The Role of Data in Achieving Precision and Value in Healthcare - - PowerPoint PPT Presentation

THE MALONE CENTER FOR ENGINEERING IN HEALTHCARE

The Role of Data in Achieving Precision and Value in Healthcare

Gregory D. Hager Mandell Bellmore Professor Director

M A L O N E C E N T E R . J H U . E D U

NAE / IOM (2006) IOM (2011) PCAST (2014)

Advancements in medicine and technology are

• nly as effective as the associated care delivery

The Malone Center Mission: Transform the Process of Healthcare Delivery By Translating Research-Based Innovations into Engineered Systems

M A L O N E C E N T E R . J H U . E D U

Health Care: Some Numbers

• Total Expenditures: 3T (18% of GDP)
• System inefficiencies estimated to be up to 1/3 cost
• Medical errors estimated to be 3rd leading cause of death in the US

Makary & Daniel. "Medical error-the third leading cause of death in the US." BMJ: British Medical Journal, 2016.

Patient workflow complexity

from Basole et al. J Am Med Inform Assoc 2015..

Decision complexity

from Engineering a Learning Healthcare System, NAM

Value = Quality/Cost

M A L O N E C E N T E R . J H U . E D U

Opportunity: A New Wealth of Data

• 5,320,357 patients in Epic
• 35,986,859 notes written in Epic in 2017
• 12,914 average patient encounters

documented in Epic per day

• 5,000+ physicians in Epic
• 2,783,734,072 DICOM objects,

12,140,267 studies for 2,421,774 patients From 2008 to 2014, hospitals with EHRs rose to 75% from 9%, and in doctors’

• ffices rose to 51% from 17%.

2300 exabytes of healthcare data will be produced in 2020 (153 in 2013)

M A L O N E C E N T E R . J H U . E D U

Data Science Opportunities at Multiple Levels

Ferlie and Shortell, 2001

Monitoring/Modeling Clinical decision making Hospital operations Public policy 4-Level Health Care System Data Science Opportunities toward better:

• health outcomes
• care value = quality / cost

Slide courtesy Scott Levin

M A L O N E C E N T E R . J H U . E D U

Opportunity Spaces

• Diagnosis
• Better decision-making
• Early warning systems
• Disentangling multiple causal factors
• Treatment
• Choosing the most effective care
• Monitoring quality of care
• Improving training and workflow
• Recovery
• Better monitoring
• Better analytics
• Technology support beyond CCE

Kata Project, Ahmad, JHU

E-Triage, Levin, JHU Schleroderma, Saria, JHU

M A L O N E C E N T E R . J H U . E D U

A Challenge: Failed Surgery

0.25M sinus surgeries per year $22B per year expenditure on chronic rhinosinusitis 25% of surgeries for nasal airway obstruction “failed” What causes nasal obstruction? When will surgery help? Hypothesis: Large population data sets, correlated with

• utcomes can provide clues to who benefits and why

M A L O N E C E N T E R . J H U . E D U

A Challenge: Failed Surgery

0.5M spinal fusion surgeries / yr $12B / year (70% increase 2001 – 2011) 7.5% compound annual growth by 2019 High Range in Variability (Quality) 53% of patients have comorbidity 8-25% of patients rehospitalized High variability in surgical outcomes à Opportunity for a more information-driven approach

Courtesy Jeff Siewerdsen (JHU BME)

M A L O N E C E N T E R . J H U . E D U

An Opportunity for “Image Clouds”

Courtesy Jeff Siewerdsen (BME)

M A L O N E C E N T E R . J H U . E D U

Statistical shape models

12

A Sinha, et al., Automatic segmentation and statistical shape modeling of the paranasal sinuses to estimate natural variations, SPIE Medical Imaging, 2016 A Sinha, et al., Simultaneous segmentation and correspondence improvement using statistical modes, SPIE Medical Imaging, 2017

⋯

"# = v## v#& ⋮ v#() "& = v&# v&& ⋮ v&() "(* = v(*# v(*& ⋮ v(*()

⋯

Deformable registration à population shape change à statistical model

M A L O N E C E N T E R . J H U . E D U

Statistical shape models

Variance along the principal mode for the maxillary sinus

Front view Left view

Variance along the principal mode for the middle turbinates

M A L O N E C E N T E R . J H U . E D U

Studying Effect of Anatomy on Outcomes

What causes nasal obstruction? When will surgery help?

⋯

s1 s2 sn

Shape Space

M A L O N E C E N T E R . J H U . E D U

Studying Effect of Anatomy on Outcomes

What causes nasal obstruction? When will surgery help?

⋯

s1 s2 sn s1 s2

Shape Space

M A L O N E C E N T E R . J H U . E D U

Studying Effect of Anatomy on Outcomes

What causes nasal obstruction? When will surgery help?

⋯

s1 s2 sn s1 s2

Shape Space

M A L O N E C E N T E R . J H U . E D U

Studying Effect of Anatomy on Outcomes

What causes nasal obstruction? When will surgery help?

⋯

s1 s2 sn s1 s2

Shape Space

M A L O N E C E N T E R . J H U . E D U

Data-Driven Anatomic Models From Endoscopy

Structure from Motion + Deep Learning

3D surfaces and normals

Deformable Registration Statistical Model Patient Model

With R.H. Taylor, A. Sinha, A Reiter, M Ishii

Move from qualitative assessment to quantitative measurement of anatomy of every patient

M A L O N E C E N T E R . J H U . E D U

Birkmeyer J.D, et al. Surgical Skill and Complication Rates after Bariatric Surgery. NEJM, 2013.

0.05% 5.20% 1.60% 2.70% 0.26% 14.50% 3.40% 6.30% Mortality Complication Reoperation Readmission Score Bottom Quartile Score Top Quartile

1 2 3 4 5 Expertise Score

Michigan Bariatric Surgery Collaborative Samples: 20 bariatric “expert” surgeons ranked by at least 10 reviewers. 10,343 patients admitted 2006-2012

The Human Element: Skill vs. Outcomes

5x Mortality Rate!

M A L O N E C E N T E R . J H U . E D U

Max over filter activations: Index of max filter:

Layer 1 Layer 2 Layer 3

Temporal Convolution Networks (TCN)

DiPietro, Robert, et al. "Recognizing surgical activities with recurrent neural networks." MICCAI, 2016. Lea, Colin, et al. "Temporal Convolutional Networks for Action Segmentation and Detection." CVPR. 2017.

M A L O N E C E N T E R . J H U . E D U

A New Lens on The Human Element

Novice Expert Capturing and structuring surgical performance data and relating it to quality of outcome. Septoplasty

M A L O N E C E N T E R . J H U . E D U

First Steps Towards an AI-Assisted OR

− − − − − − − − − − − − − − − − −8 −7 −6 −5 −4 −3 −2 −9 −8 −7 −6 −5 −4 −3 −2 −1 1

Similarity to Novice Similarity to Expert

Classifying skill in the laboratory

Less experienced operator More experienced operator Linear classifier

−5 −4.9 −4.8 −4.7 −4.6 −4.5 −4.4 −4.3 −4.2 −3.6 −3.4 −3.2 −3 −2.8 −2.6 −2.4

Similarity to Novice Similarity to Expert

Classifying skill in the operating room

− − − − − − − − − − − − − − − −

Resident Attending Linear classifier

Trained in the lab Tested in the OR

M A L O N E C E N T E R . J H U . E D U

Following Treatment and Recovery

A Plug and Play platform for quantifying clinical activities

Detecting People Identifying People

Caregiver Patient

Estimating Pose

Standing Sitting

Identifying Objects

Bed [upright ] Table

Motion Analysis Spatio-Temporal Analysis

Ma, Rawat, Reiter et al. Crit. Care Med., 45:4, 2017

M A L O N E C E N T E R . J H U . E D U

• Demonstration: Assessing patient mobility

Spatio-Temporal Analysis Detecting People Estimating Pose Standing Sitting Identifying People Caregiver Patient Motion Analysis

Following Treatment and Recovery

Ma, Rawat, Reiter et al. Crit. Care Med., 45:4, 2017

M A L O N E C E N T E R . J H U . E D U

Early Results

§ Weighted Kappa: 0.86 (95% Confidence Interval: 0.72, 1.00) § Sensor and clinician agreed on 72 out of 83 segments (87%) § Of the 11 discrepancies, 7 were due to confusion between ‘nothing in bed’ and ’in-bed activity’.

Nothing In-bed Out-of-bed Walking Nothing 18 (22%) 4 (5%) In-bed 3 (4%) 25 (30%) 2 (2%) Out-of-bed 1 (1%) 25 (30%) 1 (1%) Walking 4 (5%) Total 21 (26%) 30 (36%) 27 (32%) 5 (6%)

1200+ hours of data collected in the Johns Hopkins Weinberg ICU

Physician System

Ma, Rawat, Reiter et al. Crit. Care Med., 45:4, 2017

M A L O N E C E N T E R . J H U . E D U

Summary

• Healthcare is being transformed by data
• We are still in the early days of learning how to

effectively collect and use data to improve value

• Data science challenges are large and growing: bias,

missing data, heterogeneity and robustness, transparency, trust, and privacy

• Challenges remain at the first (data acquisition) and

last (deployment) mile.

What you cannot measure, you cannot improve – Lord Kelvin