Transcript Document

The CRIO at UW-Health
Umberto Tachinardi, MD, MSc
Chief Research Information Officer
UW Health
Associate Dean for Biomedical Informatics
School of Medicine & Public Health
Director, Biomedical Informatics Core
Institute for Clinical and Translational Research (ICTR)
UW-Health
Dean - CEOs
COOs
CRIO
BCG
CIO
HIMC
Clarity
Applications
IT
Help-Desk
CTSA BMI Core
Advisory
Commitees
Director
Umberto Tachinardi
MD, MSc
Council of CIOs
UWHC, MC, UWMF
Co-Director
TBN (MCRF)
Adm-Assistant
Sue Ann Hubanks
HIMC
Bioinformatics
Mark Craven PhD
Imaging Informatics
Elisabeth Burnside
MD
EDW
U Tachinardi MD,
MSc and P Peissig
MBA
Clinical/Health
Informatics
Eneida Mendonca
MD, PhD
Innovative
Technologies
Miron Livny PhD
IT Infrastructure
David Towers MBA
and MCRF
equivalent
User request
Triage/Routing
Packaged
Services/Prods
Yes
Packaged
Servs/Prods
available?
ICTR IT Infrastructure
No
Ad-Hoc EDW Queries
NLP
Project
Analysis
CTMS customizations
Image Processing (PACS)
Genomic Analysis
Interfacing
Data Storage
Data Security
Collaborations
Research Project
Analysis
Research/
Development
req’d?
Yes
Project
not-viable
No
BARD
Data Analysis
Regulatory/Compliance
Image Analysis
Project viable
Grant Proposal
Data Modeling
Consulting
Training
CRIS
TTRC
(Imaging Facility)
REC
BMIC
Services
Partnerships
Advanced Technologies (R&D)
Information
documentation, storage
and retrieval
Machine learning, including
data mining
Networking, security,
databases
NLP, semantic technologies
Image and Signal Analysis
Representation of logical
and probabilistic knowledge
and reasoning
Simulation and modeling
Software Engineering
Biomedical Informatics – proposed changes for next renewal
-
-
-
Goal: Define a common “BMI Core” for both CTSA and UWCCC
Opportunities
- Across the board Epic implementation (HC/MF/Unity)
- UW Health DW project (HIMC)
- OnCore integration
- Integrated IT infra-structure (servers, networks, storage)
Competitive Advantage
- HIMSS Level 7 (Epic)
- Strong Bioinformatics
- Improved Clinical Informatics (new Faculty)
- Coordinated human subject recruitment policies and tools
- Data Governance plan
- Marshfield Clinic partnership
Drivers
- Cost efficiency (non unnecessary redundancies, common standards for all
NIH funded projects)
- Research empowerment (trhough more robust and sophisticated tools and
services)
- Collaboration vs Competition (same teams working towards the same
objectives)
Core Services
HealthLink
(Clarity)
NLP
Bioinformatics
Security
Ontology
Interfaces
Finance
OnCore
(Cancer)
DW-Metadata
HealthLink
(Other)
OnCore
(ICTR)
DW-Ontology
DW-Hybrid
Molecular, images data
UWCCC Administrative Systems & IT
Infrastructure
UWCCC
Informatics
ICTR (BMIC)
UWCCC Bioinformatics
Clinical & Health, IT
SMPH
UW-Biomedical Informatics Services
Data Services (caBIG, i2B2, HL7, Portal, APIs, SQL, FTP)
Networking/Computing/Storage
Security
Honest Broker
OnCore
Epic
NLP
UW-DW
HIMC (UW-Health)
Ontologies
PACS
“omics”
Molecular Data
Images
Biosamples
Registries
External DBs
Literature
The Clinical/Research
Data Warehouse
The Hybrid Model
Semantic
ONTOLOGY
METADATA
Implementation
EAV
DIMENSIONAL
Integrated Design
Event
General
Event
Measurable
Event
Substance
RED
Observation
General
Observation
Measurable
Observation
Substance
EAV
Level 1 Rollup
Patient
Education
Community
Education
Patient Outcomes
Care Giver
Performance
Panel
Management
Case
Management
Clinical, Quality
and Safety
Measurements
Financial
Health
Measures
Quality and Safety
Improvement
Outcomes
Measurement
Population Health
Resource
Utilization
Service Line
Utilization
Improving Patient
Experience
Clinical Trials/
Research
Operational
Efficiencies
Patient Population Definitions
Clinical Patient Care
Revenue Cycle
Level 0 (Atomic Level)
Clinical
Effectiveness
Market Position
Revenue Cycle
Activity Definitions
Coverage/
Authorization
Patient
Tracking
Immunization
PCP &
Referrals
Patient
Survey
Repetitive
Services
Safety
Events
Treatment /
Procedure
Complaints
Risk
Assessments
Clinically
Administered
Medications
Tests
In-Basket
Communication
Follow-up
Retail
Pharmacy
Patient
Consult
Peer Review
Payment
Collection
Payroll
Costing
Budget
Charging
Accounts
DME
`
Pre Access/
Registration
Visits/
Appointments
Patient
History
Problem /
Medication
Patient
Resource
Needs
Vitals
Diagnosis
Plan Care/
Orders
Best Practice
Advisories
ETL and Data Integration
Registries/Data
Silos
Research
Collaborations
External to Clarity
External Data
Master Data
ADT
Transactions
Appointments /
Visits
Internal to Clarity
Tests/
Pharmacy
Accounts
Class
Relationship
Ontology Map
(26 levels)
Disease
Management
HIMC DW architecture (high level view)
Reports/Dashboards
Users
BO Universe
* Partially developed
** Planned
Security – Honest Broker **
Security – Auditing **
Ontology management **
NLP editing **
Reports
Red = non-HIMC
Blue = HIMC
Users
Security – Access Management **
Views *
Metadata *
Query
Tool **
Ontologies **
HIMC DW
Atomic layer *
Natural Language
Processing **
Other
Interfaces*
Healthlink
(Clarity/
Chronicles)
ETL*
Data mart *
The ecosystem for IT and
Analytics
IT
Analytics
Technology (i.e. Data
Center, Networks)
Data warehouse (HIMC,
Clarity)
Applications (i.e. email,
web, Epic, PeopleSoft)
Analytical tools (Natural
Language Processing,
Honest Broker,
Ontologies)
The differences between IT and anaytics
IT
Analytics
Mostly tactical and operational levels
Mostly strategic and tactical levels
Systems operated at this level tend to support operational
workflows (email, Epic, Peoplesoft), the strengths are processes
automation (i.e. CPOE, logistics, billing, HR). The technological
infrastructure (servers, networks, desktop computers) is part of this
area, it establishes the platform where the operational systems rely
upon. This area helps process optimization, efficiency and
productivity.
Systems developed in this area are mostly targeted at transforming data
into information and knowledge. This is mostly an strategic decision
support area. Reports, dashboards, simulations, depend a lot on the
business characteristics, the meaning and quality of data, the data
crunching performed on high volume datasets. This area tend to be less
technological, and much more business dependent, it will help
planning, benchmarking and discoveries.
Process drivers
Business drivers
CPOE, billing, admitting, scheduling, discharging are examples of
processes supported by this area. This area guarantees that security
and privacy are in place. It is mostly based on mission critical
systems. Highly regulated, users need to be well trained in
operating the systems. This area keeps the lights on.
Budget planning, root cause analysis, comparative effectiveness,
scenario simulation (a new hospital, a new ward, more nurses, etc),
personalized medicine, cohort finding, are examples of non-routine,
applications at the strategic and tactical levels supported by analytics.
This area helps plan for the immediate and distant future.
Engineering, customization
Architecture, development
The IT area is highly structured, depend on well defined and
implemented methods that are known by all (users and operators).
This is why most of the solutions are purchased from third-parties.
More than developing methods, this area apply them. Systems like
Epic, will be customized to be more adequate to the local
characteristics. The core of the system is untouched, only parts of
the system can be customized. This area requires more
management (big contracts, large teams, complex IT technologies)
and less development.
Different from conventional IT (data is shaped to satisfy computer
needs), analytics is closer to the human reasoning. Rather than
processes and workflows, this area deals with meaningful information
and knowledge. There are no commercial systems ready to fulfill the
needs of this area, therefore they need to be developed. There are
computational tools, available to help those developments, but they
assist mostly on how to transform and present information, identify
patterns, extrapolate and forecast. Analytics, because of the factors
described, is closer to be an architectural work.
The differences between IT and analytics
IT
Analytics
Management
Creation/Vision
The measures of a successful CIO are all related to excellence in
management. Perfect budgets, on time deliveries, regulatory
discipline, project and contract management. This well structured
mission, requires a risk-averse personality and a
standardized/predictable behavior, since most products support
mission-critical tasks in a very sensitive environment.
The CIO for Analytics, is a different type of professional. This person
needs to be highly motivated to work in a less structured environment,
he/she likes to take risks (and know how to select the reasonable ones),
be imaginative and innovative. The products of the analytics will be as
good as the capacity to predict future trends, motivate people based on
concepts and ideas, rather than repeat/improve what others are doing.
Evolutionary
Transformative
Like any other engineering process, this area produces tools and
solutions that will allow for incremental improvements (i.e. more
security, faster systems, better interfaces, more integration). Epic,
for instance, gets better in every upgrade. Networks are more
reliable, secure and faster every year.
Analytics is a parallel business to IT. While it depends on the existence
of a well built and managed IT environment, the purpose of the
analytics is to recycle data products (data stored on databases), in the
pursue of innovating, finding problems and opportunities, feeding
decision makers to change the business model, be it health care
business or clinical thinking.
Low Intellectual Property
Because most of the systems on IT are purchased from
vendors, there is very little potential for institutional IP
development.
High Intellectual Property
ROI (Return of Investment)
RONI (Risk of NOT Investing)
Conventional IT depends on positive ROI. Although, the real ROIs
are very difficult to calculate and prove, the industry is moved by
those metrics. It is a conservative approach, and CIOs are trained to
follow those predicaments (very conservative).
Health care is a business in transformation. The science is different, the
population profile is changing, the costs are not sustainable, the
patients demand different products, chronic disease management is a
new challenge. Not developing tools and solutions that will help
understand, plan and actually change the status quo becomes an
enormous risk!
The area of analytics provide an enormous potential for IP accrual.
Those IP assets can come from two sources: the methods and tools
developed in house, and also the findings (i.e. discoveries, high yield
datasets, best practices, risk management).
The “new guy”
CIO vs. CRIO…
CIO
EHR, ERP, Quality, PACS
JCAHO, AHRQ, MU, ACO, HIPAA, OSHA, HIE
Customers: Physicians, nurses, CXO, managers
HIMSS
1000s FTEs
EDW (Clinical, Quality)
CRIO
CTMS, IRB, Bio Banking, Grants Mgt, etc
I2b2, caBIG, VIVO, IRB/HIPAA, clinicaltrials.gov
Customers: Researchers, students, faculty
AMIA
10s-100s FTEs
EDW (Research, Education)
… and, IT vs. Analytics
IT
Operations and workflows
Technologies
Security
Predictability, reproducibility
Engineering, implementation
Analytics
Decision support
Knowledge
Risk
Unpredictability, innovation
Architecture, research
Research (CTMS) X Patient Care (EHR)
Regulatory
IRB
HIPAA
Storage
High Capacity /
Non fault-tolerant
Medium Size /
Fault tolerant
Data
Ownership
PI
Healthcare
provider
State
Offline
Online
Finance
Grants/Contracts
Billing
Registries/Panels/Datasets
Data Warehouse
Reporting
NLP
Semantic Mgt
Analytics (is this the CRIO?)
Data Governance
IT
Data Center
Networks
Support
Training
Security
Applications