Transcript How to Promote EHR in Research

Agenda
• Clinical Research within an EMR
• What the difference between an EMR and an EHR
• Define Clinical Research
• Practical Clinical Networks (Single Source)
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Why do we want to do this
• Recruitment
• Patient are recruited from existing local sites
• Lower Start Up Time and Cost
• Trails use existing networks and infrastructure
• Allows for simple trials in clinical setting
• Patients are more representative of general population
• Unique information for Longitudinal Data, Registries
• Lower Cost of Data Collections
• Eliminates Source Document Verification
• Eliminates redundant data entry
• Lower Monitoring Costs
• Reduces queries and time spent on query resolution
• Builds in quality checks
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Agenda
• Clinical Research within an EMR
• What the difference between an EMR and an EHR.
• EMR. = Used by a clinician to provide care
• EHR. = Synopsis of individual health related information
• Define Clinical Research
• Practical Clinical Networks
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Centralized Model – EMR
* EMR = Used by a clinician to provide care
* EMR = Hosted at a central site
* EMR = All data is resides centrally
Columbia
Duke
Boulder
Bronx Leb
U of CO
EMR
Monti
PorterStarke
Aurora
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Federated Model – Electronic Health Record
* EHR. = Synopsis of individual health related information
* EMR. = All data is collected and stored locally
* EHR. = Select clinical data (Medications, Allergies) is transferred to central site
Data
Warehouse
Firewall
Columbia
Duke
Colorado
Aurora
SVTN
Bronx Leb PorterStarke
Monti
WCMC
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Federated Model – Electronic Health Record
* EHR. = Synopsis of individual health related information
* EMR. = All data is collected and stored locally
* EHR. = Select clinical data (medications, Allergies) is transferred to central site
Data
Warehouse
Firewall
Columbia
Duke
Colorado
Aurora
SVTN
Bronx Leb PorterStarke
Monti
WCMC
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Agenda
• Clinical Research within an EMR
• Define Clinical Research
• ePRO, EDC,
• Practical Clinical Networks
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Research - Definitions
Paper
Source Medical Record
Case Report Form
Patient Rated Scales
Electronic Equivalent
Electronic Health Record (‘EHR’)
Electronic Data Capture (EDC)
Electronic Patient Reported Outcomes (ePro)
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Case Report Forms
Inclusion Criteria
Males
Age 18 – 40
Type I Diabetes
Oral Hypoglycemic
Hamilton Depression Scale
INSOMNIA EARLY
0. No difficulty falling asleep
1. Complains of occasional difficulty falling asleep, i.e. more than 1/2 hour
2. Complains of nightly difficulty falling asleep
INSOMNIA MIDDLE
0. No difficulty
1. Patient complains of being restless and disturbed during the night
2. Waking during the night - any getting out of bed rates 2 (except for purposes of voiding)
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Electronic Patient Reported Outcomes (ePRO)
How do you feel today?
Excellent
Good
Fair
Poor
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Present Process Clinical Trials - Paper
(Monitoring Cost > 15% of a trial)
Site
Monitor
Data Entry
(Data Management Cost > 15% of a trial)
QA
Coding
Analysis
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Clinical Trials – EDC (Electronic CRFs)
X
Double
Data
Entry
Source
(EMR)
Monitor
CRF
X
QA
DDE
Queries
Coders
QA
Coding
Analysis
EDC
Site
Monitor
Data Entry
Analysis
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Single Source
X
Double
Data
Entry
Single
Source
Monitor
X
QA
DDE
Queries
Coders
Analysis
X
X
CRF
EDC
Site
Monitor
Data Entry
QA
Coding
Analysis
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Single Source - Advantages
Single
Source
Queries
Monitor
Site
Monitor
Analysis
Coders
Data Entry
QA
Coding
Analysis
• Increase Recruitment
•Patient are recruited from existing local sites
• Lower Start Up Time and Cost
•Trails use existing networks and infrastructure
•Allows for simple trials in clinical setting
• Lower Cost of Data Collections
•Eliminates Source Document Verification
•Eliminates redundant data entry
• Lower Monitoring Costs
•Reduces queries and time spent on query resolution
•Builds in quality checks
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Agenda
• Clinical Research within an EMR
• What the difference between an EMR and an EHR.
• Define Clinical Research
• Practical Clinical Networks
• Barriers to Implementation
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Barriers to Implementation
The limitations to creating a single source platform are formidable.
• Technical limitations include:
• Security issues of traversing both inbound and outbound firewalls at different
institutions
• EMRs are not interoperability
• EMRs are not research ready, don’t understand concepts like a visit schedule, and
ODM protocol
• Evolving software standards form both HL7, CDISC IHE, CCHIT and other
organizations
• Regularity barriers including meeting both
• Cost to HIPPA, CCHIT, IHE and
• Cost to be FDA compliance - 21 Part 11 compliant.
• Workflow barriers include the need to function both as a
• EMR – not structured, information based, work flow
• Case Report Forms are a defined subset of elements
• Loss of efficiency for busy clinicians who see research patient.
• Functionality who should build this
• Unique data requirements of a clinical record and a clinical trail.
• EDC does not do billing, treatment plans, state and local requirements
• EDC does not what to host an EMR with liability
• EDC do not know EMR
• EMR don’t care about research and market is very small
Present System – Different Stake Holders
• Clinical
– What is the differential diagnosis?
– What is the best treatment?
• Management
– How many patients did he see?
– Am I in regulatory compliance?
• Research
– Are you collecting the data I need?
– Is the data accurate?
Clinical
Management
Research
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Role-based Configurability - Clinician
End Slide
Role Based Configuration – Research Assistant
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Practical Clinical Trial Network - Challenges
• Central Model
• Individualizing EMR to Local Needs
• ADT/Billing at Local Site
• Regulatory Needs of Local Sites
• HIPAA issues (Retention of Medical Records
• Maintenance of EMR Records
• Federated Model
• Controlling Local Variations Impact on Clinical Trials
• Data Manager – Coordinates Data transfer
• Version Controls – Keeping Application Current
• Decision Support – Distribution of Requirements
• Data Transfer – Security and Firewall Issues
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
CRF’s collect data that is not part of an EMR
Unique Data Elements
EMR
Common
EMR
Common
EDC
EDC
Diagnosis
Medications
Medical History
Adverse Reaction
Severe Adverse Reaction
Visit Schedule
Treatment Plan
Social History
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Integration of existing software
End Slide
MindLinc – Registry Manager
Data
Warehouse
Registry Manager
Firewall
Columbia
Duke
Colorado
Aurora
SVTN
Bronx Leb PorterStarke
Monti
WCMC
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Electronic Medical Record
5
4
3
MindLinc
2
1
Patient Profile
Guidelines
De-identified Data
Interface Engine
Decision Support Engine
Local Clinical Repository
Data Transport
Business Rules
Firewall
Centralized Data Warehouse
Queries
Outcomes Data
Phase 1
Phase 2
Phase 3
Study Design
1 2 3
XXX
Data Analysis
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Single Source
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Improving Clinical Care
Clinical Care
Pateint
Knowledge
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Outcomes Data Warehouse
Largest Longitudinal Psychiatric Outcome Data Warehouse in existence
180,000 patients and 1,700,000 encounters
Outcomes Data Warehouse
Data is De-Identified IRB exempt
Data is surfed (cleaned)
Commercially Available
Published data 20+ studies, journal, posters, etc.
Industry, Academics, marketing, discovery.
Example mining for medications side effects – SSRI side effects distribution (%)
Celexa
Lexapro
Paxil
Paxil-CR
Proza
c
Anxiety / Agitation
4.2%
3.7%
8.0%
2.8%
2.9%
1.9%
GI Distress Nausea
9.5%
6.7%
5.2%
7.0%
3.8%
5.6%
Increase Appetite Weight Gain
3.4%
5.0%
3.9%
1.4%
4.2%
3.3%
Sexual Dysfunction
7.3%
8.4%
8.1%
1.4%
6.5%
4.3%
Sedation
10.7%
7.0%
9.6%
8.5%
5.8%
6.3%
Patients
506
299
385
71
554
793
Zoloft
Visualization
Patient
Local Clinical Repository
De-Identified
Aggregate Data
Visualization
Decision Support Engine
Expert Consensus
Phase 1
Phase 2
Phase 3
123
XXX
Data Repository
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Decision Support
Treatment Dashboard
Patient MRN # MR1233
Depression*
Age:80 | Sex: Female | Race Black
Panic Disorder
Profile of like patients and their response
CGI I
Rx
Filters
Demographics
Treatment Recommendation - Evidence)
# of Pts
Age
Age < 18
Race
White
2.50
bupropion
1356
Medication - (CR, CP, DW)
Age > 60
Black
3.16
venlafaxine
1897
Medication - (CR, CP, DW)
Age 18-60
sertraline
2345
Medication – (EBM, CR, CP)
Other
3.67
3.67
sertraline + CBT
1755
MeD + CBT – (EBM, CR, CP)
3.67
sertraline + CBT
578
ECT – (EBM, CR, CP)
Gender
Male
Female
Risk Factors
Hospitalized
Chronic Illness
History of SI attempt
SSRI
Diagnosis
Substance Abuse
Bipolar Disorder
Relationship
Impulse
Thought Disorder
SNRI
TCA
Bupropion
ECT
3/94
1994
S
3/95
1995
3/96
1996
3/97
1997
Outpatient
1998
Inpatient
S
Stressor
Exit
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
The End
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences
Standards Based Codified Data is the Key
• Clinical
• Re-imbursement
• Efficiency
• Decision Support
Clinical
• Management
• Revenues
• Regulatory
Data
• Research/Clinical Trials
• Quality Data
• Quicker Trials
• Lower Costs
Management
Research
DUKE UNIVERSITY MEDICAL CENTER
Department of Psychiatry and Behavioral Sciences