Transcript PowerPoint slides - C-MORE

Activity Monitoring and Outcome Measurements
by Remotely Sensing Daily Mobility and Exercise
in Disabled Persons
Bruce Dobkin, MD, FRCP
Professor of Neurology
Director, Neuro-Rehabilitation Program
Geffen/UCLA School of Medicine
UCLA Wireless Health Institute
mHealth
Definition: Delivery of healthcare services via mobile
communication devices.
Opportunity: By 2017, more mobile phones than
people on the planet; currently three-quarters of the
world’s population have access to a mobile phone.
Goal: Facilitate medical and health information via
instantaneous communication anywhere/anytime.
Reduce disparities, prevent disease, improve
diagnostics & therapy, increase adherence, personalize
medical advice in chronic diseases, enhance research &
daily care, merge diverse data sets, lower costs.
UCLA WHI Strategies
• Develop reliable, highly valued tools to improve outcomes in
healthcare.
• Deploy the intellectual resources of the UCLA medical,
nursing, engineering, public health and other schools and
depts.
• Encourage faculty and students to identify needs, ideas and
pilot studies.
• Design and Test: Use an iterative approach to identify
opportunities, develop and test user-friendly tools, and show
the efficacy of sensors (or other WHI devices) within
scientifically conducted clinical trials on likely users.
• Serve as an incubator for wireless health ventures.
Lots of activity sensors, but where is the beef?
Fitbit Flex
Basis Band
BodyMedia Core2
Armband & Vue Patch
Withings Smart
Activity Tracker
PerformTek
Fitbug Orb
Fitlinxx Pebble
MDAWN
Medical Daily Activity Wireless Network
An inexpensive, wireless sensor/algorithm system
that can remotely recognize and quantify
purposeful behaviors such as walking, exercise,
and skills practice, as well as provide feedback
about performance, in persons with impaired
mobility.
Our niche: within clinical trials across diseases,
improve outcomes for disabled persons
Monitor skills practice, exercise, & mobility activities
in the home and community for compliance
and safety, and to audit clinical trial interventions.
Develop outcome measures with continuous,
rather than ordinal scales; quantify the type,
intensity, and quality of mobility.
Capture gains and declines in purposeful activities
in the real world, not just in the unnatural
environment of a lab and not with the ambiguity
of self-report scales.
Remote Sensing Systems
Accelerometer
Magnetometer Sensor
RFID
Contact,
EMG,
Goniometer
Hidden Markov
Model
Naïve
Bayes
Neural
Classifiers
Nearest
Fusion
Networks
Neighbor
Gyroscope
Decision Tree
Ambient Sound & Visuals
Singular Spectrum
Analysis
Gait Analysis;
Mean of Signal
Athletic Training
Falls
Correlation
Type, Quantity,
of Axis
Application Quality of Activity
Features
NeuroPeak
Rehabilitation
Frequency
Daily Activity
Std. Dev.
Monitoring
&
Energy of Signal
Motor Control
Feedback
Ankle accelerometers can describe walking,
cycling, exercises, and overall activity in the home
and community at low cost.
Dobkin & Dorsch. Neurorehabil Neural Repair, 2011
MDAWN for Disabled Persons
• Based upon two 10-meter walks, machinelearning algorithms enable a template for each
participant that identifies subsequent episodes
of walking or exercise throughout the day.
• Gait parameters include walking speed,
duration, distance, and limb asymmetries, which
are calculated for each walking episode.
77 year-old with chronic left hemiplegic stroke
Walking speed is 0.1m/s
Stroke Inpatient Rehabilitation Reinforcement of
ACTivity (SIRRACT)
• Can clinicians improve
walking-related
outcomes during
hospital-based rehab?
• International RCT.
• Wear ankle sensors.
• Compared 2 levels of
daily feedback about
performance.
Presented at AAN, 3/13
• 140 subjects at 15 sites.
• Showed increasing
amount of walking and
walking speed from
admission to discharge.
• Rather low mean daily
amount of training was
detected.
• Proved ease of use,
accuracy, relevance of the
data.
Daily # steps
Daily distance walked
Average walking speed
16
0.9
14
0.8
0.7
12
0.6
10
0.5
8
0.4
6
0.3
4
0.2
2
0.1
0
0
11 12 15 16 17 18 19 22 23 24 25 26 29 30 31 32 33 36 37 38 39
Days since stroke
Average walking speed (m/s)
Total time walking (min)
SIRRACT participant in Taiwan
during inpatient stroke rehabilitation
time
speed
Instrumented Devices:
UCFit low cost system for bed exercise UCFit
• Home or hospital
– Android smartphone with apps
– Portable, battery-powered,
weighs <7lbs.
Smartphone
– Strain gauge & MicroLEAP
App
sensor platform.
• Data acquisition automatic
Bluetooth
Wireless
Internet
– UCFit Server’s secure systems
acquire, archive, present data,
and provide feedback graphics
User Group’s Database MDAWN Server
UCFit light resistance cycling for disabled persons
Average Power Output
6.0
2.5
5.0
2.0
4.0
b
b
25
-Fe
26
-Fe
b
28
-Fe
b
b
27
-Fe
24
-Fe
b
26
-Fe
0.0
b
0.0
25
-Fe
1.0
b
0.5
Date
Date
Total Time Spent Pedaling
Average Cadence (RPMs)
30.0
40.0
25.0
30.0
RPM
20.0
15.0
20.0
10.0
10.0
5.0
0.0
Date
b
b
b
b
b
24
-Fe
25
-Fe
26
-Fe
27
-Fe
28
-Fe
28
-Fe
b
27
-Fe
b
26
-Fe
b
0.0
25
-Fe
b
24
-Fe
b
Time (minutes)
b
2.0
28
-Fe
1.0
3.0
b
1.5
27
-Fe
Power
3.0
24
-Fe
Distance (miles)
Distance Traveled (27" wheel)
Date
35
3.5
30
3.4
25
3.3
20
3.2
15
3.1
10
3
5
2.9
0
2.8
11
12
13
18
19
21
22
24
POD
25
26
27
32
33
34
35
Pedaling torque (Nm)
Pedaling time (min)
UCFit daily time/torque for post-op liver transplant
patient in ICU: physiological data and insight for care
time
torque
Sensors for daily medical care
• Monitor hourly or day to day fluctuations in responses to
medications, as for Parkinson’s, epilepsy, spasms, dyskinesias.
• Monitor compliance with activity-related instructions for
practice or exercise to reduce risk factors and improve function.
•Monitor for changes in activities that may reflect a decline in
functioning, due to disease exacerbation, new complications, side
effects of drugs, mood disorders.
•Provide feedback about performance to progressively improve
specific outcomes.
•Feedback and monitoring to motivate goal-setting and
compliance.
•Establish new types of measurable activity-related outcomes
and goals.
•Reduce number of visits, and cost, for care of chronic disability.
Sensors for clinical research trials
• Develop ecologically sound outcome measures of activity to
augment questionnaires and ordinal scales of disability and physical
functioning.
• Obtain continuous measures of daily activities – type, quantity,
quality. Also enables trialists to phase in an intervention so that a
baseline behavioral plateau is assured.
• Reduce the cost and increase the validity of clinical trials by being
able to remotely assess what is practiced, how much, and how well,
during a trial.
• If subjects at multiple sites or at home are being trained in a skill,
such as walking or using an affected arm and hand, monitor the
integrity of the intervention.
• Observe the effects of adverse events, such as drugs, pain or falls,
on activity.
• Increase the number of interim outcome measurements to better
develop dose-response curves.
Type, quantity & quality of activity in relation to
physiologic variables, images, social interaction,
environmental toxins, cues & feedback
Alzheimer’s
Asthma
COPD
Cancer
Depression
Diabetes
CHF
Hypertension
Obesity
Sleep
Vital signs, location, balance
RR, FEV1, oximetry, air quality, pollen
“ “
“
Adverse effects of meds & disease
Drug compliance, communication
Glucose, HgbA1c, drug use, exercise
Pulm artery pressure, weight, VS, fluid
Continuous BP, drug compliance
Smart scales, calories in/out, behavior
Sleep stages, quality, apnea
High throughput, multi-streamed, longitudinal data sets
to facilitate disease prevention, management and
behavioral changes.
Requirements for mHealth data
• 1. Collect data from technologies along with self-reported
behavioral, psychosocial, environmental, and contextual
measures. Analytics for BIG DATA.
• 2. Integrate various wireless physiologic and bio sensors
on open platforms.
• 3. Appropriately secure data at each stage of collection,
transfer, and storage.
• 4. Visualize data using customizable tools.
• 5. Analyze and report on individual or group level data
using customizable tools and reporting systems.
• 6. Maintain compliance with HIPAA, IRB and FDA.
• 7. Demonstrate efficacy and effectiveness of realworld data.
Address patient-centered outcomes research:
NIH and Medicare priority
• “Given my personal characteristics, conditions, and
preferences,…..
• “What should I expect will happen to me?”
• “What are my options, and what are the benefits and
harms of those options?”
• “What can I do to improve the outcomes that are most
important to me?”
• “How can the health care system improve my chances
of achieving the outcomes that I prefer?”
Personal activity logging and feedback
Washington, NEJM, 2011
UCLA Wireless Health Institute
www.wirelesshealth.ucla.edu
• Bill Kaiser, Majid Serrafzadeh, Deborah Estrin, Greg
Pottie, Chris Cooper
UCLA Medical and Engineering Campus
www.Wirelesshealth.ucla.edu
William Kaiser, Greg Pottie, Andrew Dorsch, Seth Thomas,
Celia Xu, Lam Yeung, Eric Yeun, James Xu, Yan Wang,