75
The Science of User- Centered Data Tools June 2, 2014 Bradford W. Hesse, PhD Chief, Health Communication and Informatics National Cancer Institute Monday, June 2, 14

Making Data Usable: 2014 presentation at Datapalooza

Embed Size (px)

DESCRIPTION

Reviews the science of data display in the context of personal and public health data.

Citation preview

Page 1: Making Data Usable: 2014 presentation at Datapalooza

The Science of User-Centered Data Tools

June 2, 2014

Bradford W. Hesse, PhDChief, Health Communication and InformaticsNational Cancer InstituteMonday, June 2, 14

Page 2: Making Data Usable: 2014 presentation at Datapalooza

What behavioral intervention was credited with saving 58,000 injuries

per year, & saving $655,000,000 per year?

Question: American Psychological Association (2004)

Monday, June 2, 14

Page 3: Making Data Usable: 2014 presentation at Datapalooza

Question: American Psychological Association (2004)

Monday, June 2, 14

Page 4: Making Data Usable: 2014 presentation at Datapalooza

Why?

Human Factors research focused squarely on user’s perspective: Cognitive & perceptual demands

Intervention based on human-system integration

Monday, June 2, 14

Page 5: Making Data Usable: 2014 presentation at Datapalooza

4.3% in rear end collisions

92,000 crashes / year

58,000 injuries / year

$655,000,000 in property damage

Return on Investment(ROI)

Monday, June 2, 14

Page 6: Making Data Usable: 2014 presentation at Datapalooza

“Travel agents selected flight on first line

more than half the time”

American Airlines Sabre System

Data design effects

November Line of Sale Analysis, memo to R. E. Murray from S. D. Nason, American Airlines, Dec. 3, 1981.

Monday, June 2, 14

Page 7: Making Data Usable: 2014 presentation at Datapalooza

Website designs reshape travel & other industries

Sabre system becomes Travelocity

Data design effects

November Line of Sale Analysis, memo to R. E. Murray from S. D. Nason, American Airlines, Dec. 3, 1981.

CONSUMER ACCESS SHIFTS MARKET

Monday, June 2, 14

Page 8: Making Data Usable: 2014 presentation at Datapalooza

Interface wars part of new business

Microsoft DOS

Graphical User Interface

Monday, June 2, 14

Page 9: Making Data Usable: 2014 presentation at Datapalooza

Targets in Medicine?

Monday, June 2, 14

Page 10: Making Data Usable: 2014 presentation at Datapalooza

Nudging Best Practice:HITECH & Behavior

Source: Hesse, Bradford W., Ahern, David K., & Woods, Susan S. (2011). Nudging best practice: the HITECH act and behavioral medicine Translational Behavioral Medicine, 1(1), 175-181.

• Incentives• Understand Mental Maps• Defaults• Give feedback• Expect error• Structure decisions

Monday, June 2, 14

Page 11: Making Data Usable: 2014 presentation at Datapalooza

Stakes are high

Monday, June 2, 14

Page 12: Making Data Usable: 2014 presentation at Datapalooza

Stakes are high

Monday, June 2, 14

Page 13: Making Data Usable: 2014 presentation at Datapalooza

Data smog

• Decisional paralysis

• Confusion

• Risky behaviors

• Frequent errors

• Avoidance

Source: Shenk, David. (1997). Data smog : surviving the information glut (1st ed.). San Francisco, Calif.: Harper Edge.

Consequences

Monday, June 2, 14

Page 14: Making Data Usable: 2014 presentation at Datapalooza

The case of Hugo Campos

Source: Dave deBronkart, through ONC (http://www.healthit.gov/)Monday, June 2, 14

Page 15: Making Data Usable: 2014 presentation at Datapalooza

The case of Hugo Campos

Source: Dave deBronkart, through ONC (http://www.healthit.gov/)Monday, June 2, 14

Page 16: Making Data Usable: 2014 presentation at Datapalooza

The case of Hugo Campos

Source: Dave deBronkart, through ONC (http://www.healthit.gov/)Monday, June 2, 14

Page 17: Making Data Usable: 2014 presentation at Datapalooza

ROI will go to best design

for each user niche

Interfaces across ecosystem

Source: Hesse BW, Hansen D, Finholt T, Munson S, Kellogg W, Thomas JC. Social Participation in Health 2.0. IEEE Computer. 2010;43(11):45-52.Monday, June 2, 14

Page 18: Making Data Usable: 2014 presentation at Datapalooza

Human System Integration

Monday, June 2, 14

Page 19: Making Data Usable: 2014 presentation at Datapalooza

History: Bad Design Led to Catastrophic Error

Monday, June 2, 14

Page 20: Making Data Usable: 2014 presentation at Datapalooza

Mechanistic World View• Actors: Engineers, biological scientists• Question: How can we create new

technologies?• Focus: Physical Object

Humanistic World View• Actors: Social scientists, physicians• Question: How can we create new

people?• Focus: Person

Competing World Views

Monday, June 2, 14

Page 21: Making Data Usable: 2014 presentation at Datapalooza

Source: Vicente, Kim J. (2003). The human factor : revolutionizing the way people live with technology (1st ed.). New York: Taylor and Francis Books.

Human System Integration

Monday, June 2, 14

Page 22: Making Data Usable: 2014 presentation at Datapalooza

Knowledge in the Head*

Knowledge in The World*

Task Relevant Schemata

General model

Monday, June 2, 14

Page 23: Making Data Usable: 2014 presentation at Datapalooza

Computer Human Interaction

Don Norman, 1988

Jakob Nilsen: 1993, 1999

Shneiderman & Plaisant: 2010 edition

DHHS: 2004

Monday, June 2, 14

Page 24: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

See: www.usability.gov

Consistency within application

Consistency across product line

Predictable Controls

Monday, June 2, 14

Page 25: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability*

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

Sir Jonathan Ive

*universal across experience, literacy, physical ability, profession

Monday, June 2, 14

Page 26: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability

• Offer informative feedback

Eight Golden Rules

Shneiderman & Plaisant: 2010 editionSeriously?!

Monday, June 2, 14

Page 27: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability

• Offer informative feedback

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

Good design: sequence, progress, action all indicated

Monday, June 2, 14

Page 28: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability

• Offer informative feedback

• Design dialogs to yield closure

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

user goal

Deep Support

Monday, June 2, 14

Page 29: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability

• Offer informative feedback

• Design dialogs to yield closure

• Prevent errors

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

Monday, June 2, 14

Page 30: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability

• Offer informative feedback

• Design dialogs to yield closure

• Prevent errors

• Permit easy reversal of actions

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

Monday, June 2, 14

Page 31: Making Data Usable: 2014 presentation at Datapalooza

• Strive for consistency

• Cater to universal usability

• Offer informative feedback

• Design dialogs to yield closure

• Prevent errors

• Permit easy reversal of actions

• Support internal locus of control

Eight Golden Rules

Shneiderman & Plaisant: 2010 edition

Monday, June 2, 14

Page 32: Making Data Usable: 2014 presentation at Datapalooza

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

Source: Hesse BW, Shneiderman B. eHealth research from the user's perspective. Am J Prev Med 2007;32(5 Suppl):S97-103.

Asking the Right Question in H.I.T.

Monday, June 2, 14

Page 33: Making Data Usable: 2014 presentation at Datapalooza

STOP asking the wrong question

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

“Expert systems in medicine turned out to be brittle, impracticable, and nontransparent. In short, they turned out to be bad medicine.”

Source: Hesse BW, Shneiderman B. eHealth research from the user's perspective. Am J Prev Med 2007;32(5 Suppl):S97-103.

Monday, June 2, 14

Page 34: Making Data Usable: 2014 presentation at Datapalooza

START Answering Right Question

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

David  Brailer,  First  Na4onal  Coordinator  for  Health  IT  

“Everyone thought Health I.T. was about computers, but we’ve refined that to say that IT is

about healthcare — it’s about the experience we really have.”

Source: 1.   Brailer  D.  Ac/on  through  collabora/on:  a  conversa/on  with  David  Brailer.  The  na/onal  coordinator  of  HIT  believes  that  facilita/on,  not  mandates,  are  the  way  to  move  the  agenda  forward.  Interview  by  Robert  Cunningham.  Health  Aff  (Millwood)  2005;24(5):1150-­‐7.

Monday, June 2, 14

Page 35: Making Data Usable: 2014 presentation at Datapalooza

STOP asking the wrong question

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

“Intelligent  Agent”  misfires

Root  Cause:  Over-­‐Reliance  on  “Autopilot”  

Monday, June 2, 14

Page 36: Making Data Usable: 2014 presentation at Datapalooza

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

START Answering Right Question

Reminder  System

Source: Hesse  BW.  Enhancing  Consumer  Involvement  in  Health  Care.  In:  Parker  JC,  Thornson  E,  editors.  Health  Communica/on  in  the  New  Media  Landscape.  New  York,  NY:  Springer  Publishing  Company;  2008.  p.  119-­‐149.

Monday, June 2, 14

Page 37: Making Data Usable: 2014 presentation at Datapalooza

STOP asking the wrong question

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

*Zuboff S, Maxmin J. The support economy: why corporations are failing individuals and the next episode of capitalism. New York: Viking; 2002.

Monday, June 2, 14

Page 38: Making Data Usable: 2014 presentation at Datapalooza

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

START Answering Right Question

Source:  Hesse  BW.  Harnessing  the  power  of  an  intelligent  health  environment  in  cancer  control.  Stud  Health  Technol  Inform  2005;118:159-­‐76..Monday, June 2, 14

Page 39: Making Data Usable: 2014 presentation at Datapalooza

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

START Answering Right Question

*Zuboff S, Maxmin J. The support economy: why corporations are failing individuals and the next episode of capitalism. New York: Viking; 2002.

Monday, June 2, 14

Page 40: Making Data Usable: 2014 presentation at Datapalooza

STOP asking the wrong question

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

Charlie Chaplin in “Modern

Times” (1936)

Monday, June 2, 14

Page 41: Making Data Usable: 2014 presentation at Datapalooza

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

START Answering Right Question

HITECH Switches Emphasis to “Meaningful Use”

Monday, June 2, 14

Page 42: Making Data Usable: 2014 presentation at Datapalooza

Wrong Question:

X What can the computer do?

X How do we automate cognition?

X What is the transactional gain?

X How do we get users to conform?

Better Questions:

✓ What can humans do?

✓ How do we augment cognition?

✓ What is the relational gain?

✓ How do we optimizesociotechnical balance?

START Answering Right Question

Source: Blumenthal, D. (2010). Guiding the health information technology agenda. Interviewed by David J. Brailer. Health Aff (Millwood), 29(4), 586-595.

David Blumenthal

Monday, June 2, 14

Page 43: Making Data Usable: 2014 presentation at Datapalooza

Research on Communicating Data

Monday, June 2, 14

Page 44: Making Data Usable: 2014 presentation at Datapalooza

Making Data Talk

Inform Support Decisions

Educate Persuade

Adapting to a World of Ubiquitous Data Systems

Monday, June 2, 14

Page 45: Making Data Usable: 2014 presentation at Datapalooza

Making Data Talk

Inform Support Decisions

Educate Persuade

Adapting to a World of Ubiquitous Data Systems

Monday, June 2, 14

Page 46: Making Data Usable: 2014 presentation at Datapalooza

Making Data Talk

Inform Support Decisions

Educate Persuade

Adapting to a World of Ubiquitous Data Systems

Monday, June 2, 14

Page 47: Making Data Usable: 2014 presentation at Datapalooza

Making Data Talk

Inform Support Decisions

Educate Persuade

Adapting to a World of Ubiquitous Data Systems

Monday, June 2, 14

Page 48: Making Data Usable: 2014 presentation at Datapalooza

Chapter 4: Visual Displays

Monday, June 2, 14

Page 49: Making Data Usable: 2014 presentation at Datapalooza

SOURCE: http://alleydog.com/topics/sensation_and_perception.php

Perceptual Basics

Monday, June 2, 14

Page 50: Making Data Usable: 2014 presentation at Datapalooza

source: Carpenter PA, Shah P. A model of the perceptual and conceptual processes in graph comprehension. J Educ Psychol. 1999, 91(4): 690-702.

• Constructive process

• Gaze goes to center for pattern

• Contiguous labels for meaning

• Left to right tendency in western culture

• Perceptual rules guide meaning

Cognitive / Perceptual Research

Monday, June 2, 14

Page 51: Making Data Usable: 2014 presentation at Datapalooza

source: Carpenter PA, Shah P. A model of the perceptual and conceptual processes in graph comprehension. J Educ Psychol. 1999, 91(4): 690-702.

• Constructive process

• Gaze goes to center for pattern

• Contiguous labels for meaning

• Left to right tendency in western culture

• Perceptual rules guide meaning

Visualizing Long Term Change

Monday, June 2, 14

Page 52: Making Data Usable: 2014 presentation at Datapalooza

• Constructive process

• Gaze goes to center for pattern

• Contiguous labels for meaning

• Left to right tendency in western culture

• Perceptual rules guide meaning

Hans Rosling, BBC

Visualizing Change Dynamically

Monday, June 2, 14

Page 53: Making Data Usable: 2014 presentation at Datapalooza

Monitoring for Change in EHR Systems

Aging In Place, IntelRule of Thumb* for “Big Data” Systems

•Overview

•Zoom / filter

•Details on demand

*Ben Shneiderman

Monday, June 2, 14

Page 54: Making Data Usable: 2014 presentation at Datapalooza

Biases & Heuristics

Monday, June 2, 14

Page 55: Making Data Usable: 2014 presentation at Datapalooza

Overcome “small numbers” bias

Monday, June 2, 14

Page 56: Making Data Usable: 2014 presentation at Datapalooza

Exceptional Case

Fallacy of small numbers;Tversky & Kahneman, 1971

Illnesses322,000,000

Hospitalizations21,000,000

Prevented

Deaths732,000

Monday, June 2, 14

Page 57: Making Data Usable: 2014 presentation at Datapalooza

Improving Decision MakingProblem: Conditional (Bayesian) probabilities are counter-intuitive, arcane for practice.

source: Gigerenzer, Gerd, & Hoffrage, Ulrich. (1995). How to improve Bayesian Reasoning without Instruction: Frequency Formats. Psychological Review, 102(4), 684-704.

For example:

Monday, June 2, 14

Page 58: Making Data Usable: 2014 presentation at Datapalooza

95 out of 100 physicians estimated 70-80% instead

of a correct 7.8%

Improving Decision MakingProblem: Conditional (Bayesian) probabilities are counter-intuitive, arcane for practice.

source: Gigerenzer, Gerd, & Hoffrage, Ulrich. (1995). How to improve Bayesian Reasoning without Instruction: Frequency Formats. Psychological Review, 102(4), 684-704.

For example:

Monday, June 2, 14

Page 59: Making Data Usable: 2014 presentation at Datapalooza

Natural frequencies work better

source: Gigerenzer, Gerd, & Hoffrage, Ulrich. (1995). How to improve Bayesian Reasoning without Instruction: Frequency Formats. Psychological Review, 102(4), 684-704.

Monday, June 2, 14

Page 60: Making Data Usable: 2014 presentation at Datapalooza

See: Fagerlin, A., Ubel, P. A., Smith, D. M., & Zikmund-Fisher, B. J. (2007). Making numbers matter: present and future research in risk communication. Am J Health Behav, 31 Suppl 1, S47-56.

Icon arrays convey natural frequencies more effectively

Monday, June 2, 14

Page 61: Making Data Usable: 2014 presentation at Datapalooza

Portraying trends to policy makersChoropleth Maps: CDC Obesity Trends, BRFSS 1985

Monday, June 2, 14

Page 62: Making Data Usable: 2014 presentation at Datapalooza

Nonsegmented geographic data

Isopleth “Weather Maps,” HINTS

Monday, June 2, 14

Page 63: Making Data Usable: 2014 presentation at Datapalooza

Juxtaposing geographic distributions

Mortality Maps (SEER):Lung Cancer Mortality

For Example:Knowledge Maps (HINTS):Does Smoking Cause Cancer?

Monday, June 2, 14

Page 64: Making Data Usable: 2014 presentation at Datapalooza

Added User Controls14 datasets spanning 6 years

NSF, NIH Collaboration

Disolving Barriers Between Clinical and Community Health

Monday, June 2, 14

Page 65: Making Data Usable: 2014 presentation at Datapalooza

Added User Controls14 datasets spanning 6 years

NSF, NIH Collaboration

Disolving Barriers Between Clinical and Community Health

Monday, June 2, 14

Page 66: Making Data Usable: 2014 presentation at Datapalooza

Disparities Frame

Framing Effects

Monday, June 2, 14

Page 67: Making Data Usable: 2014 presentation at Datapalooza

Disparities Frame

Impact Frame

Framing Effects

Monday, June 2, 14

Page 68: Making Data Usable: 2014 presentation at Datapalooza

Disparities Frame

Impact Frame

Progress Frame

Framing Effects

Monday, June 2, 14

Page 69: Making Data Usable: 2014 presentation at Datapalooza

Disparities Frame

Impact Frame

Progress Frame

Framing Effects

3.25

3.5

3.75

4

4.25

4.5

4.75

5

Progress Impact Disparity

Low MistrustHigh Mistrust

3.25

3.5

3.75

4

4.25

4.5

4.75

5

Progress Impact Disparity

Low MistrustHigh Mistrust

I want to be screened for colon cancer? Framing X Medical Mistrust Questionnaire

Best influence on behavior

Monday, June 2, 14

Page 70: Making Data Usable: 2014 presentation at Datapalooza

Bottom Line

Monday, June 2, 14

Page 71: Making Data Usable: 2014 presentation at Datapalooza

Data: New “Intel Inside*”

Source: O'Reilly, Tim. (2005). What Is Web 2.0? Design Patterns and Business Models for the Next Generation of Software.

Tim O’Reilly

Monday, June 2, 14

Page 72: Making Data Usable: 2014 presentation at Datapalooza

Data = Power(by itself)

Monday, June 2, 14

Page 73: Making Data Usable: 2014 presentation at Datapalooza

Data = PowerUser Centered*

*i.e., made understandable, actionable, accessible Monday, June 2, 14

Page 74: Making Data Usable: 2014 presentation at Datapalooza

Research methods to address gap

See: Brinck, T., Gergle, D., & Wood, S. D. (2002). Designing Web sites that work : usability for the Web (1st ed.). San Francisco: Morgan Kaufmann Publishers.

Monday, June 2, 14