Data @ Dexcom: CGM technology, IoT connectivity, and the...
Transcript of Data @ Dexcom: CGM technology, IoT connectivity, and the...
Data @ Dexcom:
CGM technology, IoT connectivity, and
the drive to insight delivery in diabetes care
Nate Heintzman, PhD
Sr. Manager, Data Partnerships
20161130
TYPE 220251980 2014
Diabetes is a global crisis
PREVALENCE
OF
DIABETES
0
700
400
MIL
LIO
NS
108MADULTS
GLOBALLY
422MADULTS
GLOBALLY
1 IN 3
ADULTSIN US WILL HAVE
DIABETES BY
2050
Source: NCD Risk Factor Collaboration, Published April 2016; CDC
WORLDWIDE DIRECT
ANNUAL COST
Source: International Diabetes Federation Diabetes Atlas 7th Edition, 2015
Germany
$35BILLION
China
$51BILLION
United
States
$320BILLION
Japan
$29BILLION
$673B - $1.2T
Diabetes is a global crisis
Are her lifestyle
changes having an
impact?
Is her $400/month
medication working
for her?
Will she need to start
taking insulin
injections soon?
Is it safe for him
to go to soccer
practice?
Should he take a
dose of insulin now?
How much?
How will his glucose
change while he is
sleeping?
Patients and their care teams live with
uncertainty about their diabetes, every day
Continuous Glucose Monitoring (CGM)We now have a whole new understanding of glucose levels
BEFORE CGM WITH CGM
Continuous
flow of data
Holistic
understanding
of time in range
Better decision-
making to increase
time in range
Patient can share
data with family
members and
clinicians
2:15 PM 2:45 PM
3:45 PM
Dexcom CGMDelivering a key vital sign to the smartphone and beyond
OCTOBER 2014
First remote mobile
device for CGM that
can transmit via app
to five “Followers”
APRIL 2015
First apps to
enable CGM
on Apple Watch
AUGUST 2015
First fully mobile
CGM that sends data
directly to smartphone
without receiver
AUGUST 2015
Software that delivers
personalized, easy to
understand analysis of
glucose trends
The status quo was to isolate the data
Care Team
Family
Medical Record
ResearchINACCESSIBLE
DATA ISOLATED
TO DEVICE
DATA ISOLATED
TO SPECIFIC
COMPUTER
Payor
INTERNET
of
THINGS
CLOUD
COMPUTINGBIG
DATA
DATA
SCIENCE
MACHINE
LEARNING
MICRO
ELECTRONICS
PLATFORM
Dexcom has invested in a different approach
Securing to
industry/regulatory
standards
Protecting patient’s
right to determine how
their data is used
Architecting with
scalability, resilience, and
multi-tenancy to support
extended ecosystem
Security Privacy AdaptabilityEasing integration by
building with industry
standards
Integration
Regulatory / Quality Compliance
Dexcom platform core requirements
Adopting best-in-class big data technologiesSolution Overview Qualities
HadoopEcosystem of tools that enable distributed processing and analytics of large disparate data sets across multiple systems
• Highly redundant• Fault tolerant • Large storage capacity (handles large datasets)• Scalable
Cassandra Distributed data store optimized for availability and high write throughput. Commonly used to back end API and store/process time series data
• Highly redundant• Fault tolerant • Low latency • Linearly Scalable
SqoopEnables integration between Hadoop and relational data sources
• Enables new technologies to leverage pre-existing investments in relational data systems
KafkaDistributed, partitioned, replicated commit log service that provides queuing system functionality.
• Highly redundant• Fault tolerant • ‘Wicked fast’• Linearly Scalable
SparkFast and general-purpose cluster computing system providing high-level APIs and also supports a rich set of higher-level tools.
• Can use it interactively with Scala, Python, and R• Can combine the power of SQL, Machine Learning,
and Streaming
Python/RLanguages used by data scientists to build and execute models and calculations
These tools can be used in conjunction with Hadoop,Cassandra, Spark, or standalone to build data products
Data Platform
What can the Dexcom platform do for you?
APIs
Data Platform
What can the Dexcom platform do for you?
APIs
Data Platform
What can the Dexcom platform do for you?
Standard API endpoints:EGVs (unit, rateUnit, egvs, systemTime, displayTime, value, status, trend, trendRate)devices (model, lastUploadDate, alertSettings, alertName, value, unit, snooze, delay, enabled, systemTime, displayTime)calibrations (systemTime, displayTime, value, unit)events (systemTime, displayTime, eventType, eventSubType, value, unit)statistics (estimatedA1C, hypoglycemiaRisk, min, max, mean, median, variance, stdDev, sum, q1, q2, q3, utilizationPercentage, meanDailyCalibrations, nDays, nValues, nBelowRange, nWithinRange, nAboveRange, percentageBelowRange, percentageWithinRange, percentageAboveRange)
APIs
Data Platform
What can the Dexcom platform do for you?
Standard API endpoints:EGVs (unit, rateUnit, egvs, systemTime, displayTime, value, status, trend, trendRate)devices (model, lastUploadDate, alertSettings, alertName, value, unit, snooze, delay, enabled, systemTime, displayTime)calibrations (systemTime, displayTime, value, unit)events (systemTime, displayTime, eventType, eventSubType, value, unit)statistics (estimatedA1C, hypoglycemiaRisk, min, max, mean, median, variance, stdDev, sum, q1, q2, q3, utilizationPercentage, meanDailyCalibrations, nDays, nValues, nBelowRange, nWithinRange, nAboveRange, percentageBelowRange, percentageWithinRange, percentageAboveRange)
3rd-party
app
3rd-party
software
APIs
Data Platform
What can the Dexcom platform do for you?
Standard API endpoints:EGVs (unit, rateUnit, egvs, systemTime, displayTime, value, status, trend, trendRate)devices (model, lastUploadDate, alertSettings, alertName, value, unit, snooze, delay, enabled, systemTime, displayTime)calibrations (systemTime, displayTime, value, unit)events (systemTime, displayTime, eventType, eventSubType, value, unit)statistics (estimatedA1C, hypoglycemiaRisk, min, max, mean, median, variance, stdDev, sum, q1, q2, q3, utilizationPercentage, meanDailyCalibrations, nDays, nValues, nBelowRange, nWithinRange, nAboveRange, percentageBelowRange, percentageWithinRange, percentageAboveRange)
3rd-party
app
3rd-party
software
APIs
Data Platform
What can the Dexcom platform do for you?
Standard API endpoints:EGVs (unit, rateUnit, egvs, systemTime, displayTime, value, status, trend, trendRate)devices (model, lastUploadDate, alertSettings, alertName, value, unit, snooze, delay, enabled, systemTime, displayTime)calibrations (systemTime, displayTime, value, unit)events (systemTime, displayTime, eventType, eventSubType, value, unit)statistics (estimatedA1C, hypoglycemiaRisk, min, max, mean, median, variance, stdDev, sum, q1, q2, q3, utilizationPercentage, meanDailyCalibrations, nDays, nValues, nBelowRange, nWithinRange, nAboveRange, percentageBelowRange, percentageWithinRange, percentageAboveRange)
3rd-party
app
3rd-party
software
Treatment
Data
Geolocation
DataRisk
Data
Cost
Data
Data
Science
Device
Integrations
Enabling integration, insight, and innovation…
Give high-risk patients
additional support
Alert family of danger
and locate patient
Evaluate treatment and
intervention efficacy
Determine impact of
glycemic control on
cost of care
Identify patterns to
inform new treatments
Closed-loop technology for
automated insulin delivery
…and empowering everyone involved to
improve outcomes and reduce costs
CLINICIANS
PATIENTS
POPULATIONS
PAYORS
Navigate daily self-
management decisions
with feedback loops for
personal understanding of
glucose response
Identify patients who need
more attention and develop
individualized precision
treatment plans
Deeply understand
efficacy of treatments to
target care more efficiently
and cost-effectively
Gain insights on which
approaches have driven
positive impacts for
different kinds of patients,
and why
Economics
of sharing
data
Regulation
Privacy &
Security
Cross
pollination
Industry
status quo
Which data
returns true
value?
Getting there is not simple
Driving together to the future of diabetes care
EHR
Payors
Data @ Dexcom:
CGM technology, IoT connectivity, and
the drive to insight delivery in diabetes care
Nate Heintzman, PhD
Sr. Manager, Data Partnerships
20161130
Basic Copyright Notice & Disclaimer
©2016 This presentation is copyright protected. All rights reserved. You may download or print out a hard copy for your private or internal use. You are not permitted to create any modifications or derivatives of this presentation without the prior written permission of the copyright owner.
This presentation is for information purposes only and contains non-binding indications. Any opinions or views expressed are of the author and do not necessarily represent those of Swiss Re. Swiss Re makes no warranties or representations as to the accuracy, comprehensiveness, timeliness or suitability of this presentation for a particular purpose. Anyone shall at its own risk interpret and employ this presentation without relying on it in isolation. In no event will Swiss Re be liable for any loss or damages of any kind, including any direct, indirect or consequential damages, arising out of or in connection with the use of this presentation.