Interoperability Solution - Hybrid Update -- From Pahe II and III to Post Market Medical Record Systems

2/2014
MedDATA FOUNDATION © 2013 –
ORIGINAL Distributed August 2013
1
"For an idea that does not at first seem insane, there is no hope."
- Albert Einstein
The views expressed herein are solely those of Stephen A. Weitzman, J.D. LL.M.
Executive Director of MedDATA Foundation.

 Leaflets and PPIs (Physician Package Insert ) cannot provide
full prescribing information given what we know today
about the response of patients by phenotype, genotype and
other omics.
 Current labeling therefore cannot meet FDA or other legal
standards of “Adequate Directions for Use” or absence of
False and Misleading Information (including omissions).
 In the age of personalized precision medicine are blanket
warnings or precautions adequate now that we know that
individual patients, because of “omics,” respond differently
in terms of adverse events (in degree) and effectiveness
(degree)?
 In that case is there sufficient information about the patients
who participated in the clinical studies for the prescriber to
make the "risk benefit decision" for their patients?
2/2014 MedDATA FOUNDATION 2

 How do we make maximum use of the data down
to the granular level for the benefit of patients?
(PCAST 2010)
 Given the legal framework noted above embraced
in the FD&C ACT and the comparable statutes of
all nations, what is the baseline for sharing dossier
or NDA submitted information?
 How do we achieve maximum use of data without
harming incentives for research and discovery?
 How do we overcome obstacles to exchanging
electronic clinical data, bothe HIT & Governance.
2/2014 MedDATA FOUNDATION 5

 A platform and methods for sharing data in a
way that it can be analyzed for all the good
purposes
 Standards and Common Data Models for all
disease areas
 Incentives for the Pharma and Healthcare
Systems Silos to share data (The Silos include
holders of post market patient medical records
and pharma companies that hold the pre-market
data that shows safety and efficacy or
lack thereof)
2/2014 6
MedDATA FOUNDATION

2012
1956 2025
MedDATA FOUNDATION 8
2/2014

Federal Aid Highway Act of 1956
MedDATA 9
I Guess Not!
2/2014

POST MARKET MEDICAL RECORD DATA
PREMARKET CLINICAL DATA SHOWING SAFETY AND
EFFECTIVENESS OF THERAPIES
MedDATA FOUNDATION 2/2014 11

It is argued that we do not have data standards. That
is not true. We do have medical record formats in
current use by pharma companies by which data is
collected in clinical studies and submitted to FDA or
EMA for evaluation of safety and efficacy. If we use
these data structures then we can collect and merge
post market data with premarket data in the same way
that FDA evaluates data.
It is time to create incentives for pharma to make
disclosure – full transparency – of protocols and
clinical data of approved therapies available to
advance creation of the next generation of therapies.

Central Database
Distributed Database
Hybrid Database

 Data in medical records are collected into a
Central Database for Querying and Analysis
 The database is the GPRD/CPRD with millions
of patients and over 60 million records
 The database is to be expanded to 55 million
patients

1. Data is kept in the hands of the original data holders
2. Decrease proprietary and liability concerns
3. Decrease risk and severity of data breaches
4. Data holders know their data; improve value and better
interpretation of findings
5. Minimize data transfer; minimum necessary
6. Voluntary – data partner autonomy
7. Reciprocity – value for participation
8. Partnership
9. Well-defined purpose

1- User creates and
submits query
(a computer program)
2- Data partners retrieve
query
3- Data partners review
and run query against
their local data
4- Data partners review
results
5- Data partners return
results via secure
network
6 Results are aggregated
17
2/2014

1. Data must be kept in the hands of the original data holders –
In the U.S. we will never get a central database – but we can get close!
2. Decrease proprietary and liability concerns – Can be handled
3. Decrease risk and severity of data breaches – Disagree
4. Data holders know their data; improve value and better
Interpretation of findings – Disagree
Data in distributed system is not uniformly indexed or coded
5. Minimize data transfer; minimum necessary – Security Issue
6. Voluntary – Data partner autonomy - Same as 1
7. Reciprocity – Value for Participating: Access more data
8. Partnership
9. Well-defined purpose

1 – Mirror
Data and
2 Index
1. Data held by partners is
mirrored at their location
(Silo)
2. Mirrored data is "reindexed"
24/7 in a uniform manner
using NLP and
Auto-Coding
3. Indexes (inverted files) of
partners are aggregated in
central computer 24/7
4. User selects data sources
and creates and submits
query to "central" portal
5. Query locates data in the
partner sites through the
central index
6. Data relevant to the query is
aggregated in a cloud
7. Analytics is applied to
generate the report
8. Results are obtained and
published with reference to
sources of data (trail)
9. Data is erased
Data Partner 4 – Select
Data Sources;
Run Query
8- Obtain
Results
Data Partner
Mirrored
Data and
Index
Mirrored
Data and
Index
Mirrored
Data and
Index
Mirrored
Data and
Index
5 - Central
Catalog - Index
Data Partner
Data Partner
Data Partner
Data Partner
7-Aggregate
Data, Analyze,
and
Index Path
Data Path
9- Erase Data
3
6
2/2014
Mirrored
Data and
Index

1. Researcher formulates
logical query
2. Researchers system
translates query using
3. Metadata services
4. Researcher identifies data
sources and submits query
to "central index" portal
5. Data in the partner sites is
located
6. Data relevant to the query
is aggregated in a cloud
7. Common data elements are
matched (ala SHARP) and
analytics applied to
generate the report (User
can use its own analytics
engine)
8. Obtain results and publish
with reference to sources of
data (trail) - and log query
9. 9. Erase data
Researcher 1
5- Run Query
Central
Catalog Index
8- Obtain results
9- Erase Data
4
Data Partner 1
6
2. Researcher Formulates
logical query
Translates logical query
to physical query
Mirrored and
Enhanced Data
Researcher Formulates
Logical Query
Translates logical query
to physical query
6
5
Data Partner (2)
Mirrored and
Enhanced Data
7-tAggregate
data for answer and
analysis
5
Researcher (n)
4
Data Partner (n)
Mirrored and
Enhanced Data
Researcher selects databases, uses the chosen query system, uses the chosen analytics.
2/2014
5
5
6

 The System is Data Agnostic, and Query System Agnostic
 Can access all available data for that user based upon data use agreements
 Data is kept in the hands of the original data holders (Same as distributed)
 Hybrid system is more efficient - Scalable (New Silos add Pointers to Index, “Catalog”)
 Hybrid system can obtain results faster
 Hybrid system can be multi-purpose
 Outcomes Research (CER)
 Drug Safety Signaling (surveillance)
 Personalized medicine
 Make Clinical Research More Efficient
 Rapidly design and implement observational trials
 Quickly and affordably conduct randomized studies
 Significantly reduce usual expenses associated with start-up and shut-down of
clinical research studies
 Identify patients for clinical studies
 Data is uniform – NLP and Coded to Snomed-CT
 Reciprocity – value for participation (Same as distributed)
 Partnership (Same as distributed)
 Well-defined purpose (Same as distributed)

1. Recognition that pharma is global and a solution
needs to be adopted globally with EU and U.S.
pharma taking the first steps on trial data and
owners of medical record systems agreeing to
share their data. (Or is it not really the patients
who own the data?)
2. Adoption and expansion of CDISC standards for
all disease areas based on BRIDG and finish
SHARE (Shared Health and Research Electronic
Library).
3. Capitalize on the SAS platform for clinical trial
data.
2/2014 23
MedDATA FOUNDATION

4. Designate a trusted 3rd NGO party to run a
global entity to administer data sharing in an
efficient sustainable model.
5. The 3rd party coordinates data sharing so that
qualified researchers can pose questions and
do analytics without release of personal
information to provide research papers based
upon information.
2/2014 24
MedDATA FOUNDATION

Interoperability Solution - Hybrid Update -- From Pahe II and III to Post Market Medical Record Systems

More Related Content

What's hot (20)

Similar to Interoperability Solution - Hybrid Update -- From Pahe II and III to Post Market Medical Record Systems (20)

Recently uploaded (20)

Interoperability Solution - Hybrid Update -- From Pahe II and III to Post Market Medical Record Systems