Medical Design & Manufacturing Executive Summit 2014

Medical Design & Manufacturing Executive Summit 2014

Evren Eryurek Software Chief Technology Officer, GE Healthcare February 11, 2014

© General Electric Company, 2013. All Rights Reserved.


Now we get Air Speed Calibrated Altitude Carrier Name Core Compartment Cooling Valve Position Core Speed Date GE Received Departure Station Destination Station EGT Exhaust Gas Temperature EGT1 EGT2 EGT3 EGT4 EGT5 EGT6 EGT7 EGT8 Engine Serial Number FMV Position FMV Position Demand Fuel Filter Delta Pressure Fuel Flow Ground Speed Mach Number Message Type N1 Fan Speed without modifier N1 Indicated Fan Speed N1 Maximum Fan Speed N1 Modifier Oil Filter Delta Pressure Oil Pressure PS3 Compressor Discharge Pressure T25 Compressor Inlet Temperature T3 Compressor Discharge Temperature Time GE Received Total Air Temperature Aircraft Total Air Temperature Engine VBV Position

AVSV Position Aircraft ID Aircraft Pitch Angle Aircraft Roll Angle Control System Status Word 1 Control System Status Word 2 DMS Cumulative Chip Count DMS Per Flight Chip Count EMU Status Word 1 ENGINE_RATING Engine Bump Engine Position FADEC AS Software Version Id FADEC Hardware P/N FSV Demand FSV Main Demand FSV Position Fuel Boost Strainer Delta Pressure Fuel Density Center Tank Fuel Flow Demand Fuel Manifold Pressure Fuel Manifold Temperature HPTACC Demand HPTACC Position LPTACC Demand LPTACC Position MSV Demand MSV Position Minor Airframe Model Word Oil Quantity Aircraft Oil Supply Temperature P0 Ambient Pressure PS25 Compressor Inlet Pressure PT2 Inlet Total Pressure Report Code T12 Inlet temperature TBV Demand Aircraft Corrected Angle of Attack

Temperature Static Air APU Bleed Isolation Valve Close Switch APU Bleed Isolation TBV Position TCORE Nacelle Temperature TOIL Oil Scavenge Temperature Throttle Resolver Angle VBV Demand VSV Demand Vib Broadband #1 Bearing - Engr Units Vib Broadband TF - Engr Units Vibration N1 #1 Bearing - Engr Units Vibration N1 Turbine Frame - Engr Units Vibration N2 #1 Bearing - Engr Units Vibration N2 Turbine Frame - Engr Units A/I valve position Wing Valve Open Switch AVM BB vibration Cockpit units AVM Fan vibration Cockpit units AVM High Wins Select Vibration - SS AVM LPT vibration Cockpit units AVM N2 vibrations Cockpit units Air Speed Calibrated - SS Altitude - SS Bleed Flow rate Control System Status Word 4 Control System Status Word 6 Copy Number Core Speed - SS EAI Switch Position - SS EAI Switch Position Open/Closed ECS Pack 1 Flow ECS Pack 1 High/Low ECS Pack 2 Flow ECS Pack 2 High/Low ECS Pack 3 Flow ECS Pack 3 High/Low EEC Subsystem Status Word 2 EGT Exhaust Gas Temp Unshunted - SS


OT is virtualized

Analytics become predictive

Machines are self healing & automated

Monitoring and maintenance is

mobilized

Employees increase productivity

Entertainment is Digitized

Social Marketing Emerged

Communications

Mobilized

IT Architecture Virtualized

Retail & Ad Transformed

2. Rise of the Machines

WHAT HAPPENED WHEN

1B PEOPLE BECAME CONNECTED?

WHAT HAPPENS WHEN

50B Machines BECOME CONNECTED?

CONSUMER INTERNET INDUSTRIAL INTERNET


3. Real-Time Analytics


4. Changing Architecture

Distributed Compute & Analytics

Cloud for Efficiency & Agility

Mobile: Anytime,

Anywhere Access

Predictive Insights from

Big Data

Transition to Brilliant Machines

Consistent & Meaningful Experience

Industrial Software Application Requirements

End-to-End Security


The Power of 1…

Increasing Freight Utilization

Rail

Predictive Maintenance

Healthcare

Predictive Diagnostics

Power

$27B industry value

by reducing system inefficiency

$63B industry value

by reducing process inefficiency

$66B industry value

with efficiency improvements in gas-fired

power plant fleets

Outcomes that matter

Power of 1

Limits re-use

• Wasted resources

• No cross-pollination among developers and SMEs

• Varying look and feel

• Solution overlap

Takes too long

• Multiplier effect: 30-50% of time spent on repetitive tasks

• Difficult to add new features, upgrades

• Slow time to market = missed opportunity

Costs too much

• Repetitive tasks and lack of reuse increases CAPEX and OPEX

• No economies of scale

• Opportunity cost

Increases risk

• Many additional sources of error

• Higher security risk

• Greater risk of cost and time overruns

• Poor decision support

Challenges of Fragmented Development


Software & Analytics at GE

Page 11

GE Predix™ Software Platform

GE Aviation

GE Energy Management

GE Healthcare

GE Oil & Gas

GE Power & Water

GE Transportation

Op

era

tio

ns

o

pti

miz

ati

on

Ass

et

op

tim

iza

tio

n

Op

era

tio

ns

o

pti

miz

ati

on

Ass

et

op

tim

iza

tio

n

Op

era

tio

ns

o

pti

miz

ati

on

Ass

et

op

tim

iza

tio

n

Op

era

tio

ns

o

pti

miz

ati

on

Ass

et

op

tim

iza

tio

n

Op

era

tio

ns

o

pti

miz

ati

on

Ass

et

op

tim

iza

tio

n

Op

era

tio

ns

o

pti

miz

ati

on

Ass

et

op

tim

iza

tio

n

Pla

tfo

rm

1

Pla

tfo

rm

2

…

Pla

tfo

rm

a

Pla

tfo

rm

b

…

Pla

tfo

rm

i P

latf

or

m ii

…

Pla

tfo

rm

x

Pla

tfo

rm

y

…

Pla

tfo

rm

A

Pla

tfo

rm

B

…

Pla

tfo

rm

. P

latf

or

m ..

…

Predix GE Industrial Internet Capabilities

PredictivityTM Solutions Operations optimization Asset optimization

Big Data & Analytics @ Healthcare


GE Healthcare

Global business unit of

GE $18B

Number of employees

worldwide 53,000

Core Strengths in bio-sciences,

diagnostics technologies, operations

$1B+ Annual investment in R&D

14 GE Title or job number

2/7/2014

Clinical Quality Reporting Use Cases

Thousands of physicians benchmarking millions of patients in a secure database.

Providing clinical insight in a proactive manner.

Analytics lead physician directly to action.

High Value Questions:

- Chief Medical Officer: How are my clinicians performing

vs. hospitals like mine?

- Clinician: How am I performing for my patients?

- Clinician: Which patients need help now?


2/7/2014

Single Cell Analysis for Cancer Diagnosis Goal: Develop timely tests using biomarker analytics to diagnose cancer cells.

Output: Automated per cell quantitative biomarker to correlate and classify tumor type and outcomes.

Scale: We will analyze 10,000 cells per slide, and each cohort will be 500 slides. Per cohort, this means about ~5TB of raw images and ~50GB of structured data.

High value questions:

Does this sample contain a diagnosis of cancer?

What are the types of tumors we can see?

How does the type of tumor and associated cells impact outcomes?

Nuclei Cytoplasm Membrane

Segmented

ERK Her4 p4EBP1

Trisomy Risk Calculation

Platform: a recipe for speed

Updated ultrasound algorithm

Created new analytics UI

Deployed app to cloud

Sent new data to cloud

Delivered solution….

….in one week .


2/7/2014

Frictionless Patient Experience…

….for a single hospital …for an entire region

Which hospital can treat this patient? Which clinic is the patient in? Are we having an epidemic?

Dynamic patient placement …

Which ward should treat this patient? How long have my patients been waiting? Exactly when will a patient be discharged?


2/7/2014

Collaboration for Radiologists in the Cloud

High Value Questions:

Which Radiologist can help me read this image right now?

Do I have a business agreement with this radiologist?

How long do I want this image in the cloud before it expires?

Cloud-based clinical collaboration services & medical imaging applications

Zero footprint image viewing

Collaboration among radiologists for patient care


2/7/2014

Common Concerns about Cloud…

Will data go outside my geographic region?

Where is my data?

Failures and unavailability

of access to data How can I ensure my data is secure?

Chain involving multiple processors and subcontractors

Can I permanently delete my data? Will my data be kept separate from

other data?

Who has access to my data?

Thank You


Health & Medicine

Medical Design & Manufacturing Executive Summit 2014