This document discusses challenges and opportunities in managing data for personalized medicine. It begins with an overview of personalized medicine and the role of information and biomarkers. There is currently a deluge of diverse data from sources like omics, IoT, social media and mHealth. Biomarkers and computational techniques help reduce complexity and support integrative models. However, effective data capture, integration and interpretation require addressing issues like interoperability, security and privacy compliance. Personalized medicine is transforming healthcare to be more data-driven and patient-centric.
VIP Call Girls Sector 67 Gurgaon Just Call Me 9711199012
Nikhil anmol pres_092014_1.0_final_2222
1. Biomarkers, Social Media and Personalized
Medicine: Security and Integration Challenges in
Managing the Data Deluge and Data Scarcity:
Nikhil Kumar, President & Founder, ApTSi
Anmol Limaye, Research Intern, ApTSi
1
2. Content
• Overview
• Personalized Medicine – a context
• Information and Personalized Medicine
• Security and Personalized Medicine
• Q & A
3. Why Personalized Medicine?
• Some Food for Thought
– Which is one of the greatest killers in the US today?
– What percentage of new drugs have serious, undetected adverse effects at the time of
approval?
– How many of the recently FDA approved medications were subsequently withdrawn
from the market or given a black box warning?
– What percentage of Americans have gene-based variations that significantly increases
the risk of having an ADR?
– What percentage of rare diseases are genetic in origin?
4. Key Guidances / Reflection Papers
• The FDA:
– FDA Guidance - Enrichment Strategies for Clinical Trials to Support
Approval of Human Drugs and Biological Products - Dec’12
– FDA report - Paving the way for Personalized Medicine, Oct 13
• The EMA
– EMA - Reflection paper on methodological issues with pharmacogenomic
biomarkers (for consultation until 25 Nov 2011)
– EMA - Approving oncology drugs in the era of personalized medicines,
Dec’11
• The Industry
– AMS, MHRA, Industry - Realizing the potential of stratified medicine. Jul 13
5. "When I use a word," Humpty Dumpty said in rather a scornful tone, "it
means just what I choose it to mean -- neither more nor less."
"The question is," said Alice, "whether you can make words mean so many
different things."
Figure from https://www.cs.cmu.edu/~rgs/alice-VII.html
6. So let’s define what we mean..
• Personalized medicine.. ‘a medical model that proposes the
customization of healthcare using molecular analysis - with
medical decisions, practices, and/or products being tailored to
the individual patient, often employing diagnostic testing to
personalize the treatment’ OR
• “the right patient with the right drug at the right dose at the right
time.”(EU) OR
• “Health care that is informed by each person’s unique
clinical, genetic, and environmental information” (AMA)
– The prevailing person-centric holistic model of modern Personalized
Medicine, where Omics play a part
Personalized medicine is more than pharmacogenomics, covering the
space of the omics and including biotic, chemical, physical and genomic
aspects
7. Some more terms..
• A companion diagnostic test
– essentially a biomarker test that enables better decision making on the use
of a therapy (and usually accompanies the PM drug)
• Biomarker..
– ‘An indicator of normal biological processes, pathogenic processes, or
pharmacological responses to a therapeutic intervention.’
• Stratification..
– ‘a medical model that proposes the customization of healthcare using
molecular analysis - with medical decisions, practices, and/or products
being tailored to the individual patient, often employing diagnostic testing to
personalize the treatment’ ..
PM is more than pharmacogenomics
Biomarkers and patient involvement play a growing role in Personalized
Medicine
8. Data
Analytics
Biomarkers…
• The Pharma industry and regulators are further emphasizing the
role of biomarkers in drug development
• For example
– Personalized Medicine Coalition Report, 2014
– A 57% increase in personalized drugs/ treatments from 2006 – 2014
– 30% of biopharma require all developing compounds to have a biomarker
– 50% of all clinical trials collect DNA from patients for use in biomarker
development
– Today, 137 FDA-approved drugs have pharmacogenomic information in
their labeling, and 155 total pharmacogenomic biomarkers are included on
FDA-approved drug labels
• Starting with Herceptin for breast cancer – to Vectibix recently
for metastatic colorectal cancer – we are moving ahead
1 Salter et al, 2014
OMICS
IOT, Social
Media
mHealth Wellness
9. PM Today..
Kumar1
OMICS
IOT, Social
Media
Data
Analytics
mHealth Wellness
10. Analytics
A Genomic System Model
OMICS
IOT, Social
Media
Data
mHealth Wellness
Methylomics
Transcrip-tomics
Proteomics
Methylation
Transcription
Genomics Metablomics
De-
Methylation
-mRNA Expression/ Splicing
- Alternative Splicing
- Allele specific expression
- microRNA Expression and
Discovery
Synthesis,
Degradation,
Transportation,
Translation Etc.
11. Content
• Overview
• Personalized Medicine – a context
• Information and Personalized Medicine
• Security and Personalized Medicine
• Trends and Advances
12. The Evolving Healthcare
Ecosystem is Person-Centric
FHIM focus
Payers
Healthcare IT
Patient
Internet of
Things (IOT)
BRIDG
HIPAA Business Associate & Covered Entity
Regulatory and Compliance
Providers
IDNs
Labs
Analytics
HIMMS & Continuaa introduce personal connected care
The new world of healthcare is person-centric
Pharma
Companion
Dx
ONC Direct Connect
PBM
Pharmacy
Social
Media
Data
Standards / initiatives
A Person-Centric model based on seamless interoperability, regulatory
compliance and security are the cornerstones of modern healthcare
OMICS
IOT, Social
Media
mHealth Wellness
13. The future of Healthcare…
Data
Analytics
The Modern Clinical World based on Personalized Medicine
Aetna accepts 100 genetic tests for genetic testing
http://www.aetna.com/cpb/medical/data/100_199/0140.html
FDA issues Personalized Medicine
guidance – 2013
OMICs Data
Clinical
Decisions
CDx narrows scope
Clinical Data
IOT, Lab and other
data
Clinical
Decisions
Systems
Physician
Engagement
Patient
Engagement
Outcomes
Wellness
The physician of the future is going to use Cdx’s and Clinical Decision
Systems to take Clinical Decisions.. And a focus on wellness and patient
engagement is going to shift the process and the quality
OMICS
IOT, Social
Media
mHealth Wellness
14. The advent of personalized
medicine….
• “personalized medicine” is here to stay
• There is a deluge of data
• Biomarkers, bioinformatics and IT are making this actionable
• Companion diagnostics and IT(big data, SOA) facilitate adoption
• Limited by business model issues, limited data sources and the
ability to analyze and use it reliably
• Enabled by support from regulatory bodies
• Person-Centric:
– Supports the empowered consumer and wellness!!
Personalized medicine is here to stay. It IS the future of medicine. And it
is data centric & person-centric. Capturing, translating, and interpreting
data are key success factors
15. A deluge of data…
Data
Analytics
“..healthcare is 17 percent of the US economy. It's
upwards of $3 trillion. The costs of healthcare are a
problem, not just in the United States, but all over the world,
and there are a great number of inefficiencies in the way we
practice healthcare. ” – Jason Lee, Director, Healthcare
Forum, The Open Group
• There is an exponential increase in data
• ePRO, internet of things and social media add to the variety!
• Predictive analytics and Integrative models gaining adoption
• Balance this against cost, agility and quality considerations!!!!
“..$1000 sequencing …$1,000,000 interpretation” Ken Davies
We need to reduce complex data into a model
that is accessible for human comprehension
Bryn Roberts
“All research data at Roche up to 2010 amounted to about 100 TB”.
During 2011/12, we ran a project called CELLO, where the genomes
from about 300 cancer cell lines were sequenced. Together with other
data from the cells, we generated 100 TB of data in this single
‘experiment’—equal to 100 years of Roche research up until 2010!” ..
Bryn Roberts
The effective capture and interpretation of this information will change
the practice of medicine
OMICS
IOT, Social
Media
mHealth Wellness
16. Biomarkers & computational
techniques are enablers
• …the problem cannot be solved (reasonably) with
Data
Analytics
CDER Biomarker
Program
traditional brute force techniques. So we must use new
ones..
– Biomarkers help reduce complexity and incorporate disease etiology
– New computational techniques provide a foundation for supporting
integrative models and bench to the bedside – necessary for successful
adoption of HIT
• Machine learning
• Reverse Markov models
• The list goes on…
– Standards provide a framework for interoperability
– Ontologies, vocabularies and metadata link it together
The appropriate use of biomarkers, ontologies, metadata and modern
computational techniques provide a framework to harness the data
FDA Guidance
on Biomarker
Development
(2014)1
1 FDA - Qualification Process for
Drug Development Tools
OMICS
IOT, Social
Media
mHealth Wellness
17. Content
• Overview
• Personalized Medicine – a context
• Information and Personalized Medicine
• Security and Personalized Medicine
• Q & A
18. The SOA Ecosystem pervades
HIT Ecosystem
… “Service Orientation”
is disruptive and here
Enterprise SOA
Cloud Computing
Modern
SOA
Ecosystem
Legacy APIs
(CORBA/DCOM)
EAI
Business Adoption and Impact of Service Orientation
A world of SOAs
Micro Service
Architectures/ APIs &
IOT
Low Increasing High
SOA RA
Kumar 1
Service orientation in its different flavors is creating a HIT fabric for
information exchange
1 Derived from Kumar, 2014
… And will be the
cornerstone of the HIT
world
OMICS
IOT, Social
Media
Data
Analytics
mHealth Wellness
19. New ways for gathering data...
• IOT – do we even know if the device is right?
– Who owns the data?
– Is it secure?
– OK now I have it – what does it mean?
• ePro and Social Media
– It really works in the world of wellness
– It really works in the world of drug adherence
– So how do we capture it and interpret it?
The coalescing of SOA and Business requires stakeholders from both IT
and the business to think Service Oriented. This presentation should
provide an introduction of the concepts involved.
0110
0110
01001
0110
1110
0110101
Unstructured data is 80% of data (Seth Grines)… and growing
20. Information Reference Model
Behavioral Data Interoperability
(Rules and Data behavior)
Semantic Data Interoperability
Syntactic Data Interoperability
Physical Data Store Data when assigned st ructure (syntax)
and semant icsbecomes
information
Behavioral
Data
Interoperability –
Data behavior is
consistent
RDF/OWL
Semantic Data
Interoperability –
Data shares the same
semantic implication
UDEF
Syntactic
Interoperability –
Data structures are
rationalized
Canonical Forms
&
Schemas
Data
interoperability
is a critical success
factor in the
effective leveraging
of data.
It is also a key
factor in the
reduction of the
overhead of data
mapping and the
creation of a
virtualized data
model.
Interoperability Reference Model
Structured Data Unstructured Data
Characteristics of the
interoperability layers
Kumar 2009
0110
0110
01001
0110
1110
0110101
21. Interoperability….
1. The evolving HIT world
involves a plethora of
ontologies.
2. Ontologies are controlled
vocabularies with
relationships between the
terms
3. Controlled vocabularies are
an accepted list of terms
4. Translation between
ontologies is a painstaking
but necessary process
5. Metadata is a fundamental
base for interoperability
6. In the future communicating
processes and services will
depend on this
interoperability
Without interoperability the data deluge is noise. Interoperability must
address structure, syntax and semantics.
0110
0110
01001
0110
1110
0110101
22. The practice of integration
• Metadata is key
• Assessing completeness for reliable decision
• Computational models to manage integration – address the
kinds of data
• Integration in practice (confidence, traceability, fact vs source of
truth)
0110
0110
01001
0110
1110
0110101
23. A Big Data Model for
Personalized Medicine
Regulatory (Governance, Security and Monitoring)
Acquisition
(Landing
& Staging)
Latency
Mediation,
MDM,
Transformation
&
Formatting to
Enterprise Model
Analytic Storage
Analysis/
Decision/
Consumption
Diverse
Data Sources
(Structured,
Unstructured)
at
Diverse Velocities
Kumar1
Kumar, 2013
0110
0110
01001
0110
1110
0110101
24. Integrative models and their
role
• What is an integrative model?
• Revisiting – PM is more than pharmacogenomics
• Computational implications of integrative models
0110
0110
01001
0110
1110
0110101
25. Content
• Overview
• Personalized Medicine – a context
• Information and Personalized Medicine
• Security and Personalized Medicine
• Q & A
26. CFR-21
NIST
The Elephant in the room…
• Security
• Privacy
• Ethics and adoption
Security and Privacy can be serious sources for overhead. Use them to
establish clarity and plan early
PRIVACY
Common
Controls
HIPAA
CERT
27. Compliance
• CFR21 Type 11 – access and retention
• HIPAA
– business associate or covered entity
– Privacy implications
• Safe Harbor
CFR-21
NIST
PRIVACY
Common
Controls
HIPAA
CERT
28. Why anonymity
Voters list, sales
prospects, etc.
Healthcare data
Ethnicity
Address
Diagnosis
Procedure
Sex
Birth date
CFR-21
NIST
PRIVACY
Common
Controls
HIPAA
CERT
‒Perfect anonymity
can never be
guaranteed
‒ But we can make it
hard
‒ Regulations require
it (HIPAA, CFR-21,
Data Protection Act,
etc.) !!!
29. Compliance and realities
• Compliance
– CFR21 Type 11 – implications
– HIPAA –business associate
• Trends
– Deidentification cannot be complete
– ePRO may or may not be private (PatientsLikeMe.com)
• What’s involved
– Secure your access, Don’t trust … insiders!!!
– Address Safe Harbor if you send the data out
– Log
CFR-21
NIST
PRIVACY
Common
Controls
HIPAA
CERT
30. Content
• Overview
• Personalized Medicine – a context
• Information and Personalized Medicine
• Security and Personalized Medicine
• Q &A
33. Why anonymity
Voters list, sales
prospects, etc.
Healthcare data
Ethnicity
Address
Diagnosis
Procedure
Sex
Birth date
CFR-21
NIST
PRIVACY
Common
Controls
HIPAA
CERT
‒Perfect anonymity
can never be
guaranteed
‒ But we can make it
hard
‒ Regulations require
it (HIPAA, CFR-21,
Data Protection Act,
etc.) !!!
34. De-id. and annonymity
• What is required?
CFR-21
PRIVACY
– Does not identify a person
– No reasonable basis to believe that the
information can be used to id. an individual
• 2 techniques
– Expert determination (obfuscation)
– Safe harbor (removal of id. parms)
NIST
Common
Controls
HIPAA
CERT
35. CFR-21
With large volumes of data…
• Both are used
• Expert determination includes:
PRIVACY
– K-anonymity coupled with t-closeness are
well known and normally acceptable
– Add obfuscation (one-way)
• Once encrypted you can’t identify it
NIST
Common
Controls
HIPAA
CERT
Hinweis der Redaktion
Biomarker.. ‘An indicator of normal biological processes, pathogenic processes, or pharmacological responses to a therapeutic intervention.’
Personalized medicine.. ‘a medical model that proposes the customization of healthcare using molecular analysis - with medical decisions, practices, and/or products being tailored to the individual patient, often employing diagnostic testing to personalize the treatment’
A companion diagnostic test is essentially a biomarker test that enables better decision making on the use of a therapy3
Numerous initiatives – public and private
interoperability driven by
e.g. Direct Connect
E.g. Standards standardization