Contenu connexe Similaire à The Future of Personalized Health Care: Predictive Analytics by @Rock_Health (20) The Future of Personalized Health Care: Predictive Analytics by @Rock_Health1. R E S E A R C H
2014 OCT 27
PREDICTIVE ANALYTICS
THE FUTURE OF PERSONALIZED HEALTH CARE
2. A R O C K R E P O R T B Y
E DISTINCTLY REMEMBER THE MOMENT THAT SCIENTISTS CLAIMED
Wvictory against all nature of future disease after the human genome had successfully been decoded. However, over
the ensuing decade-plus, it has become clear that our health is
not quite that deterministic. Clinicians must weigh not just a
string of nucleotides when making decisions about our care, but
must also incorporate a growing set of health data that is
generated and controlled by patients. Incorporating this data
into health care to enable better decisions is at the heart of this
report. The benefits of using predictive analytics are the same
as many categories of digital health: better care and lower costs.
The difference is that the path to realizing these benefits—
through personalized care—is only possible by implementing
these technologies. The concern that care will be reduced to a
set of algorithmically-derived probabilities is important and
real. But the promise is as well.
AUTHORED BY WITH HELP FROM
MALAY GANDHI
@mgxtro
TERESA WANG
@teresawang6
ROCK HEALTH is powering the future of the digital health ecosystem,
bringing together the brightest minds across disciplines to build
better solutions. Rock Health funds and supports startups building
the next generation of technologies transforming healthcare.
ROCK HEALTH partners include Abbott, Blue Shield of California,
Boehringer Ingelheim, Deloitte, GE, Genentech, Harvard Medical
School, Kaiser Permanente, Kleiner Perkins Caufield & Byers, Mayo
Clinic, Mohr Davidow Ventures, Montreux Equity Partners,
Qualcomm Life, UCSF and UnitedHealth Group.
LEARN MORE AT rockhealth.com
LAUREN DEVOS
@lauren_devos
3. PRESENTATION © 2014 ROCK HEALTH
Contents
SECTION
4 Background Definition of predictive analytics and personalized health care
Scope of report
6 Landscape Core technologies used in predictive analytics
Venture funding of predictive analytics companies (2011-Q3 2014)
Landscape of predictive analytics companies
16 Direction Examples of predictive analytics in health care
Requirements for personalized health care
22 Challenges Key advancements in predictive analytics in health care
Case studies of digital health companies
31 Considerations Healthcare industry use cases
Regulatory and adoption constraints
38 Acknowledgements Contact information
4. [Genome science] will
revolutionize the diagnosis,
prevention and treatment of
most, if not all, human
diseases.” PRESIDENT BILL CLINTON
“
Remarks on the completion of the first survey of
the entire human genome (June 26, 2000)
5. Nearly fifteen years later, it is obvious that health care is far more
complex than simply understanding our DNA
PERSONALIZED MEDICINE PERSONALIZED HEALTH CARE
Treatment (through drugs) FOCUS Prevention, intervention, and treatment
Molecular DATA Demographic, social, administrative, clinical,
“ If I wanted to be a doctor today
PRESENTATION © 2014 ROCK HEALTH
Right MANTRA Best
Deterministic MODEL Probabilistic
5
molecular, patient-generated/reported
Figuring out how to get the right
drug to the right person at the
right dose at the right time.”
I’d go to math school not to
medical school.”
“
DR. FRANCIS COLLINS VINOD KHOSLA
DIRECTOR, NATIONAL INSTITUTES OF HEALTH VENTURE CAPITALIST
7. Predictive analytics is not reinventing the wheel. It’s
applying what doctors have been doing on a larger
scale. What's changed is our ability to better measure,
aggregate, and make sense of previously hard-to-obtain
or non-existent behavioral, psychosocial, and
biometric data. Combining these new datasets with
the existing sciences of epidemiology and clinical
medicine allows us to accelerate progress in
understanding the relationships between external
factors and human biology—ultimately resulting in
enhanced reengineering of clinical pathways and truly
personalized care.”
VINNIE RAMESH
Chief Technology Officer
Co-founder
Wellframe enables
health plans and
healthcare providers to
better manage clinical
and financial risk, while
augmenting the impact
of their existing care
resources
“
8. PREDICTIVE ANALYTICS is the process of learning
from historical data in order to make predictions
about the future (or any unknown)
FOR HEALTH CARE, predictive analytics will enable
the best decisions to be made, allowing for care to
be personalized to each individual
9. Our report focuses on how predictive analytics is directly
impacting patient care
PRESENTATION © 2014 ROCK HEALTH
THIS NOT THIS
• Clinical decision support
• Readmission prevention
• Adverse event avoidance
• Chronic disease management
• Patient matching
9
• Actuarial modeling for rate /
premium setting
• Advertising and purchasing
• Customer satisfaction and retention
• Business decision modeling
• Fraud
10. The goal of predictive analytics in any field is to reliably predict
the unknown
WHEN WILL I DIE?
PRESENTATION © 2014 ROCK HEALTH
10
PREDICTION
CERTAINTY
WHAT DID I EAT
TODAY?
HOW WILL MY
BLOOD SUGAR
CHANGE?
HOW MUCH
WEIGHT WILL I
GAIN?
WILL I GET
DIABETES?
WHAT COMPLICATIONS
MIGHT I SUFFER FROM?
KNOWN UNKNOWN
11. In fact, “predictive analytics” underlies most of traditional
medicine and health care, whether technology-enabled or not
PRESENTATION © 2014 ROCK HEALTH
11
TRAINING DATA
AGGREGATION
• Cleanse
• Tag and/or label
• Structure
RELATIONSHIP
SEARCH
• Identify attributes that act as predictors
• Develop algorithms
Acute
Chronic and
preventive
CASE DATA
COLLECTION
• Collect predictive attributes for specific case (e.g., a
patient) Symptoms Risk factors
INDIVIDUAL CASE
CHARACTERIZATION
• Apply algorithms derived from training data to
case attributes of the patient
• Describe an unknown Diagnosis Stratification
RECOMMENDATION
CONTEXTUALIZATION
• Apply specific recommendations based on ‘who’,
‘when’, ‘where’, etc. Treatment Intervention
PERFORMANCE
CAPTURE
• Define success
• Record results relative to recommendation
• Improve algorithms for characterization and
recommendations
Outcome Outcome
1
2
3
4
5
6
Note: Preventive care includes management of chronic diseases
IN TRADITIONAL
MEDICINE AND
HEALTH CARE
12. The overabundance of data and widespread availability of tools
has catalyzed predictive analytics in health care
PRESENTATION © 2014 ROCK HEALTH
BIG DATA
Expected growth in healthcare
data, 2012-2020 (petabytes)
25,000
500
Source: American Medical Informatics Association
DATA MINING
DATABASES/WAREHOUSES
BIG DATA PLATFORMS
12
2012 2020
ALGORITHM PRODUCTION
SERVICE PROVIDERS
AGGREGATE SERVICE PROVIDER VENTURE FUNDING: $1.8B
13. Investors certainly believe in the promise, pouring $1.9B into
companies that purport to use predictive analytics
MOST ACTIVE INVESTORS
PRESENTATION © 2014 ROCK HEALTH
Venture funding for companies using predictive analytics (2011-Q3 2014)
$902M
13
PREDICTING FUNDING
$520M
$300M
$201M
2011 2012 2013 Q3 2014
NOTABLE
DEALS
• Khosla Ventures
• Merck Global Health
Innovation Fund
• Norwest Venture Partners
• Sequoia Capital
• Social+Capital Partnership
Source: Rock Health funding database
Note: Only includes deals >$2M
14. Funded companies claiming to use predictive analytics are highly
focused on providers, practically ignoring patients
ENTERPRISE SHARED PATIENT
PRESENTATION © 2014 ROCK HEALTH
14
KYRON
USER OF ANALYTICS
COMPANIES
Source: Company websites
Note: Only includes companies that received venture funding from 2011 to Q3 2014;
companies are selected, not comprehensive
15. New data streams, including those direct from patients, are
beginning to be used by companies for predictive analytics
6%
SO MUCH DATA
Percentage of venture-backed predictive analytics companies using various types of data (2011-Q3 2014)
CLINICAL CLAIMS PATIENT-GENERATED PATIENT-REPORTED RESEARCH MOLECULAR CLINICAL TRIALS
PRESENTATION © 2014 ROCK HEALTH
15% 14%
26%
42% 42%
71%
15
Source: Company websites
Note: Percentages do not sum to 100%; companies may collect multiple data types
Current data sets generally revolve
around claims but that’s going to be
changing with lots of clinical data and
transactional information with lifestyle
becoming more readily accessible.”
SAM HO, M.D.
Chief Medical Officer, UnitedHealthcare
17. Familiar methods of predictive analytics with a long history in
other technology services are also appearing in health care
PRESENTATION © 2014 ROCK HEALTH
CORRELATION CONTEXT ACTION
Source: “Giving Viewers What They Want” The New York Times (February 24, 2013)
17
• Movie preferences (by rating,
viewing history, etc.) are
gathered across all users
• Viewers who liked movie A also
liked B, C and D and since you
like A, so you’ll probably also
like B, C, and D
• Historical viewing is labeled
and identified by individual
viewers
• You tend to watch movies on
weekends and TV shows on
weekdays, so a movie should
be suggested on Saturday
• Larger data sets on preferences
that are based on real world
viewing are collected
• Audiences have a high likelihood
of enjoying a type of TV show, so
an entire season can be
purchased instead of just a pilot
Hom-Lay Harish Add Profile
18. Symptom calculators are the “recommendation engines” of
health care, helping millions of consumers diagnose themselves
PRESENTATION © 2014 ROCK HEALTH
Source: Mayo.com
Note: Other use cases are representative, not comprehensive 18
HOW IT WORKS
Consumers enter in their
symptoms, and related factors,
and in turn receive the diagnoses
with the “most matches”
OTHER USE CASES
• Triage
• Comorbidity identification
• High cost patient identification
• Physician-patient matching
CORRELATION
19. Lacking appropriate context, clinical indicators—including vital
signs—can generate false positives or negatives in alert systems
HOW IT WORKS
HOW HEART RATE RESPIRATORY RATE Lucile Packard Children’s Hospital
CONTEXT
Stanford adjusted its early warning
algorithms to match actual vital
signs from hospitalized children
versus textbook definitions
Using textbook definitions, 14% to 38%
OTHER USE CASES
of heart rate observations and 15% to
• Decompensation
30% of respiratory rate observations
would have resulted in false alarms
• Readmission prevention
• Behavior change
Source: “Development of Heart and Respiratory Rate Percentile Curves for Hospitalized Children” Pediatrics (2013),
“A ‘Green Button’ For Using Aggregate Patient Data At The Point Of Care” Health Affairs (2014)
Note: Other use cases are representative, not comprehensive
19
PRESENTATION © 2014 ROCK HEALTH
20. Genetic screening companies similarly know the inherent risks
before a child is conceived, allowing decisive action
PRESENTATION © 2014 ROCK HEALTH
20
ACTION
HOW IT WORKS
A couple planning its family submits
DNA to Counsyl, which provides
probabilities on 100+ health
conditions that could be passed from
parents to children
OTHER USE CASES
• Disease prevention
• Population health management
and early intervention
• Treatment selection
Source: Counsyl.com
Note: Other use cases are representative, not comprehensive
21. Building models that break the curve of uncertainty will lead to
personalized care, but it is not without significant challenges
KEY REQUIREMENTS
PRESENTATION © 2014 ROCK HEALTH
21
MOVING FORWARD
PREDICTION
CERTAINTY
KNOWN
UNKNOWN
Using predictive analytics to personalize health care
• Incorporation of new data
types and sources
• Reliability of predictive
models
• Timeliness of data
• Transparency in prediction
• Convenient (and in context)
recommendations
• Rapid learning and
improvement
Personalized care
will emerge from high
confidence algorithms that
can predict actionable
interventions that improve
long-term health outcomes
UNKNOWN
23. “The keystone of any successful predictive analytics
model is the ability to improve the prediction based on
a feedback loop.
Within seconds, Google knows whether its search
engine prediction is correct. But in health care, the
feedback loop—which is often measured in terms of
impact on biometric or cost outcomes—can take
years.”
CHRISTINE LEMKE
Co-founder and CEO
The Activity Exchange is
the connective tissue
between healthcare
companies and their
populations to build and
manage relationships to
improve outcomes.
24. Startup companies are attacking the key challenges in predictive
analytics, advancing the space and creating differentiation
PRESENTATION © 2014 ROCK HEALTH
1
2
3
4
5
6
TRAINING DATA
AGGREGATION
RELATIONSHIP
SEARCH
CASE DATA
COLLECTION
INDIVIDUAL CASE
CHARACTERIZATION
RECOMMENDATION
CONTEXTUALIZATION
PERFORMANCE
CAPTURE
BASIC ADVANCED EXAMPLES
Limited Disparate
Traditional data Novel data
Lagged / point Real-time / continuous
Obfuscated Transparent
Generic Personalized
Disjointed Closed loop
There are a whole bunch of
variables and very few
observations.
The number one thing holding
predictive analytics back is the
lack of data: the fact that things
are not easily measured,
collected, or accessible.”
URI LASERSON
DATA SCIENTIST, CLOUDERA
PHD IN GENOMICS
24
“
25. Aggregating, cleansing, and labeling data from disparate sources
is the building block for developing non-obvious predictions
CASE STUDY: ONCOLOGY CARE EXAMPLE: CHALLENGES
DATA AGGREGATION RELATIONSHIP SEARCH DATA COLLECTION INDIVIDUAL CASE CONTEXTUALIZATION PERFORMANCE CAPTURE
PRESENTATION © 2014 ROCK HEALTH
BASIC ADVANCED
Single oncology data source (e.g.,
clinical trials, claims, or electronic
health records)
DATA TYPES Aggregating data from EHR,
laboratory and billing systems
Integrating and matching claims with
patient trial data
Avoids data that isn’t already
cleansed, structured, and labeled
(e.g., claims, pre-designated fields in
EHRs, etc.)
CLEANSING Identify high value data (e.g. EHR
notes) and cleanse, structure, and
label it as part of the aggregation
process
Data is historical with inherent bias
from unintended use and lag
associated with claims processing
FREQUENCY Data is loaded on a nightly basis and
processed continually, near real-time
Source: Company website 25
• Ability to access meaningful, historical
data sets and normalize for inherent
biases and validity concerns
• Integrating with current clinical workflow
to collect real-time, point of care patient
data
• Learning to manage and process new and
existing forms of unstructured, siloed data
• Addressing HIPAA and privacy related
concerns to guarantee patient anonymity
26. Using new data sources creates an opportunity to surface better
(i.e., more accurate, timely, or cheaper to collect) predictors
CASE STUDY: CARDIAC REHAB EXAMPLE:
CHALLENGES
BASIC ADVANCED
Printed packets of information and
cardiac rehabilitation guidelines are
handed to patients to follow
DATA AGGREGATION RELATIONSHIP SEARCH DATA COLLECTION PERFORMANCE CAPTURE
PRESENTATION © 2014 ROCK HEALTH
26
• Identifying existing and novel data points
that can better predict outcomes
• Identifying reliable and not spurious
relationships
• Rapid data collection and continual
integration to train and iterate algorithms
• Ability of predictive analytics to integrate
and impact a clinician’s work flow
• Limited data sources are analyzed by care
providers
INDIVIDUAL CASE CONTEXTUALIZATION
Source: Company website, interviews
ENGAGEMENT
MODALITY
Mobile app tracks patients’
interaction with the cardiac rehab
program, which is linked in real-time
to a care management dashboard
Engagement and clinical data
collected infrequently through office
visits and in-person interactions
ACQUISITION Collects additional data via activity
trackers, meal logging, and non-diagnostic
mental health questions
Poor, incomplete data sets limits a
clinician’s ability to identify patients
likely to be readmitted or suffer
adverse event
PRIORITIZA-TION
Algorithm predicts patients who need
more attention and sends alerts to
clinicians or care coordinators to take
action
27. Real-time data collection reduces traditional intervention
response time
CASE STUDY: HIGH-RISK PREGNANCY EXAMPLE:
CHALLENGES
DATA AGGREGATION RELATIONSHIP SEARCH DATA COLLECTION PERFORMANCE CAPTURE
PRESENTATION © 2014 ROCK HEALTH
Source: Company website, interviews 27
• Current care model predates that data
collection happens in discrete intervals
with an additional lag due to claims
processing
• Data collected may vary in reliability and
accuracy if based solely on patient
reporting or non-clinical devices
• Using real-time data in a meaningful
manner requires new infrastructure and
workflow
INDIVIDUAL CASE CONTEXTUALIZATION
BASIC ADVANCED
Regular check-ups generate claims
data that get processed several weeks
to months later
PREDICTOR
VARIABLE
SOURCE
Patient self-reports weight and mood
data on a frequent basis, which is
immediately accessible to care
provider
Infrequent, missed appointments
results in missed data points
RELIABILITY Decreased lag time between weight
measurement and processed
information
Lagged and infrequent data results in
late recognition and interventions
TIMING Timely data allows for early
stratification and intervention to
avert high-risk complications
28. Improving the transparency of methodologies and the data
behind analytics better supports physicians in decision-making
CASE STUDY: CDS TOOLS CHALLENGES
DATA AGGREGATION RELATIONSHIP SEARCH DATA COLLECTION INDIVIDUAL CASE PERFORMANCE CAPTURE
PRESENTATION © 2014 ROCK HEALTH
Source: Company website, interviews 28
CONTEXTUALIZATION
BASIC ADVANCED
Clinical decision support based on
limited set of protocols and
guidelines
BREADTH Incorporates the current scientific
research and clinical practice data for
analytics
Guideline updates significantly lag
clinical research and require approval
through centralized bodies
ADJUSTMENT Real-time analytics and continuous
updates based on outcomes from
observational data
Medical practice highly paternalistic
and substantiated through
experience versus evidence
VISIBILITY Transparency via medical knowledge
graph to support physician decision-making
regarding symptoms,
medications, risk factors, and
diagnoses
• Visualization challenges in displaying all
relevant data for time sensitive decision-making
• Finding the balance between black box
engines and information overload tools
• Recency and accessibility of data to
develop medical, evidence-based
recommendations
• Physician and patient adoption of
“algorithms” dictating care
EXAMPLE:
29. By tailoring both recommendations and timing, companies can
motivate consumers via a personalized toolset
CASE STUDY: 10,000 STEPS EXAMPLE:
CHALLENGES
DATA AGGREGATION RELATIONSHIP SEARCH DATA COLLECTION CONTEXTUALIZATION PERFORMANCE CAPTURE
PRESENTATION © 2014 ROCK HEALTH
Source: Company website, interviews 29
• Using advanced algorithms and behavioral
economics theory requires large
individually tagged datasets
• Determining the “right intervention” is
challenging and requires trial and error
• Consumer concerns around privacy and
identity
INDIVIDUAL CASE
BASIC ADVANCED
Tracking and visualization of gross
progress or milestones against time
(e.g., by day, week, month)
MEASURE-MENT
Tracking progress and challenges
relative to consumer behavior and
engagement patterns across all
devices and services
Focus on identifying trends at
population level and applying
learnings top down, demonstrating
quick success for majority
APPROACH Analyzing when and how individual
consumers respond to incentives to
allow for personalized notifications,
or “interventions”
Engagement and subsequent
effectiveness weans and new
population interventions are
deployed
EFFECT Results are sustainable as
interventions continuously adapt to
individuals, rolling up to significant
population change
30. Companies that are able to quickly improve algorithms through
closed loop models build significant long-term defensibility
CASE STUDY: POPULATION HEALTH EXAMPLE:
CHALLENGES
BASIC ADVANCED
Relevant data is accessible but split
across multiple entities
ACCESS All relevant financial, clinical, and
customer data is stored within a
single structure or warehouse
Predictive capability restricted by
dated relationships between
attributes and recommendations
TESTING Patients are randomized at point of
intervention to allow rapid testing of
population health interventions
Retrospective observational reviews
are conducted to assess effectiveness
of interventions
TIMING Performance is measured in near real-time
to link patient predictive
attributes to recommended
interventions to outcomes (e.g.,
engagement, health, financial)
RELATIONSHIP SEARCH DATA COLLECTION PERFORMANCE CAPTURE
PRESENTATION © 2014 ROCK HEALTH
Source: Company website, interviews 30
• Control of data collection along the
continuum from predictor data to
treatment/intervention to health outcome
• Ability to aggregate historical data and
patient information at point-of-care for
real-time performance measures
• Integration into clinical workflow for
intervention testing and performance
capture
• Health outcomes are inherently lagged,
limiting timely assessment of effectiveness
DATA AGGREGATION INDIVIDUAL CASE CONTEXTUALIZATION
32. “Healthcare providers don’t just want predictive
analytics to output graphs and statistics. They need
something that’s actionable.
You have to distill it down to what matters and is
actionable. Because there’s a hundred thousand
things that come into play in health care, predictive
analytics has to tell us what matters and how we can
act on it.”
ANIL JAIN
Chief Medical Officer
Explorys offers a
software platform
solution that helps
healthcare systems
aggregate, analyze, and
manage their big data
33. Personalizing care through predictive analytics represents a
significant opportunity to reduce costs in the healthcare system
$192B $128B $35B
OVERTREATMENT FAILURES OF CARE DELIVERY LACK OF CARE COORDINATION
PRESENTATION © 2014 ROCK HEALTH
• Eliminating care that cannot help
patients—care that is outmoded,
supply-driven, and eschews science
• Restricting treatment and intervention
to the patients who will benefit based
on the individual and the context
• Continuously studying care to identify
what works for whom and in what
context
• Scaling best practices including
preventive care and early warning
systems that demonstrate effectiveness
• Ensuring those at the highest risk of
costly medical episodes are identified,
monitored, and cared for between visits
and following hospitalization
Source: “Eliminating Waste in US Health Care” Journal of the American Medical Association (2012) 33
34. It will largely fall onto the healthcare industry to recognize the
value of predictive analytics and implement critical use cases
IDEALIZED USE CASE OVERTREATMENT CARE DELIVERY COORDINATION
PRESENTATION © 2014 ROCK HEALTH
34
PAYERS Construct personalized medical policy (what is and isn’t
covered) and benefits (how costs are shared by parties)
Match interventions to individuals to scale behavior change
programs (wellness, chronic disease management, etc.)
PROVIDERS Provide point of care access to historical data in the context
of a patient in ambiguous situations (“Green Button”)
Reduce treatment variation and improve outcomes
Manage risk of population health management programs
under accountable care
BIOPHARMA Predict individual responsiveness to treatment (within R&D
and post-market contexts)
Conduct pharmacovigilance
35. The industry might be waiting to implement predictive analytics
as the FDA decides how best to regulate clinical decision support
“
Any software that analyzes data and supports
clinical decision making, including:
• Computerized alerts, reminders and warnings
• Computer-aided diagnosis
• Treatment recommendations
Regulation will be agnostic to information source
(manual entry, automated, etc.)
Our question: How will the FDA regulate the practice of medicine when algorithms prove more accurate than clinicians?
PRESENTATION © 2014 ROCK HEALTH
This guidance does not address the
approach for software that performs
patient-specific analysis to aid or
support clinical decision-making.”
LIKELY SCOPE OF FUTURE GUIDANCE POTENTIAL FRAMEWORKS
Source: FDA.gov;
“FDA regulation of clinical decision support software” Journal of Law and the Biosciences (2014) 35
Bipartisan Policy Center (BPC) proposed CDS be
subject to a new oversight framework:
• Adherence to and implementation of designated
standards
• Participation in safety monitoring
• Aggregation and analysis of trends to mitigate
future risk
Food and Drug Administration Safety and
Innovation Act (FDASIA) working group advised:
• Different frameworks dependent on risk, with
low-risk categories exempt from pre-market
approvals/clearances
• Clarification amongst multiple agency regulation
(e.g., FDA/ONC/FCC)
36. Beyond regulation, the biggest risk to predictive analytics being
used in health care is adoption as power dynamics shift
2 1
PATHWAYS ADOPTION CHALLENGES
1 Software-based clinical decision support
Patient provides data to the doctor, who
incorporates it into a decision support
algorithm for diagnosis or treatment
2 Patient-controlled
Patient generates and submit their own
data into the predictive algorithm, allowing
them to directly receive clinical insights
Our question: Can user experience and design influence decision making so deeply as to be regulated?
PRESENTATION © 2014 ROCK HEALTH
36
0110001001
1010010110
1111011010
0101101110
0110011001
PREDICTIVE
ANALYTICS
HEALTHCARE
PROFESSIONAL
PATIENT
3
1
3 Traditional
Patient provides the clinician with the data
they need to diagnose and treat based on
their own judgment
• Loss of decision making power
• Direct integration into clinical workflow
• Transparency of complex algorithms
• Management of liability
• Convenience of accessing algorithms
• Accuracy and reliability of recommendations
• Management of privacy concerns
• Regulatory burden
37. We are underestimating the potential impact of
predictive analytics in process tools to help physicians
make better decisions.
Every week, at the airport, I get on an airplane, and I
don’t worry about flying at all. There are so many tools
deployed to assist the pilot. I was talking with a pilot
about the new 787–and the pilot said he basically
monitors the plane. We’re going to see more of that in
health care.
Physicians will be monitoring algorithms.”
KEVIN FICKENSCHER
President, AMC Health
Former President, AMIA
AMC Health provides
customized, scalable
telehealth solutions for
organizations serving at-risk
populations through
remote patient
monitoring programs.
“
38. ACKNOWLEDGEMENTS
We are indebted to our industry partners who not only support
our work every day but provided invaluable feedback on an
early draft of this report.
A number of industry, startup and venture folks also offered
their expertise. Special thanks to Karina Babock, Benjamin
Berk, Archit Bhise, Joe Boyce, Matt Butner, Chris Coloian, David
Crockett, Ash Damle, Asif Dhar, Bill Evans, Kevin Fickenscher,
Luca Foschini, Ryan Goldman, Josh Gray, Sam Ho, Lucian
Iancovici, Anil Jain, Donald Jones, Allen Kramer, Uri Laserson,
Christine Lemke, Dave Levin, Dan Martich, Phil Okala, Trishan
Panch, Vinnie Ramesh, Leah Sparks, David Tamburri, Euan
Thomson, Abhimanyu Verma, Nate Weiner, and Jack Young for
their time and insights.
Finally, we are fortunate to work with the most encouraging and
passionate team in digital health. We are certain that no one
would even be reading this report if not for the tireless
marketing efforts of Halle Tecco and Mollie McDowell.
research@rockhealth.org
@rock_health
PRESENTATION © 2014 ROCK HEALTH