Enhancing Social Networks to Improve Innovation

Download Report

Transcript Enhancing Social Networks to Improve Innovation

Getting Unstuck:
Creative Problem Solving
Ken Abrams, Ph.D.
You-Know
Copyright  2007 You-Know
1
Objectives
• Explore why we need to be creative and
how to get that way
• Propose a method to leverage new
technology to improve problem solving
creativity
Copyright  2007 You-Know
2
The Challenge
Think different!
Copyright  2007 You-Know
3
• To get better solutions, we need to
– Escape from pre-conceptions
– Come up with more creative, complete solutions that
fit more of the data
Copyright  2007 You-Know
4
Creativity: Why now?
•
Internet
–
–
–
–
–
•
New technology challenges and threats
–
–
•
Deal with global warming
Discover opportunities
Healthcare
–
–
•
Genetic engineering
Nano-technology
Pressing need for alternative energy
–
–
•
Disseminating ideas and innovation
Place to find knowledge and express opinion
Universalizing access to education
Speeding globalization
Flattening the world
More effective use of resources
More efficient delivery and payment
Security
–
–
–
Clash of civilizations
Understand other cultures and thinking
Cross-culture collaboration
Copyright  2007 You-Know
5
Markets that need more creative
problem solving:
– Medicine
– Education
– Automotive
– Energy
– Military
– Government
– Non-profits
– NGOs
Copyright  2007 You-Know
6
Brief History of Problem-Solving
Creativity
Copyright  2007 You-Know
7
Functional Fixedness (Maier 1931)
• Inability to use an object in an atypical way to
solve a problem: stuck in the frame
– Problem: Tie together
two ropes hanging too
far apart to grab at the
same time
– Solution: Use wrench
as a pendulum weight
to swing distant rope
within reach
• Increase effect
– If S tightens a nut with the wrench, less likely to see new use
• Decrease effect
– If E bumps into rope and starts it swinging, S is more likely to
use wrench as a pendulum weight
Copyright  2007 You-Know
8
Alternative Uses (of a Brick) Task
Guilford 1967
• Originality
– Fewer people who give a response, the more original
• Fluency
– Total number of ideas
• Flexibility
– Number of different categories ideas fall into
• Elaboration
– Detail of each idea
Copyright  2007 You-Know
9
Remote Associates Test
Mednick & Mednick 1967
• Measures ability to find a common word or
concept that links three otherwise
unrelated words
• Example
– “Falling Actor Dust”
STAR
Copyright  2007 You-Know
10
Divergent/Convergent Problem Solving:
Brain-storming (Osborn & Parnes 1967)
• Step 1. Goal finding
– D: What do we want?
– C: Rank importance.
• Step 2. Fact finding
– D: What need to know?
– C: Which first issues?
• Step 3. Problem finding
– D: How to? Challenge?
– C: Work on which?
• Step 4. Idea finding
– D: New ways to do this?
– C: Best ideas? new? risky?
• Step 5. Solution finding
– D: Decision points? Value?
– C: Most relevant? Likely
succeed?
• Step 6. Acceptance finding
– D: Obstacles? Aids?
– C: Change needed? Success
measures? Implement steps?
Copyright  2007 You-Know
11
Lateral Thinking
deBono 1971
• To break out of functional fixedness, and come
up with new solutions, start with a new thought,
a new metaphor
• Lateral Thinking: techniques for generating a
new starting place to create a novel solution
–
–
–
–
Recognize dominant ideas
Search for different perspectives
Relax control
Use chance to create new starting points
Copyright  2007 You-Know
12
Watanabe’s Ugly Duckling Theorem
• Any one thing has an infinite number of attributes
• Any two things have an infinite number of attributes in
common and an infinite number of attributes in only one
• Similarity is selecting the set of attributes to look at.
Choosing Set 1 vs. Set 2: yields different similarity levels
• Similarity is in the eye of the beholder; it is not intrinsic in
the objects themselves
• Perception and thought affect judgment, because one
set of attributes is favored over another
Copyright  2007 You-Know
13
Copyright  2007 You-Know
14
Biases in Decision-Making Under
Uncertainty (Tversky & Kahneman 1973)
• People prefer avoiding a loss over making a gain
• Availability
– Judged likelihood is proportional to how easy it is to
think of relevant (recent) examples
• Confirmatory bias
– Once have a hypothesis, then accept only confirmatory
data
• Gambler’s fallacy
– Random event is less likely to occur again, because it
happened recently
Copyright  2007 You-Know
15
Judgments of emotional tone are fast (Gladwell)
• Expert can assess emotion in person or
relationship in two seconds
– Fast, efficient unconscious processing
– Students can decide how good a professor is from
2 sec of video tape: After 2 sec judgment is highly
correlated with judgments after a semester w. prof
• Expert can identify the style and emotional
valence of a person or relationship in seconds
– By walking through a person’s home
– By viewing video of a married couple
Copyright  2007 You-Know
16
Instinctive Expert: “thin slicing”
• Makes accurate perceptual judgments, takes quick
effective action, executes skilful movement
• Can’t explain how she knows or performs
– Fraud experts, tactical battalion commanders, CEOs, athletes,
machinists, poets, performing artists
• Knowledge is the act of perception or performance
– There is no knowledge of how the act is performed
• Knowledge is not structured as verbal principles
– Non-verbal knowledge is not accessible for general problem
solving
• Not committed to theory: pragmatic, quick
• Disconnect between verbal explanation and non-verbal
knowledge
– Explanation distracts from the task and degrades performance
Copyright  2007 You-Know
17
Systematic Expert
• Synthesizes multiple channels of data
– Visual, tactile, text, numerical, computational
• Writes reports and papers
– Doctor, lawyer, academic, other professional, guru
• Organizes knowledge in a verbal representational
structure under general principles
– Even if supporting evidence or data is non-verbal
• Knowledge is verbal and accessible for solving problems
– That can be represented in natural language
• Can articulate principles or evidence behind one’s
thinking
– Because knowledge representation is organized under general
principles and is verbal
• Invested in the structure of one’s knowledge
– Gives up theory only
with extreme reluctance
Copyright  2007 You-Know
18
Systematic vs. Instinctive expertise
• Best judgments are a balance of both
• Depends on the domain
– Diagnosis (internal medicine, military strategy) =
systematic
– ER triage and stabilization = instinctive
– Dynamic (skiing, skirmishing, performing) = instinctive
– Judging emotional relatedness (sales, interaction) =
instinctive
– Articulating cues in face to face interaction =
systematic
Copyright  2007 You-Know
19
Physicians making decisions
• Rapid decisions (minutes)
– Surgeons suppress uncertainty through action
– Emergency physicians: act fast to stabilize
– Radiologists: pressured by referring doc to
commit to diagnosis
• High case load forces rapid decision
• Describing observations with no immediate
relevance to the question asked can have value,
but it takes longer
Copyright  2007 You-Know
20
Physicians making decisions
• Systematic decisions
– Cardiology Algorithm (decision tree of best
practices) is most accurate predictor of which
patient is having a heart attack at the moment
– Enforces systematic approach
– More accurate than non-experts
– Extra information reduces accuracy
Copyright  2007 You-Know
21
Systematic decisions: encourage
others adopt (Welch 2005)
• Leaders should probe and push with a
curiosity that borders on skepticism…
• Every conversation ..about a decision, a
proposal, or .. market information has to
be filled with .. “What if?” “Why not?” and
“How come?”
Copyright  2007 You-Know
22
How Doctors Think (Groopman)
• Even the most accomplished physician can miss
a key clue about a person’s true diagnosis
• Misdiagnosis is a window into the medical mind
revealing why
–
–
–
–
Doctors fail to question assumptions
Their thinking is closed or skewed
They fall into cognitive traps
Deliver poor care
• More than15% of all diagnoses are inaccurate
(Comparing diagnosis with autopsy, 1995)
Copyright  2007 You-Know
23
Cognitive errors
• Framing bias
– Doctors use shorthand to categorize patients
• I have a case of diabetes and renal failure
• I have a drug addict here in the ER with fever and
a cough from pneumonia.
– Accepting frame as given can be a serious
error, because
• A frame sticks esp. when pronounced by an expert
• An erroneous frame constrains thinking delays an
accurate diagnosis, sometimes for years
Copyright  2007 You-Know
24
Senior Attending: instinctive decisions
• When encountering a patient in an emergency: doesn’t
reason at all
• Expert clinician forms idea of what’s wrong in 20
seconds
– Pattern recognition: immediate perception leads to a gestalt
• Physician begins diagnosis from the first second
– Pallor or ruddiness, tilt of head, movement of eyes and mouth,
how sits and stands, timbre of voice, depth of breathing
• Notions of what is wrong evolve in the next minutes
– Peer into eyes, listen to heart, press the liver, inspect initial xrays
• Come up with 2-3 possible diagnoses from the outset of
meeting a patient
– Talented docs use heuristics to generate 4-5 diagnoses from
incomplete information
Copyright  2007 You-Know
25
Cognitive errors
• Attribution error
– Patient fits negative stereotype (alcoholic,
drug user, vagrant,…), rather than someone
who can’t respond because he’s ill
• Representativeness error (framing)
– Thinking constrained by a prototype: no
contradictions considered
Copyright  2007 You-Know
26
Cognitive errors
• Availability error
– Diagnosis comes readily to hand, because one’s seen numerous
cases of this infection or this addiction over recent weeks
– Pattern recognition is distorted by local ecology
• Confirmation bias:
– Strong belief in hypothesis causes cherry-picking to find data
that fits and ignore data that doesn’t
• Anchoring
– Latch onto a single possibility, confident that the anchor has
been thrown down just where it needs to be
– Don’t consider alternate explanations
– Leads to skewed reading of the facts
– Ex: Estrogen helps with menopause
• Gynecologists and cardiologists are on opposite sides
Copyright  2007 You-Know
27
ER: rapid action is valued, but systematic
works better
• Swift and decisive action saves lives, BUT
• Studied calm - consciously slowing thinking and
action - avoids being distracted by the chaotic
atmosphere
• Shooting-from-the-hip causes anchoring and
availability bias
– Can lead to misdiagnosis
• Ask What’s the worst it can be? To slow down
and broaden thinking
Opportunity: Intelligent diagnostic aid
Copyright  2007 You-Know
28
Cognitive errors
• Momentum of the diagnosis (band-wagon effect)
– Once an expert (specialist) fixes a label to the problem, it stays
attached, because the expert is usually right
• Worked-up the yin-yang error
– Multiple specialists and tests discourage innovative divergent
thinking
– What new could be possibly found?
• Denial of uncertainty
– Retain control by making the world more certain
– Even at the expense of falsifying data
– Denying uncertainty makes action possible: breaks paralysis
Copyright  2007 You-Know
29
Cognitive errors
• Commission bias
– Move toward action, rather than inaction
– When over-confident, ego inflated, desperate
– But no action can be the best course
• Satisfaction of search error
– Search stops when first thing is found
– But may be more than one thing to be found
Copyright  2007 You-Know
30
Cognitive errors
• Inside-the-box error
– Can’t think fresh, when test and clinical data
don’t fit
– Sometimes have to investigate multiple
causes to explain the data
– Ask What else could this be?
• Even in face of the obvious
Opportunity: Intelligent aid
Copyright  2007 You-Know
31
Conclusions about cognitive errors
• ER doctors: shoot-from-the-hip instinct is prized
• Radiologists: finding “the gestalt” is a mark of good
training
• Studies of radiologists show high error rates
– Going on first impressions missed important findings
– Screen for normality: 60% failed to note a clavicle was missing
– Searching cancer (look at all structures carefully): 17% failed to
note missing clavicle
– Confidence is no indicator of accuracy
• Poor performers were as confident as best
– Average diagnostic error interpreting medical images = 20-30%
• Studies of internists show high error rate
– Physical exam for cyanosis: 73% error
– Reading EKG for myocardial infarction and other abnormalities:
46% error
Copyright  2007 You-Know
32
How to avoid errors
• Radiology
– Slow down the process of perception and analysis
– Be systematic
– Note observations before drawing conclusions: sometimes an
exact diagnosis can’t be made: resist over-diagnosing
– Provide the complete clinical story; don’t ask for an answer to a
single focused question
– Generate a short list of alternatives
– Use a structured checklist
• Observations that seem irrelevant can have clinical import
• Going with your gut sometimes doesn’t work
• A checklist forces one to work in a stepwise way. It leads to more
accurate diagnosis, even though examining the image to address
only the specific question asked is quicker.
– Use computer assistance
• Improved detection of cancer: decreased false negatives 14-24%,
but increased false positives 10%
Copyright  2007 You-Know
33
How to avoid errors
• Radiology
– Train experts to use a controlled vocabulary
• Calibrate language to code perception accurately
and consistently
• Clarifies what is NOT implied
– “X is not enlarged” /=> X is normal
– Computer enhance the image to improve
contrast
• Clear border rather than blurry edge
• Boost clarity of “objects” in normal tissue
Copyright  2007 You-Know
34
A New Approach
Copyright  2007 You-Know
35
Develop intelligent aids
• Capture best practice algorithms
• Includes hints and suggestions to encourage lateral
thinking to explore alternate causes of observed data
• Solicits input from experts in the community of practice
to resolve or synthesize standards of treatment in
different hospitals and regions
• Pushes relevant vetted studies to the clinician as she
works on the diagnosis
• Uses sharing over the Internet to unify the standard of
care within a national (or international) community of
practice based on empirical studies of outcome
• Allows clinician with a difficult case to request input from
experts with relevant experience
Copyright  2007 You-Know
36
Advantages and Benefits
• Brings new ideas to the table; new starting points, new combinations
• Information and opinions are drawn from an expanded range of
starting assumptions and different regions of the problem space; a
wider range of perspectives
• More interaction and synthesis of ideas
• Develops a persistent repository and an audit trail of what ideas
went into a decision
• Transparency (can be set at different levels, but more is better)
• Encourages clinicians to consider a wider range of considerations in
their solutions
Copyright  2007 You-Know
37
Considerations
• How can this approach deliver enough value in more efficient
diagnosis and treatment to have it adopted by clinicians throughout
the community practice?
• How will payers react?
• How to accelerate adoption? Become a standard of best practice?
Cf. InterQual
• Can the system if adopted by the COP, become a channel of wider
collaboration?
Copyright  2007 You-Know
38
Contact
• Ken Abrams, Ph.D
You-Know
Tel 617-519-0500
[email protected]
Copyright  2007 You-Know
39