Going With The (Knowledge) Flow: The Future Of Decision Making In Improving Outcomes
Ian Z. Chuang, MD, MS, CCFP
SVP, Healthcare Informatics and Chief Medical Officer
Dennis Morrison, PhD
Chief Clinical Officer
CARERECORDTM: A platform providing a
framework for EMR, billing, scheduling, and clinical workflows
CARECONNECTTM: Beyond
interoperability, focused on integration, a messaging bus
ensuring an open and free flowing network
CAREVIEWTM: Providing specific
needs to extend capabilities,
connect disparate communities and optimize user experiences
CAREPATHWAYSTM: The use
of data to drive optimal outcomes and operational efficiencies
CAREMANAGERTM: Coordinating
care to ensure healthcare access, coordination, affordability and outcomes
PLEXUSTM: On demand services,
technology and delivery providing predictable results as needed, when needed
CAREPOINTETM: Person-centric solutions to
enable and empower the consumer,
providing a seamless integrated connection in the pursuit of recovery, health and
wellness.
CARERECORDTM: A platform providing a
framework for EMR, billing, scheduling, and clinical workflows
CARECONNECTTM: Beyond
interoperability, focused on integration, a messaging bus
ensuring an open and free flowing network
CAREPOVTM: Providing specific
needs to extend capabilities,
connect disparate communities and optimize user experiences
CAREPATHWAYSTM: The use
of data to drive optimal outcomes and operational efficiencies
CAREMANAGERTM: Coordinating
care to ensure healthcare access, coordination, affordability and outcomes
PLEXUSTM: On demand services,
technology and delivery providing predictable results as needed, when needed
CAREPOINTETM: Person-centric solutions to
enable and empower the consumer, providing a seamless integrated connection in the pursuit of recovery, health and wellness
We are our own worst enemies
Social Services MH/SU Medical
Recipients of Services Clients Consumers Families People Individuals Client Consumer Patient Patient Providers of
Services Direct Service Providers Social/Case Workers
Clinicians Therapists Providers Doctor Nurse Provider
Interventions Supports Services Treatment Therapy Treatment Therapy
Intervention Documentation
Service Plan Permanency Plan Individualized Service Plan
Treatment Plan
MH/SU
Medical
Family &
Social
Services
Key Point about CDSS
An EHR is NOT an electronic
recapitulation of a paper record
The EHR can and should do things for
you that the paper record can’t.
It’s not just about EHRs anymore
CDSS 1.0
Why do we need CDS?
It is now humanly impossible for unaided
healthcare professionals to possess all the
knowledge needed to deliver medical care
with the efficacy and safety made possible
by current scientific knowledge.
Definition Clinical Decision Support
Systems
Clinical Decision Support Systems are:
“…active knowledge systems which use
two or more
items of patient
data
to
generate case-specific
advice
"
Definition Clinical Decision Support
Systems
Put Another Way:
If
X
is true and
Y
is true:
CDSS 2.0
Who gets the decision support information?
• CDSS 1.0 – Clinicians/caregivers • CDSS 2.0 – Consumers – Family Members – Clinicians/caregiversHealth care and Social Services need to be
transformed
Less expensive professionals
to do
more sophisticated things
in
The Least Expensive Provider is
CONSUMERS ARE ALREADY
DOING DECISION SUPPORT
Data Deluge
What’s the new role of care
providers when consumers have
more data?
Eric Topol, MD
“…[the doctor’s] role will be progressively
morphed into providing guidance, wisdom,
experience on
how to transform data and information to
knowledge and judgment.”
Netsmart Clinical Model
External Data Source (EBT) Mobile/Telemed Social Networking Consumer Input Portals/PHRs Internal Data Source (PBE)Pre-Intake Intake Service
Plan ProgNotes Outcomes
Clinician Decision Support
Consumer/Family Decision Support
DECISION SUPPORT
FUNCTIONALITY
Meaningful Use Opens the Door for CDS
• CDS = Clinical Decision Support • Meaningful Use = a start
– Phasing: Stage 1 vs Stage 2 – Functional capabilities focused
• Necessary but NOT sufficient
– Success is based on implementation of specific CDS interventions tied to Clinical Quality Measures
(CQMs)
What is a Clinical Decision Support System?
• Information technology – based functionality
designed to improve clinical decision-making
– System generated information based on user data input and/or available data
– Presented back to the user – Functionality Focused
• How a decision support will work within an EHR system?
• How will the clinician user interact with the decision support functionality?
Clinical Decision Support: The Bigger Picture
• Objective Focused
– Process – Outcome
• Knowledge-Driven Care Process
– What are we trying to achieve? Desired outcomes? – How can an intelligent system prompt, guide or
influence the clinician user to the desired outcomes
• Knowledge Flow is key
Clinical Decision Support as a Process
• Referencing the way the nervous system works:
– Alerts and information display interventions are the efferent arms
• Based on the available data, specific message is pushed to the
clinician user at the point-of-care
– Dashboard, reports, and benchmark are the afferent arms
• Data from the point-of-care come back into the central data base • Analysis for knowledge-driven process improvement
• Both perspectives are necessary to close the loop
CareGuidance
• A broad set of integrated decision support
capabilities (not just alerts) that interact with the right user at the right time to have the most
impact of making optimal decisions and actions
– At the point-of-care, the caregiver is the target user – At home, and in the community, the consumer, their
proxy, health coaches are the target users
– Care Managers and Operations Managers are target users at a system and population level
Clinician Workflow Opportunities for CareGuidance
Pre-Intake Intake Service Plan Progress Note Outcomes
• Risk identification • Assessment • Reference • Work routing • Progress • Performance • Assessment • Progress • Compliance • Treatment Guideline • Therapy • Test • Medication • Reference • Targeted cohort • Scheduling • Hallmark events
Components of CareGuidance
• Knowledgebase
– Reference data in computer-usable form
• Codesets • Criteria • Metadata
• Trigger
– Event
– Date/time vs time interval – Data
• Software algorithm • Notification output
– Recommendation/guidance/content output
Information Based Decision Support
• Four levels of information detail
– Identify – Inform – Educate – Guide
• Different levels of functional intervention
– Passive info display
– Workflow alerts/interruption
– Workflow interruption with action
Decision Support imbedded into Workflow Functionality
• Reference information or guidelines are actionable content with the EMR
• Complex set of decision support functionality structured like EMR functional components, such as Order Entry
– Order sets
Interactive Forms/Diagnostic Tool
• Interactive functionality for real-time decision support
• Clinician provides data as requested by the tool • Upon completion of data input, the tool runs an
algorithm to generate an output
– Clinician uses the information
– The output may be something that is actionable and ready for user selection
• One advantage is the ability to run different scenarios
Online Reference
• Digital reference libraries
– Research paper – Textbooks
– Info aggregators, such as Up-to-Date
• Functionality
– Info on the side
Prognosis/Prediction
• Predicting health risk or events, such as risk of death
– Framingham – APACHE
Challenges and Obstacles
• Information overload
• Lack of precision/lack of relevance • Lack of time
• Topically correct, but wrong user audience
• Decision support is only as good as the data that is available/inputted
The Data Matters for CareGuidance
• Garbage in-garbage out
– Erroneous data
• “cold” as in feeling cold or a viral respiratory infection?
– Imprecise data
• “diabetes”, meaning diabetes mellitus or diabetes insipidus
– If diabetes mellitus, which one? Type I, Type II, Gestational?
– Data omission
• Is the absence of a specific data interpreted as “not present” or
“not stated”? Does it matter?
• Rules-based CDS can be biased towards specificity (more
precision), or sensitivity (more inclusion but higher rates of false positive) depending on how they are designed
CareGuidance Data Strategy
• Structure or codify the data needed for CareGuidance in pre-defined ways
– Support synonym
– Support related concepts described at different levels of granularity via concept identifiers and data
dictionary
• “anemia” “folate deficiency anemia”, and “severe iron deficiency
anemia” are related if we don’t care about the type of anemia
• String match won’t work so easily
• Same decision support rule can support different but related data
Additional Requirements for CDSS
• Data codified for software algorithm
• Data codified for monitoring and reporting • Recording decision support actions for
monitoring and reporting
• Data codified to track outcomes
Two Sides of Knowledge-Driven Care Process
• If we want more evidence-based practice, we need more practice-based evidence.
Evidence-Based Practice Guidance/ Suggestion
Practice-Based Evidence
“do that for this reason”
When All is Said and Done
“If it was your mom receiving care in the middle of a complex health system, given the option,
wouldn’t you prefer to have a decision support system checking on optimal care as a back-up?”