caBIG® POSSCON March 23, 2011
John Speakman Chief Program Officer, CBIIT, NCI
john.speakman@nih.gov @speakman (views mine, not NCI’s)
The National Cancer Institute (NCI): cancer.gov
Why NCI is in the informatics space: 21st Century Biomedicine, all the P’s
How’re we doing? • Estimated US cancer deaths 2009: 562,340 (American Cancer Society)
• Estimated new US cancer cases 2009: 1,479,350 (American Cancer Society)
• Cost of cancer deaths: $960.7 billion in 2000, estimated $1,472.5 billion in 2020 (Journal of the National Cancer Institute, Dec. 9, 2008)
Biomedical Research is Decades Behind the “Knowledge Economy” Curve The cancer Biomedical Informatics Grid® (caBIG®) is a virtual network of interconnected data, individuals, and organizations that redefines how research is conducted, care is provided, and patients/participants interact with the biomedical research enterprise.
20th Century Research > Care Paradigm
• Biological pathways • Target identification and validation
• Candidate selection and Optimization • Pre-clinical testing • Phase I, II, III • New Drug application and Approval#
• Product launch
• Reporting of serious/fatal ADRs
• Clinical adoption #
• Re-labeling (or recall) as needed • Additional indications as warranted#
21st Century Learning Health System
Role of private sector • Ernst & Young, Progressions 2011: • Pharma 3.0: The health outcomes ecosystem • Growing pool of data will be generated through electronic health records, social media, online communities, wireless devices and smartphones, “meaning that no single entity will own or control all of the data about any company, product, disease state or patient behavior.” • “Value will come from developing insightful solutions with the data.” • Open, defined interfaces mean a “common market” • Focusing on interfaces, NCI has a core collection of non-viral, open source code (not abandonware) that can be reused
NCI also has an architecture: Services-Aware Interoperability Framework (SAIF) Framework for the development and use of data standards in a Services Oriented Architecture • • • • •
Human-readable statements about APIs Machine-testable definitions Platform-independent specifications Expanded metadata infrastructure Robust services framework
UTILITY”
“Infra /
“CORE” “CAPABILITY”“PROCESS”
NCI Enterprise sSOA Periodic Table of Services R
Pt
Oc
Po
E
Ae
Hx
Dx
Ds
Ra
Registration
Protocol
Study Outcomes
Patient Outcomes
Eligibility
Adverse Event
Hx and Physical
Discharge Note
Decision Support
Referral and Authorization
Cr
S
Tp
I
L
Rx
Sc
Credentialing
Specimen
Treatment Plan
Image
Lab
Pharmacy
Scheduling
Sd Qr SDTM
Data Query
C
O
P
Pa
D
A
Correlation
Organization
Person
Protocol Abstraction
Disease
Agent
Km Cm Ev Va Tx Au Knowledge Management
Contract Management
Enterprise Vocabulary
Validation
Translation
Audit
Id Id Management
Mp Ay Master Problem List
Tr
11Trust
Management
Allergy
Aa Py Authorization Authentication
Policy
caBIG® Core Principles • Open Access • Open Development • Open Source • Federation • Over 70 tools developed
Open Source License for NCI-funded caBIG® Software: Lawyer Talk for “BSD-style, non viral” •
• • • •
•
• • •
Source code, object code and related documentation must be available at no charge to the public with “unlimited rights”, i.e., the rights to disclose, reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly in any manner and for any purpose (FAR Section 52.227-14, “Rights in Data – General”). End users have no rights to use NCI-owned trademarks or the names “The National Cancer Institute” or “NCI” to endorse or promote products derived from NCI-funded caBIG® Software. There is no affirmative obligation to distribute or otherwise make available the source code of derivative works. “Non-viral” – no reciprocity requirement for distribution of derivative works (“modifications”) Choice of distribution method for source code of derivative works is an economic decision rather than a contractual term required by NCI. Closed distribution may represent a valid choice depending on the type of software product (applications vs. databases vs. infrastructure). Distribution of NCI-funded caBIG® Software that is bundled, combined, integrated or packaged with components that are licensed under reciprocal open source licenses (e.g., GNU GPL) would subject NCI-funded caBIG™ Software to such provisions, and thus conflict with NCI funding terms. Instead, developers of NCI-funded caBIG® Software that needs such components to function should provide instructions to end users to install those components separately. End user documentation must acknowledge the author. If there is no end user documentation, derivative software product itself must acknowledge the author. Licensee is solely responsible for the consequences of incorporating caBIG™ Software into thirdparty software Comprehensive disclaimer of expressed and implied warranties.
2011 Inflexion point: Scope has increased
2011 Inflexion point: People want to contribute
2011 Inflexion point: Appetite for funding of all software development
What we think we need • Community • We need to build and maintain a true open-source development community • We also need to maintain engagement with the user community (i.e., not just actual/potential developers)
• Content • We need to coordination “community” (writ very large) development of software, including extensions to existing caBIG applications and infrastructure
• Connectivity • We need to support and guide developers • We need to review submitted code to ensure consistency with caBIG principles and architecture
Other things we have learned / are learning • Not everyone who says they need functionality, need functionality • Maybe an open source community, with sufficient separation from NCI, can serve as a market
• Not everyone who thinks they can code, can code • Maybe an open source community can be a nonjudgmental way
• Making something open source doesn’t necessarily make it cheaper • But maybe it can make it easier to scale
We’ve been here already
More .org, less .gov
Find us at: • http://cabig.cancer.gov/ - public site • http://cabig.nci.nih.gov/ - community site • john.speakman@nih.gov • @speakman (views mine, not NCI’s)