Date post: | 06-May-2015 |
Category: |
Health & Medicine |
Upload: | philip-payne |
View: | 133 times |
Download: | 0 times |
The CIELO Project: Towards a Research Analytics Commons
Philip R.O. Payne, PhD, FACMI
Professor and Chair, College of Medicine, Department of Biomedical InformaticsProfessor, College of Public Health, Division of Health Services Management and Policy
Associate Director for Data Sciences, Center for Clinical and Translational ScienceExecutive-in-residence, Office of Technology Commercialization and Knowledge Transfer
2
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
3
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
Critical Dimensions of a Learning Healthcare System: Systems Thinking Applied to Patient Centered Research
4
Environment and Culture
• Instrumenting the clinical environment
• Generating hypotheses
• Creating a culture of science and innovation
Precision Medicine
• Rapid evidence generation cycle(s)
• ‘omics’• Analytics/decision
support
Data Science
• System-level analyses• Data science• Visualization• Reproducible
analytics
Integrated and High Performing
Healthcare Research and Delivery Systems
Learning from every
patient encounter
Leveraging the best
science to improve care
Identifying and solving
complex problems
Rapid Translation
Our Focus!
5
The Need for Reproducible Research
Creating a high performing healthcare research and delivery system requires both economies of scale and increased efficiencies Timeliness Resource utilization Data “liquidity”
Central to this argument is a need to exchange research findings and evidence between and among stakeholders in a consumable manner Design Data Analysis
Doing so allows for reproducible research with cumulative benefits
6
Why is it Hard to Reproduce Research?
Data sharing alone is insufficient to this task How was data pre-processed? What analytical workflows were utilized? What additional parameters influenced data analysis? How were results “packaged” for dissemination?
Many socio-technical barriers to addressing these questions, including: Intellectual property and data-level concerns Availability of technology platforms/tools Documentation Metadata Standards Many, many other issues…
7
A Community Dialogue
8
BD2K and the Vision for a Research Commons
Phil Bourne’s Vision (Associate Director for Data Science, NIH) “To foster an ecosystem that enables biomedical
research to be conducted as a digital enterprise that enhances health, lengthens life and reduces illness and disability”
Creation of a commons providing for: Cloud infrastructure for data and computing Search Security Reproducibility standards App store
Source: Phil Bourne, “Ask Not What the NIH Can Do For You; Ask What You Can Do For The NIH”
9
Source: Phil Bourne, “Ask Not What the NIH Can Do For You; Ask What You Can Do For The NIH”
10
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
11
Translating a Problem into a Solution: The Problem Definition Process
Establish the Need for a Solution
Justify the Need
Contextualize the Problem
Write the Problem
Statement
Adapted from: Spradlin, “Are You Solving The Right Problem?”, HBR, September 2012
12
CIELO: Enabling Collaborative Data Analytics in Patient-Centered Research
Project Goals:1) Provide members of the research community with
access to an open-source/-standards “app store” for data analysis and software sharing
2) Reduce time and cost of research while enhancing the reproducibility and transparency of data analysis.
3) Evolve and meet emerging community needs
Blue-Sky: not grounded in the realities of the present: visionary <blue–sky thinking> (Merriam Webster Dictionary)
13
Process To-Date for CIELO
Ideation
MVP Development
Stakeholder Review and
Requirements Gathering
MVP Re-Engineering
Process and Outcomes Measure
Team formation and
proposal development
CIEHLO Prototype
Review andFeedback fromStakeholders
IterativeUser-Centered
Design
You Are Here!
Contextualizes
14
CIELHO Conceptual Model
15
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
16
Building on Existing Tools and Approaches
Sharing of Technical Artifacts
Social NetworkingMetadata
Github/Gitlab
Activity FeedsDiscussion Forums
FolksonomySemantic Search
Partitioning of access Bundling code and data Data model harmonization Cross-linkage (URIs/APIs)
Project-level feeds Linkage to metadata
Current ontologies Linkage to social functions
17
CIELHO Workflow Model
18
Community-Defined Requirements
Integration with analogous platforms and tools Ex. Sage Bionetworks Synapse
Incorporation of data security/confidentiality controls Particularly in the context of analyses involving PHI or similarly privileged data
sets
Convergence towards common data model for submission and reuse of data sets
Ex. OMOP
Multi-tiered sharing model Open access Limited access Private (for defined collaborators)
Semantic search and discovery of code and data
Connectivity to linked open data sets
Social networking at a project and individual level
Community-Defined Requirements: Focus for Public Beta Integration with analogous platforms and tools
Ex. Sage Bionetworks Synapse
Incorporation of data security/confidentiality controls Particularly in the context of analyses involving PHI or similarly privileged data
sets
Convergence towards common data model for submission and reuse of data sets
Ex. OMOP
Multi-tiered sharing model Open access Limited access Private (for defined collaborators)
Semantic search and discovery of code and data
Connectivity to linked open data sets
Social networking at a project and individual level
19
20
How Will We Evaluate the CIEHLO?
21
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
22
Future Directions: Shared Execution Environment (VAULT)
23
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
24
Meeting Objectives (1)
Provide a cross-section of stakeholders with a review of current technical functionality and design decisions surrounding the CIELO platform;
Identify needs for future functional/technical extensions to the platform, with a particular emphasis on:1) Shared analytic tool execution environments and
mechanisms2) Data model promotion/harmonization3) Crowd-sourced feedback and rating systems4) Minimum standards for “bundle” population (code
and example data)
25
Meeting Objectives (2)
Identify socio-cultural barriers and opportunities as they relate to creating and sustaining a CIELO user community;
Identify opportunity to promote and fund the ongoing development and adoption of CIELO as it can be positioned as a solution to enhancing research credibility and reproducibility.
26
Meeting Deliverables (1)
Stakeholder verification/validation of current CIELO functionality;
An enumeration and prioritization of future functional needs/requirements;
An enumeration of socio-cultural barriers and opportunities as they related to the creation and sustainability of an adopter/adapter community;
27
Meeting Deliverables (2)
An enumeration of communication, advocacy, and funding targets intended to position CIELO as a solution to enhancing research credibility and reproducibility;
An enumeration of targeted end-users and their communities;
A whitepaper and project plan that formalizes all of the preceding deliverables and provides a “roadmap” for future CIELO development and dissemination efforts.
28
Overview
1) Background and Motivation
2) Conceptualization of CIELO
3) Functional and Technical Architecture
4) Future Directions
5) Today’s Objectives
6) Discussion
29
“Information liberation + new incentives = rocket fuel for innovation” – Aneesh Chopra (The Advisory Board Company)
Philip R.O. Payne, PhD, [email protected]
"Without feedback from precise measurement, invention is doomed to be rare and erratic. With it, invention becomes commonplace” – Bill Gates (2013 Gates Foundation Annual Letter)
“Data is beyond simply quantifying, it is seeing measurement as the intervention” – Carol McCall (GNS Healthcare)