Publication of results of all trials (structured summary, layperson summary and in case of

Marketing Authorisation Application in the EU the clinical study report).

• The structured summary of results and laypersons summary will be made public 12 months

after the end of each trial, regardless of the Marketing Authorisation status of the medicines studied in the trial.

• Only possibility of justified deferral for summary results in case of category I trials up to a

maximum of 30 months post end of trial (i.e. maximum 18 months deferral).

• Detailed rules for dossier documents linked to category of trial

37

Summary - Clinical Trial Transparency – and EMA

• Clinical Trials authorised in EU/EEA:

– Growing body of clinical trial information and results summaries in EU Clinical Trial Register for trials authorised since 2004. – New clinical trial Regulation - Extensive information on clinical trials from authorisation to the trial summary results of all trials authorised in EU/EEA under the new Regulation.

• Clinical Study Reports submitted in Marketing Applications in EU:

– All clinical study reports included in marketing applications to EMA from 1 Jan 2015 on – Clinical study reports for all trials authorised in EU/EEA under the new clinical trial Regulation and included in a marketing application in EU.

38

Thank you for your attention

European Medicines Agency

30 Churchill Place • Canary Wharf • London E14 5EU • United Kingdom

Telephone +44 (0)20 3660 6000 Facsimile +44 (0)20 3660 5555 Send a question via our website www.ema.europa.eu/contact

Fergus.Sweeney@ema.europa.eu Further information

Follow us on @EMA_News

Public data and information about medicines, their development and authorisation

• Generate trust – information is available;
• Build confidence – I understand what is happening;
• Empower – knowledge enables decision-making

Governance Workgroup

Rebecca Li, PhD Moderator: Barbara Bierer, MD

March 21, 2016

• Working Group Main Objec2ves:

41

MRCT Governance Working Group

Informa2on Technology (IT) Working Group Business Models Working Group

Data Sharing Strategy Phase: Governance

4/5/16

Develop a high-level charter Develop principles for organiza4onal leadership

• Governance
• Scope and exper2se
• Opera2ng guidelines

Strategic policy decisions

• PlaUorm
• Governance
• Communica2on

Workgroup Deliverables

• Key Workgroup Deliverables:

– Governance structure – Review Process – PlaUorm par2cipa2ng trials – Data packages – Data Sharing guidance and resource kit

4/5/16 42

Proposed Organiza2onal Structure for Non Profit Vivli

Board of Directors Finance & Audit CommiJees Advisory Technical CommiJee External Advisory CommiJee President / Exec Director

4/5/16 43

Proposed Governance Model: Hybrid

• Func2onal Hybrid Model

– Incorporated as Non-profit Organiza2on – Membership of the Board is combina2on of representa2on of contributors to the organiza2on (ojen financial) and “at-large” posi2ons – Representa2ves can be nominated by a contributor, subject to approval of the board (ojen termina2ng when organiza2on is financially stable and independent) – At-large members, appointed by the nomina2ng commiJee of the Board (ajer ini2al appointment) according to personal investment in the project, e.g:

• Academic representa2on
• Government
• Pa2ent/Pa2ent Advocate, etc.

4/5/16 44

Workgroup Deliverables

• Key Workgroup Deliverables:

– Governance structure – Review Process – PlaUorm par2cipa2ng trials and data packages – Data Sharing guidance and resource kit

4/5/16 45

Review Process Proposal

Vivli will act as a “data library” which respects the needs of requesters and concerns of donors Several review op2ons for shared data were developed:

• Tier 1: Data generator would donate data which would not

require a review process.

• Tier 2: Data generator would cede review to the plaUorm’s

selected IRP.

• Tier 3 3: Data generator will maintain own review process.

PlaUorm will forward requests for review.

4/5/16 46

Data Request Model: Phase 1

4/5/16 47

Data Request Models: Aspira2onal

4/5/16 48

Workgroup Deliverables

• Key Workgroup Deliverables:

– Governance structure – Review Process – PlaForm parHcipaHng trials and data packages – Data Sharing guidance and resource kit

4/5/16 49

Par2cipa2ng Trials and Par2cipants

• What defines a parHcipaHng trial on the plaForm?

– Registered in ClinicalTrials.gov, WHO ICTRP or other trial register – Metadata (searchable terms) for par2cipa2ng trials are available and curated – IPD is or will be made available

• PlaUorm will not be prescrip2ve in the types of trials or datasets

that will be included and where appropriate link to exis2ng guidance and regula2ons.

• Par2cipa2ng trials - at a minimum we encourage the inclusion of

large pivotal trials of marketed products

4/5/16 50

4/5/16 51

250 trials

Sponsor no,fies pla0orm of # of par,cipa,ng trials

PlaUorm curates metadata for the 250 trials IPD SHARED

Searchable terms match available IPD Review process

PlaUorm Par2cipa2ng Trials (e.g Sponsors)

PlaUorm Par2cipa2ng Trials (e.g. Publica2on-based)

• Publica2on

prepared for submission

Submission to Journal

• Data Package

Prepared for PlaUorm (anonymized)

Submission to PlaUorm

• IPD / Data

Package and Publica2on held in Embargo

Ar2cle Published & Data available

• n plaUorm

4/5/16 52

Data Packages – What to share?

• A Data package is the set of data and associated

materials about a study that is to be shared via the PlaUorm

May include:

• Analyzable individual par2cipant-level dataset (IPD)
• Final study protocol + amendments
• Sta2s2cal analysis plan
• Annotated case report forms
• Analy2c code suppor2ng the published results
• Redacted CSR

4/5/16 53

Data Packages – Recommenda2ons

Basic Package (recommended for inclusion)

• Analyzable individual parHcipant-level dataset (IPD)
• Final study protocol + amendments
• StaHsHcal analysis plan

May also be included if available to share:

• Annotated case report forms
• Analy2c code suppor2ng the published results
• Redacted CSR

A checklist will be available by the pla0orm to provide op,ons

4/5/16 54

Workgroup Deliverables

• Key Workgroup Deliverables:

– Governance structure – Review Process – PlaUorm par2cipa2ng trials and data packages – Data Sharing guidance and resource kit

4/5/16 55

Data Sharing Resource Kit (See Appendices)

• Data Use Agreement
• Data Contributor Agreement
• ICF

4/5/16 56

Successful Case Studies of Secondary Use of Data

Data Generator

Data Recipient

Impact

(1) ITN Network – RAVE Trial – NIAID (funder)

Inves2gators at the Ins2tute for Computa2onal Health Sciences, University

• f California, San Francisco

Reanalysis of the RAVE trial demonstrated a correla2on between pa2ent’s granulocyte subsets at baseline with the likelihood of remission at 6 months following either cyclophosphamide or rituximab in pa2ents with ANCA- Associated Vasculi2s.

(2) Pfizer, GSK, Lundbeck

Inves2gators at the University of Gothenberg, Sweden, Swedish Research Council (funder)

In a post-hoc analyses of 18 trials of SSRIs u2lizing a single ra2ng scale versus the Hamilton depression ra2ng scale, SSRIs were shown to be more consistently effec2ve than placebo in the treatment of pa2ents with major depression.

(3) Medtronic – YALE YODA (intermediary)

Inves2gators at the University of Colorado Department of Orthopedics

Metaanalysis of 4 trials demonstrated that, for pa2ents with degenera2ve disc disease following lumbar arthrodesis, the presence of radiographic fusion was correlated with improved clinical outcomes compared to radiographic non-union (Oswestry Disability Index - ODI; Numeric Ra2ng Scales (NRS) for back and leg pain).

(4) J&J, Amgen, Hoffman-LaRoche and 5 independent inves2gators

Inves2gators at University of Bern; German Federal Ministry of Educa2on and Research, Medical Faculty of University of Cologne and Oncosuisse (funder)

Metaanalysis of 18 trials (13,933 pa2ents) demonstrated that administra2on of erythropoiesis-s2mula2ng agents in oncology pa2ents increased mortality compared with transfusion therapy alone. The increase in mortality must be balanced against the poten2al benefits of erythropoiesis -s2mula2ng agents in this pa2ent popula2on.

(5) Pfizer, Sanofi

Inves2gators at University of North Carolina

Mitoxantrone added to prednisone in the treatment of pa2ents with post-docetaxel, metasta2c, castrate-resistant prostate cancer showed no survival benefit over the use of prednisone alone and may be associated with increased toxicity.

57

The MRCT Center’s Data Sharing Workgroup Members

4/5/16

Governance Work Stream

Co-Chairs: MRCT Center Wellcome Trust Arnold FoundaHon

Team Members: Mark Barnes (MRCT Center) Barbara Bierer (MRCT Center) Kris Bolt (MRCT Center) Stuart Buck (Laura and John Arnold FoundaHon) Marla Jo Brickman (Pfizer) Nina Hill (Pfizer) Lauren Johnson (MRCT Center) Rebecca Li (MRCT Center) Nick Lingler (DeloiSe ConsulHng) Heather Marino (MRCT Center) Sandra Morris (Johnson & Johnson) Jennifer O’Callaghan (Wellcome Trust) Nicola Perrin (Wellcome Trust) Paul Seligman (Amgen) Ida Sim (UCSF) Jessica ScoJ (GlaxoSmithKline) Catrin Tudur Smith (U. of Liverpool) Natalie Zaidman (Pfizer)

58

Ques2ons and Discussion

4/5/16 59

Business Models Workgroup

Presenter: Rohin Rajan, DeloiJe Consul2ng LLC Moderator: Frank Rockhold, Duke University

4/5/16 60

MRCT Clinical Trial Data Sharing Business Model Working Group Outputs

March 2016

Executive Summary

Background and context

63

Key objectives for the Wellcome Trust / MRCT Business Models workstream

1) To develop sustainable business models for Vivli 2) To align with the governance and IT workgroups on how to best to develop and capacitate the not-for-profit entity, including in regard to its financial viability

64

Framework and Approach

Proposals discussed in the Business Models workstream need to be aligned to IT and Governance workstreams to finalize operating model and timeline

scan of existing clinical trial data sharing technologies and platforms

functionality and capabilities of the desired platform / system / entity

market capabilities, gaps and short-term needs; validate scope

• f Vivli

the development of Vivli’s business model

• perating models that

would enable the capabilities required for Vivli

execution modes based on cost, speed, risk, revenue streams Determine near-term and long-term

• pportunities to flexibly

evolve Vivli

sustainable business model to address gaps, seek out potential partners, and build as needed

Landscape Capabilities Operating Model Timeline

Context Discussion

65

Assumptions and considerations

§ Should be a “not-for-profit” organization, but can collaborate with for and non-profit companies § Requires sustainable support (resources, funding) from industry, academics and other stakeholders § Over the next 18 – 36 months, the objective is to create valuable offerings to attract users and define

• pportunities for growth

§ Cannot reinvent technologies already in place; but in the near-term, potentially:

– Connect existing platforms in an extensible and collaborative manner – Improve interoperability by facilitating data warehousing, integration, and curation – Create a marketplace for academic researchers and smaller, non-participating companies

§ Revenue opportunities will be evaluated continuously as offerings become mature § Long-term, Vivli may be able to unify various data sources into a single multifaceted platform

66

Key findings from the external scan

§ The scan revealed that numerous initiatives either exist or have been created § No single effort has become the ‘silver bullet’ § Several major biopharmaceutical companies are involved in at least two data sharing initiatives § Smaller players and academics do not have similar modes for sharing clinical trial data

67

Objectives for today’s Business Model presentation

§ Consider and discuss the various operating models presented; determine if other, distinct models exist § Review and align on near, mid and longer-term options for the evolution of Vivli § Discuss and determine which functional capabilities can be achieved through (post presentations):

– Partnerships with existing systems or platforms – Assembly and coordination of current marketplace offerings – Creation / development of unique services and capabilities

Executive Summary

Overview of potential

• perating model options

69

Potential data sharing operating models

A number of different routes could be taken to achieve the goal of increasing clinical trial data sharing & creating a data aggregation platform

Engage in advocacy for policy and practical changes to increase transparency and cooperation in clinical trial data sharing with existing efforts (platforms / systems / regulators, other stakeholders) Advocate & Lobby Build and develop capabilities as needed to facilitate the exchange of clinical trial data Build & Develop Find, support, and functionalize specific solutions to address existing and emerging gaps in clinical trial data sharing, to facilitate exchange between platforms Assemble Identify and partner with existing initiatives (e.g. Project Data Sphere, CSDR) to enable immediate access to current capabilities / functionality as a foundation for growth Partner § Generalized options to be aligned to specific IT and Governance goals and partnership opportunities § Options not mutually exclusive, and could be reconfigured to meet evolving needs § “Advocate and Lobby” model would continue throughout the evolution of Vivli § “Advocate and Lobby” option does not comprise a data sharing entity, and could be an extension of MRCTs

Assumptions

70

Operating model dimensions

Operating models were evaluated using a set of comparison dimensions

Capability • Technological capabilities offered by Vivli to facilitate data sharing across multiple stakeholders Dynamism

• Flexibility / elasticity with which Vivli can quickly shift strategic priorities based on the market

Investment

• Up-front capital and ongoing operational investments to fund Vivli

Stakeholder Risk

• Data Contributors: risk of providing data to stakeholders who may undermine efficacy and safety of

sponsor’s drug leading to regulatory action and / or revenue losses

• Data Users: risk of using data that may not be accurately collected, cleansed, integrated and

associated implications on users reputation and credibility

• Vivli: risk of losing the reputation as a credible, trusted and impartial curator of clinical trial data and

inability to sustain profits over the long term; also risk of competition to existing entities / Vivli depending

• n operating model and capabilities chosen

Control

• Ability to exercise direct control over the direction of technological capabilities and services

Time to build • Speed with which the solution can be implemented / achieved Build / Develop Partner Advocate & Lobby Assemble

71

Advocate & Lobby

Engage in advocacy, lobbying for policy and practical changes to increase transparency and cooperation in clinical trial data sharing with existing efforts (platforms / systems / regulators and other stakeholders)

Ranking Detail

Capability

• Low capability to directly expand clinical trial data sharing

Dynamism

• Opportunity to quickly react to changing trends and dynamics in data

sharing, and influence the broader environment Investment

• Limited investment needs to cover communications / marketing efforts

Stakeholder Risk

• Low risk level as Vivli will not directly facilitate data sharing, but may not

address the needs of stakeholders leaving Vivli with reputational consequences Control

• Lowest control given role limited to advocacy / engagement

Time to build

• Efforts largely underway already as part of MRCT scope and mission

Build / Develop Partner Advocate & Lobby Assemble

Less Favorable More Favorable

72

Assemble

Ranking Detail

Capability

• Vivli may not be the central platform for data sharing, but offer capabilities

that facilitate sharing by addressing current gaps Dynamism

• Operating model is focused on reacting to changing dynamics and new

trends in clinical trial data sharing, and to find / fund evolving solutions Investment

• Investment need could vary and would be at Vivli’s volition

Stakeholder Risk

• Risk due to unpredictability of investing in and / or effectively connecting 3rd

party solutions and driving value

• Risk to existing platforms by means of disruption / loss of autonomy

Control

• Limited choice over existing solutions / platforms given position as a third

party, but control achieved through customization and coordination Time to build

• Accelerated timeframe as efforts would target existing data sharing

technologies, but still may take time to derive and build solution Build / Develop Partner Advocate & Lobby Assemble

Neutral

Less Favorable More Favorable

Find, support, and functionalize specific solutions to address existing and emerging gaps in clinical trial data sharing, to facilitate exchange between platforms

73

Partner

Ranking Detail

Capability

• Partnering would secure existing technical capabilities, with additional build

conducted as needed (including to achieve interoperability) Dynamism

• Ability to chose a platform / system as desired, but flexibility (initially) is

limited to capabilities offered by the platform / system Investment

• Upfront investment to secure partnership, with further investment variable

(depending on level of upgrade targeted)

• Could leverage non-profit status to enhance investment negotiations

Stakeholder Risk

• Risk exposure from required upfront investment
• Adaptability of current capabilities for further customization & interoperability

needs Control

• Ability to select the platform / system with existing capabilities and potential

for customization Time to build

• Fastest route to access current capabilities and create a foundation for

enhancements Build / Develop Partner Advocate & Lobby Assemble

Neutral

Less Favorable More Favorable

Identify and partner with existing initiatives to enable immediate access to current capabilities / functionality as a foundation for growth

74

Build / Develop

Ranking Detail

Capability

• Capabilities can be developed and customized as needed based on current

gaps, best practices from existing solutions, and stakeholder needs Dynamism

• Limited flexibility and ability to adapt model once build is underway

Investment

• Significant resource / capital investment required to build model and develop

technological solution Stakeholder Risk

• High capital outlay
• Potential duplication of systems and platforms already in existence
• Competition for users / unclear value proposition / differentiation

Control

• High degree of control afforded to Vivli to customize solution as needed

Time to build

• Significant time required to design, source, and build solution
• Prime target for disruption as other, well-positioned companies able to

deploy solutions over a faster timeline Build / Develop Partner Advocate & Lobby Assemble

Less Favorable More Favorable

Build and develop capabilities as needed to facilitate the exchange of clinical trial data

75

Comparison across dimensions

Capability Dynamism Investment Stakeholder Risk Control Time to build Cumulative

Dimension Rankings

Build / Develop Partner Advocate & Lobby Assemble

(N) (N)

Less Favorable More Favorable

76

Learnings from the operating model comparison

§ The “Build” option offers greatest control

– “Partner” & “Assemble” models could achieve similar capabilities more flexibly, faster and at lower cost – “Partner” offers established capabilities and contain the largest data-sets – Certain unique capabilities may require build as they are currently unavailable not suitable in current form (e.g. due to cost, capabilities, content etc.) - to be discussed post IT presentation

§ The “Advocate and Lobby” option is the lowest investment and could be an extension of MRCT

– Offers no technological capabilities in the near-term, but serves to align other initiatives and grow users – As a lobbying entity, can influence the key technology, service and regulatory priorities

Executive Summary

Potential near / medium term operating model evolution & revenue considerations

78

Proposed timeline for the evolution of Vivli

Near-term (0 – 6 mos) Mid-term (6 – 18 mos) Long-term (18 - 36+ mos) Advocate & Lobby

§ Clearly define mission, scope, and timeline for Vivli § Widespread awareness and buy in to Vivli’s mission amongst the broader industry stakeholders (academics, regulators/ editors etc.) § Service offerings and revenue targets identified Partner Assemble § Agreements / partnerships achieved with other data sharing

• rganizations

§ Clearly defined scope / roles and responsibilities between Vivli and complementary organizations § If possible Vivli will endeavor to not be in “addition to” existing platforms but will assume their role/ services to reduce burden § Participation / sign on by biotechs & academic organizations Build / Develop (as needed) § Substantial and growing user base § Technological capabilities established to facilitate data sharing § Sufficient revenue flow to sustain

• ngoing operations of Vivli

achieved

79

Timeline Details: Near-term – 0 to 6 months

0 – 6 Months 6 – 18 Months 18 – 36+ Months Key Considerations / Requirements Investment Required (annualized)

§ Screen potential targets for partnership / assembly; conduct capability gap analysis § Achieve partnership with a major data sharing platform § Identify required investment level, revenue sources, and

• perating model of Vivli

§ Determine scope of Vivli – Type of companies / willingness to participate – Type of capabilities targeted (technology, services) – Scope / type of data to be shared § Align with ongoing industry data sharing platform / effort § Establish a clear value proposition for Vivli § Create a detailed project plan for Vivli § Seek immediate implementation to support gaps (e.g. for academic ICMJE compliance) if needed § Capital: 0 § Operational: 2-2.5 FTEs ~ $400 – 500K – Leadership: 1 FTE – Project manager: 1 FTE – SME: 1 FTE (mix of partial or full-invested time)

Potential Revenue Streams

§ Grants and donations

Advocate & Lobby

80

0 – 6 Months 6 – 18 Months 18 – 36+ Months

+

Timeline Details: Medium term – 6 to 18 months

Key Considerations / Requirements Investment Required (annualized)

§ Facilitate transition / partnering of existing platform to Vivli § Launch cross-platform services by facilitating connectivity and uniformity of approach (IRP, DSA etc) § Achieve user base growth / positive indications for user base growth for cross-platform/academic data) § Create and release RFP to conduct technology build § Enact detailed project plan (earlier in phase) § Begin to defining services and capabilities for which Vivli may be able to generate revenue (e.g. data warehousing for academics) § Capital: $5M - $10M § Operational: 19 – 25 FTEs ~ $2M – $4M* – Leadership / management: 5 FTEs (3 from prior stage plus two additional) – Technology: 5-7 FTEs – Service: 3-5 FTEs – Operations: 6-8 FTEs

Potential Revenue Streams

§ Grants and donations § Data repository services for academics and

• thers

§ Data Anonymization / De-identification § Data submission support services § Annual Subscription § Proposal Review & Refinement § Pay Per Use / Request § Data retrieval § Value add services § Rare disease populations § Advertising

Advocate & Lobby Partner & Assemble

81

0 – 6 Months 6 – 18 Months 18 – 36+ Months

Timeline Details: Longer term – 18 to 36+ months

Key Considerations / Requirements Investment Required (annualized)

§ Expand technological capabilities to improve efficiency and capability of platform § Achieve sustainable revenue generation § Increase industry / academic organization participation § Expand service offerings § Proceed with detailed project plan § Interface / engage with regulators this should be in lobby/ advocate and should be earlier § Capital: TBD § Operational: TBD – Leadership / management: 5+ FTEs (TBD) – Technology: 7+ FTEs (TBD) – Service: 5+ FTEs (TBD) – Operations: +8 FTEs (TBD)

Potential Revenue Streams

§ All revenue streams from prior phase § Access to Tools § Customized report generation services

Partner & Assemble Build / Develop Advocate & Lobby

+

82

Revenue considerations & proposed approach

§ Over the next 36 months “revenue” should refocused towards grants and charitable donations – If near-term includes support for ICMJE requirements, earlier revenue options may available § The rationale is as follows: – Services will take time to launch, and should be refined and enhanced before charging users – Services and capabilities should be extensible for interoperability with Vivli / other platforms – The barrier to entry for investigators should be kept low – Time will allow positive interactions with partners, reviews and press coverage of effort to build – Prices should be in-line with what industry and stakeholders are willing to pay – “Non-profit” title will facilitate collaboration and should de-incentivize competition § Once a substantial user base / repository size is established revenue generation can be expanded

Executive Summary

Summary & next steps

84

Summary

§ Building a technology will require significant effort and investment and may not allow flexibility to change directions if the marketplace changes dynamically § Assemble and Partner offer significant flexibility for marketplace demands / disruptions and may be cheaper, faster options in the near-term § In the immediate-term (~6 months), the lowest risk option is Advocate and Lobby with an emphasis

• n working towards an Assemble and Partner model

§ Despite feasibility of launching revenue streams in the medium term (6-18 months), focus on securing grants, donations, and other philanthropic sources of support

85

Key questions for discussion

Potential operating model options:

§ Are there other operating models that could be perceived as distinct from the ones presented? § Does the group agree on the choices and tradeoffs therein?

Near term operating model evolution:

§ Does the near term evolution of Vivli seem sensible? § Are there assembly / partnership activities we can take on in the near term? What other near term considerations should we keep in mind? § What capabilities do we want to focus on to achieve: – Partner? – Assemble? – Build? Can the build stage begin during an earlier phase than currently targeted?

86

Data Sharing Workgroup Members

Co-Chairs: Wellcome Trust and MRCT Center Team Members: § Mark Barnes (MRCT) § Barbara Bierer (MRCT) § Patrick Cullinan (Takeda) § Rob Frost (GSK) § Rebecca Li (MRCT) § Peter Lyons (Deloitte Consulting) § Nicola Perrin (Wellcome Trust) § Rohin Rajan (Deloitte Consulting)

Executive Summary

Appendix

88

Potential evolution of Vivli’s operating model

Vivli Imperatives & Goals

§ Engage in advocacy § Align with other industry efforts § Engage with other clinical trial data sharing platforms § Screen other platforms for assembly / partnership potential § Define platform and specifications for Vivli’s efforts to expand / host clinical trial data sharing § Begin scoping potential IT “build needs” and identifying services Vivli will provide § Identify preliminary targeted revenue streams for Vivli § Socialize data sharing ecosystem with non-participating pharma / biotechs / academics § Begin developing functionality to enable Vivl’s vision and mission § Create partnerships with other platforms / data sharing organizations § Launch data sharing and value-add services § Define additional desired add-on functionality requiring further build § Engage with biotechs / academics and garner buy-in / participation § Begin facilitating exchange of data with and between non-participating biotechs / academics § Facilitate alignment with other data sharing efforts § Build additional add-on functionality (as desired) and improve and expand service offerings § Refine and operationalize the platform § Increase user base and frequency of use of platform § Streamline process and reduce burden for all through technology (donators and requestors) § Launch efforts to monetize and/or generate revenue to sustain Vivli § Test viability of existing revenue streams and determine new

• pportunities

Critical Success Factors

§ Define mission, scope, and timeline to establish Vivli § Widespread awareness and buy in to Vivli's mission amongst the broader industry/stakeholders (academics, regulators/editors etc) § Service offerings and revenue targets identified § Agreements / partnerships achieved with other data sharing organizations § Clearly defined scope / roles and responsibilities between Vivli and complementary organizations § If possible Vivli will endeavor to not be in “addition to” existing platforms but will assume their role/services to reduce burden § Substantial and growing user base § Technological capabilities established to facilitate data sharing § Sufficient revenue flow to sustain

• ngoing operations of Vivli achieved

Partner Assemble Build / Develop 0 – 6 Months 6 – 18 Months 18 – 36+ Months Advocate & Lobby

EMA Policy 70 Implementa2on

Presenter: Frances NuJall, European Medicines Agency

4/5/16 89

European Medicines Agency Policy on publication of Clinical Data for medicinal products for human use (policy 0070)

MRCT CENTRE – Wellcome Trust meeting The Future of Clinical Trial Data Sharing

Frances Nuttall, Senior Policy Adviser, EMA, 21 March 2016

Policy purpose + objective

Clinical Data Publication policy 91

• Access to Documents – individual request, route continues

To date:

• 2 October 2014, Clinical Data Publication (human medicinal products)

New policy:

• Public presentation of clinical data, clinical reports

What is it:

• Transparency, continued EMA commitment, regulatory decision basis
• f medicinal products
• Proactive publication enables public scrutiny: establishes trust,

confidence

• Better public information: Public access enables application of new

knowledge in future research, increases efficiency of medicine development, learning from experience

• Avoids clinical trials duplication: limits unnecessary patient exposure
• Enhanced scientific knowledge/value of secondary analysis: sharing

scientific knowledge, contribution to public health, confirmation of regulatory decisions/need to review regulatory decisions taken.

Benefits:

Clinical data publication policy & Clinical Trial Regulation

Clinical data publication policy: all the clinical reports (trials located in EU or non EU) in the regulatory submission to EMA

Clinical Data Publication Policy 92

Clinical Trial Regulation: EU clinical trials

Clinical data publication policy

Clinical Trial Regulation

Clinical Data Publication policy 93

Policy effective: 2015

1 January 2015: Marketing authorisation applications

• Withdrawn

applications pre

• pinion included -

Innovation

1 July 2015: extension of indication and line extensions

To add later: post authorisation regulatory categories Not in scope: Legacy data, pre 2015

Clinical Data Publication Policy scope

Two phase implementation

Phase I

• Clinical reports = clinical overview, clinical

summary, clinical study reports, protocol & amendments, sample case report form, documentation of statistical methods

• EMA is working on Phase I implementation

Phase II

• Individual patient data (IPD)
• To approach later

Clinical Data Publication policy 94

Policy implementation

Introduction, scope, definitions

External guidance on the procedural aspects related to the submission of clinical reports for the purpose of publication in accordance with EMA policy 0070 Guidance on the identification and redaction of commercially confidential information (CCI) in clinical reports submitted to the EMA Guidance to pharmaceutical industry on the anonymisation of clinical reports for the purpose of publication in accordance with EMA policy 0070

Published on EMA website: 3 March 2016

Clinical Data Publication Policy 95

Clinical Data Publication Guidance

Clinical reports public presentation

Commercially confidential information (CCI) EMA position: majority of clinical report content is not CCI Redaction principles set out in the policy The company justifies each redaction, EMA reviews redactions & decides if accepted or not

Clinical Data Publication policy 96

Anonymisation Data utility: important for researchers, EMA encourages utmost data utility, balance to protect personal data, EMA guidance recommends methodology to avoid (re)identification of clinical trial participants, various techniques, evolving area. Anonymisation report: requirement, reviewed by EMA + published, shows the company’s approach for the specific case.

Access to clinical data published

Watermark:

The EMA applies a watermark to the clinical reports before their publication to emphasise the prohibition of their use for commercial purposes The watermark text is: “www.ema.europa.eu This document cannot be used to support any marketing authorisation application and any extensions or variations thereof”

Clinical Data Publication Policy 97

General information purpose:

§ View on screen only (no save, download, print, copy). § Digital rights management applied. § Text searchable. § Minimum user registration requirement. § Specific Terms of Use

Academic/non-commercial research purpose:

• Save, download, copy etc.
• Text searchable.
• More personal data on registration.
• Specific Terms of Use

Watermark Samples

Clinical Data Publication Policy 98

These examples follow how the watermark will be presented on the published information

Clinical Data Publication Policy 99

Other companies are being informed to wait, for now, until EMA contacts them . EMA will work forward chronologically from this point. It will take time to recover the ‘waiting list’ . EMA will contact the first company with a product coming under the policy granted a CHMP opinion (September 2015).

Operational start up beginning

Learning curve for all parties

Preparation to Go-Live

Clinical Data Publication policy 100

Implementation status + issues

• End September 2016

Go live current forecast:

• Download + View only permanent public access linked to the relevant

regulatory decision

Public website:

• Industry co-operation
• Data utility: Anonymisation vs redaction only, improve CCI

understanding

• Secondary analysis:
• Regulatory effect, public scrutiny of published material
• Any negative public health impact, publications with misleading

conclusions

• Advance info to EMA, not obligatory, reciprocal transparency
• Public/stakeholder response
• EMA Annual report on experience of the policy’s implementation,

to include a list of the non-compliant companies

• Non-compliance: In the event of non-compliance by companies at

the various stages of the process, EMA will take remedial action including, where appropriate, publication of a non-compliance notice

Consequences

Conclusions

Clinical Data Publication policy 101

• EMA commitment to transparency, regulatory decision basis of

medicinal products

• Proactive publication enables public scrutiny: establishes trust,

confidence

• Better public information: Public access enables application of new

knowledge in future research, increases efficiency of medicine development, learning from experience

• Avoids clinical trials duplication: limits unnecessary patient

exposure

• Enhanced scientific knowledge/value of secondary analysis: sharing

scientific knowledge, contribution to public health, confirmation of regulatory decisions/need to review regulatory decisions taken.

Policy Benefits:

Thank you for your attention

European Medicines Agency

30 Churchill Place • Canary Wharf • London E14 5EU • United Kingdom

Telephone +44 (0)20 3660 6000 Facsimile +44 (0)20 3660 5555 Send a question via our website www.ema.europa.eu/contact

Further information

Follow us on @EMA_News

Data Sharing PlaUorm Workgroup

Presenter: Ida Sim, University of California, San Francisco Moderator: Brian Bot, Sage Bionetworks

4/5/16 103

The MRCT Center’s Data Sharing Workgroup Members

4/5/16 104

IT Work Stream Co-Chairs: Ida Sim (UCSF) Barbara Bierer (MRCT Center) George Alter (U of Michigan) Munther Baara (Pfizer) Kris Bolt (MRCT Center) Brian Bot (Sage Bionetworks) Anne Claiborne (IOM) Khaled El Emam (U of OSawa) Nick Ide (NIH) Ghassan Karam (WHO) Michael Khan (U of Colorado) Sean Khozin (FDA) Rebecca Kush (CDISC) Rebecca Li (MRCT Center) Gene Lichtman (HCRI) Michelle Mancher (IOM) Chris Mavergames (Cochrane) Eric Perakslis (Takeda) Frank Rockhold (Duke)

Data Sharing PlaUorm IT Workgroup: Main Objec2ves

• Workgroup objec2ves

– develop the Data Sharing blueprint – make recommenda2ons that would enable broader data sharing of clinical trials data

• In this presenta2on we will define the suggested scope,

technical requirements, administra2ve recommend- a2ons and suggested developmental stages of the Vivli plaUorm.

• We are not making any up front recommenda2ons or

assump2ons about building the solu2ons versus partnering with exis2ng technologies

105

106

Approach to the Work

4/5/16

Use Case Documents –Define scope, u2lity and feature set of IT plaUorm including:

• Data Analysis Use Cases
• Outline how data users will interact

with the plaUorm to find, request, and analyze data

• Analyzing published and

unpublished data from studies

• Data Submission Use Cases
• Outlines how data generators will

deposit data in the plaUorm

• Addresses anonymiza2on and data

standardiza2on issues

MRCT Center IT Workgroup

PlaForm Use Cases Data Sharing PlaUorm

107

MRCT Center IT Workgroup

PlaUorm Use Cases Data Sharing PlaForm

Approach to the Work

4/5/16

• Review where new func2onali2es enable

greater sharing and new paths to data analysis

• Host in-person workgroup mee2ng at IOM
• Create Data Sharing PlaUorm blueprint

Most Important Topics that Came Up

PlaForm Topics IPD Intake and CuraHon Must be able to intake and curate data with flexibility to scale to new types of data Catalog Metadata and Dataset IdenHfiers Must have accurate structured metadata and unique iden2fiers for each dataset, to enable precise granular searching Review Process for Requests Must be able to accommodate different review requirements and processes AnonymizaHon and DUAs Cri2cal to maintain par2cipant privacy Analysis Tools and Workspace Must have a secure analysis space, enabling as much cross-dataset aggregra2on as possible, with the flexibility to develop accommoda2on of many types of tools

4/5/16 108

Use Case: Data Browse, Request, Access

PrecondiHons PlaForm AcHons Post CondiHons

• Requester has an area of

interest

• Account is not needed for

searching and browsing Search interface supports granular queries about study design features Requester iden2fies studies for which they are interested in the IPD

• Requester has account on

the PlaUorm

• Requester has analy2c plan

for data request Work flow to support review of request Request approved or denied

• Requester has an approved

request Confirm that IPD is available and curated to comply with request; authen2cate access; provide secure analy2c environment Requestor has sufficient access to analysis tools and IPD and associated study materials to complete desired analysis

109

Data Requester searches PlaUorm, iden2fies studies of interest, submits a request for access to available Full Data Package and/or Post Publica2on Data Packages

New Func2onality to Enable New Paths to Data Analysis

• Current search interfaces do not support accurate

precise queries on study design features, e.g.,

– studies tes2ng meUormin in women with PCOS and Stage 1-3 CKD – studies tes2ng carvedilol for heart failure 6 and 12 months ajer acute myocardial infarc2on

• Need to provide ability to express more precise queries

– e.g. beyond medicine tested, medical condi2on or study phase

• Need study design informa2on (i.e., study metadata) to

be more structured and computable, and accurately completed

4/5/16 110

Vivli can bring a value-add by building an enhanced catalog that provides structured computable informa,on on study features not searchable elsewhere When data are more discoverable and more usable, they will enable more analysis

4/5/16 111

Discovering Datasets

Catalog Metadata Elements

• Will only curate Metadata for those

studies where IPD is or will be made available

• Defining a data model that adequately

captures the catalog metadata elements in a computable way

• Capturing eligibility criteria is a more

difficult task

• Capture criteria on lab values,

descrip2ons of prior treatments, details of included or excluded co- morbidi2es, and even temporal restric2ons

1. Sponsor Internal Unique ID 2. Study Title 3. Acronym 4. Secondary IDs (e.g. NCT number) 5. Study Design Type 6. Study Start Type 7. Study Comple2on Data 8. Sponsor 9. Responsible Data Contributor 10. Sample Size 11. Loca2on of Study 12. ParHcipant Features 13. IntervenHon Features 14. Outcome Features

112

See Appendix 2 of Blueprint

Use Case: Submission of Catalog Metadata

PrecondiHons PlaForm AcHons Post CondiHons

• Study registered (e.g.

ClinicalTrials.gov or ICTRP)

• SubmiJer has access and

authority to complete submission process Assign PlaUorm Iden2fier Human (Phase 1) or machine-assisted (Phase 2) cura2on of study documents New study is fully described and discoverable in plaUorm PlaUorm can aJest to some level of accuracy in data based on cura2on process

113

PI of study or Delegated data submiJer(s) with established account requests to submit data and sign contributor agreement

Metadata Cura2on Pipeline

Phase 1: Human CuraHon

• Learning from

ClinicalTrials.gov, ICTRP, etc.

• Quality assurance to validate

assump2ons

• Mechanisms for contributors
• r sponsors to review and

validate

Phase 2: Semi-automated CuraHon

• phase 2 = natural language

processing, text mining

• More computable elements

extracted from protocols

• Enough metadata curated to

allow for analysis on just the metadata

4/5/16 114

Use Case: Submission of IPD

PrecondiHons PlaForm AcHons Post CondiHons

• Catalog metadata for the

study is already in the PlaUorm

• Data contributor has

decided which sharing scenario (A or B) to use for this IPD dataset Mint Digital Object Iden2fier Update all study DOI links Enter study into IPD cura2on queue Study priori2zed in IPD cura2on queue Study is next up for cura2on in IPD cura2on queue If Scenario A:

• QC for anonymiza2on
• mapping to IPD standard

format If Scenario B:

• mapping to IPD standard

format Anonymized, curated IPD is fully indexed to the other contents of the PlaUorm, and are accessible for PlaUorm func2onality

115

PI of study or Delegated data submiJer(s) With established account requests to submit data and sign contributor agreement

Overview of Vivli plaUorm data sharing scenarios

116

Vivli PlaForm

Catalog Metadata Non-IPD

e.g., protocol, SAP,

Individual PaHent Data (IPD) Analy2c Environment and Tools SAS R etc.

Query, Request, and Analy2cs User Interface IPD IPD Non-IPD R

Scenario A Scenario B

Non-IPD SAS

Possibility to aggregate datasets across SAS environments

Data Requester

PlaUorm Schema2c: Scenario A

117

• Allows Non-IPD and IPD data to be

stored by the PlaUorm for analysis and combina2on with all other data that is shared on the plaUorm

• Ajer submission of IPD and other data

data, host does not have further

• ngoing maintenance costs for data

sharing.

Data Requester

Vivli PlaForm

Catalog Metadata Non-IPD

e.g., protocol, SAP,

Individual PaHent Data (IPD) Analy2c Environment and Tools SAS R etc.

Query, Request, and Analy2cs User Interface IPD

Scenario A

Non-IPD

PlaUorm Schema2c: Scenario B

118

• Allows non-IPD data to be copied to the plaUorm

for analysis and combina2on with all other data shared on plaUorm

• Scenario B does not allow IPD data to be

combined/analyzed with data outside of their analysis environment which provides tools and space for requesters

Vivli PlaForm

Catalog Metadata Non-IPD

e.g., protocol, SAP,

Individual PaHent Data (IPD) Analy2c Environment and Tools SAS R etc.

Query, Request, and Analy2cs User Interface IPD Non-IPD R

Scenario B

SAS

Possibility to aggregate datasets across SAS environments

Data Requester

Use Case: Submission of IPD

PrecondiHons PlaForm AcHons Post CondiHons

• Catalog metadata for the

study is already in the PlaUorm

• Data contributor has

decided which sharing scenario (A or B) to use for this IPD dataset Mint Digital Object Iden2fier Update all study DOI links Enter study into IPD cura2on queue Study priori2zed in IPD cura2on queue Study is next up for cura2on in IPD cura2on queue If Scenario A:

• QC for anonymiza2on
• mapping to IPD standard

format If Scenario B:

• mapping to IPD standard

format Anonymized, curated IPD is fully indexed to the other contents of the PlaUorm, and are accessible for PlaUorm func2onality

119

PI of study or Delegated data submiJer(s) With established account requests to submit data and sign contributor agreement

IPD Cura2on Queue: Priori2za2on and Cura2on Flow

120

PrioriHzaHon order for datasets submiJed under Scenario A

• 1. Any dataset for which there is an approved request
• 2. IPD datasets that are already anonymized and standardized
• 3. IPD datasets that are already anonymized, but need mapping to an IPD standard
• 4. IPD datasets that are neither anonymized nor standardized

CuraHon flow to achieve curated, anonymized IPD

3

Individual Participant Data (IPD)

in SDTM not in a standard model to be put into SDTM by Vivli by sponsor leave unstandardized

curated, anonymized IPD

anonymized in SDTM anonymized in SDTM basic anonymization

1 2

2a 2b

Phased Development Approach

4/5/16 121

Authen2ca2on / Iden2ty Management

Review process for requests Review brokering: referral to qualified external reviews Single plaUorm managed review process Harmonized review process Repository metadata & data set iden2fiers Unique study iden2fiers Secure storage Data set DOI Smart dataset cita2on & academic credit, archiving Dataset provenance External APIs

(e.g., to/from services)

Data intake & cura2on Core rich metadata IPD data model standard Human cura2on Harmonized common elements, non-registered studies, genomics, non- interven2onal studies Machine assist core metadata, imaging data Anonymiza2on and DUAs Standard DUA, contributor agreement, anonymiza2on verifica2on PlaUorm performed anonymiza2on service. Pa2ent level 1 way hash

Referral to approved 3rd party anonymiza2on service(s)

Analysis tools & workspace R, SAS & others Visualiza2on, advanced analy2cs, collabora2ve work, scien2fic workflows External APIs

(e.g., to open source tools)

BROKER

search, browse, requests, access, analysis Phase 1 Min: Dec, 2017 Phase 1+ Phase 2 possibili2es

Complex Assembly

• Interlocking technical, policy and business

considera2ons

• Long-term sustainability

– Services to be developed

• Exis2ng plaUorms and func2onali2es

– Iden2fy possible partners

• Considera2on for proposed publishing requirements,

forthcoming policy discussions

4/5/16 122

Q&A

• Agreement on approach
• Challenges and opportunity
• Scenarios
• MVP
• Use cases

4/5/16 123

Facilitated Discussion

4/5/16 124

Facilitated Discussions

• Please join your assigned group for in-depth discussion

and opportunity to provide feedback on Governance, Business Models and Data Sharing PlaUorm

4/5/16 125

A Vision Comes Together

Presenter: Mark Barnes, MRCT Center

4/5/16 126

Wrap Up Remarks

Presenter: Jeremy Farrar, Wellcome Trust

4/5/16 127

Dinner: 6 PM

Royal College of General Prac,,oners 30 Euston Square, London

4/5/16 128