I. Introduction to the psychology of medical decision making. (Robert M. Hamm, PhD)

• A. Why should doctors know about decision psychology?
  

• 1. Medical decision making is not optimal
  
  2. The psychology of doctors' decision making is one partial cause of that. Green and Yates example: What knowledge is used? What is wrong? Is there a psychological explanation for this?
  
  3. Understanding this psychology is necessary for improving it.
  
  
  • a. Improvement of the doctor's thinking (through education, training, medical culture)
    
    b. Improvement of the system (provide information, time, incentives).
    
    

• B. What are the standards we compare doctors' performance to? (See IV-A-4.)
  

• 1. Decision theory - expected utility, multi-attribute utility, cost effectiveness
  
  2. Invariance - same decision should be made when same situation is described differently.
  
  3. Accuracy - diagnosis agreement with right answer, gold standard.
  
  4. Logical consistency.
  
  5. Use of all pertinent information that is available.
  
  6. Adherence to rules/standards that the field endorses.
  
  

• C. What ideas of psychology are useful for understanding expert medical doctors' decision making and for planning ways to improve it or maintain its quality.
  

• 1. The role of memory in medical thinking.
  
  
  • a. Complex, organized knowledge
    
    b. Rapid "pattern recognition" of situations
    
    c. Responses difficult to verbalize.
    
    d. Expert performance depends on memory.
    
    
    • i. Expert memory for cases is good.
      
      ii. Illness scripts: mental structures efficient for familiar problems.
      
    
    
• 2. Association based errors in decision relevant judgments.
  
  
  • a. Pattern recognition involves activation of associated ideas.
    
    b. Some associations may be irrelevant for optimal decisions. Examples:
    
    
    • i. Recency of event makes it appear more likely to happen again
      
      ii. Vividness of past experience makes it come to mind easily.
      
      
• 3. Decision making strategies.
  

• a. Strategy choice:
  

• 1. Accuracy costs.
  
  2. Strategies and their choice not always conscious.
  
  

• b. Strategies affect errors.
  

• 1. People may err in selecting which strategy to use.
  
  2. Analytical versus intuitive thinking. Different types of strategy have different characteristic errors.
  
  

• 4. Psychophysical processes in decision making.
  

• a. Judgments of probability
  
  b. Judgments of value
  
  c. Combination of information.
  
  
  • 1. SEU: combining probability and value Standard: expected utility judgments:
    
    2. Appropriate weighting of information Standard: theory based model or empirical model
    
    

• 5. Not covered here: Motivation, learning and training, social psychology -- perception of others and behavior toward others, emotion, psychodynamics, individual differences in personality or in ability, attitudes, doctor- patient relationship.
  
  

• D. The rest of our afternoon.
  

• Section 2: Diagnosis, associative memory processes. Al Connors, Jr.
  
  Section 3: Strategies in diagnosis and decision making. Rob Hamm.
  
  Section 4: Decision making, psychophysical processes. Gretchen Chapman.
  
  Section 5: Strategies for improvement. Gretchen Chapman.
  

II. Diagnostic thinking in medical decision making. (Alfred F. Connors, Jr., MD)

• A. The ideal diagnostic process.
  

• 1. Identify possible diagnoses.
  
  2. Assess the probability of each diagnosis.
  
  3. Seek additional information.
  
  4. Reassess probability of diagnosis.
  
  5. Choose the most probable diagnosis.
  
  

• B. Impediments to diagnostic accuracy.
  

• 1. Heuristics.
  
  2. Biases.
  
  3. Cognitive errors.
  
  

• C. Examples of heuristics, biases, and errors in diagnostic reasoning.
  

• 1. Availability.
  
  2. Representativeness.
  
  3. Anchoring and adjustment.
  
  4. Ego bias.
  
  5. Confidence, the illusion of validity.
  
  6. Value-induced bias.
  
  
  • a. Regret
    
    b. Wishful thinking
    
    
• 7. Hindsight bias.
  
  8. Cognitive errors
  

• a. Insensitivity to sample size.
  
  b. Gambler's fallacy.
  

III. The role of memory and cognitive strategies in diagnosis and decision making. (Robert M. Hamm, PhD)

• A. Memory and cognitive strategies.
  

• 1. Big knowledge base.
  

• a. Many concepts
  
  b. Complicated concepts
  
  
  • i. Chunks, big units.
    
    ii. Rules: If (x is true), do (y).
    
    
• c. For clinicians: scripts for familiar cases:
  

• i. what to look for.
  
  ii. what to do.
  
  

• d. Experienced clinicians have good memory for details of cases.
  
  

• 2. Access through pattern recognition.
  

• a. Pattern recognition is rapid.
  
  b. It is hard to explain how one does it.
  
  c. It is hard to notice errors, to change responses.
  
  

• 3. Heuristic strategies, satisficing.
  
  

• B. Judgment, when many sources of information must be combined. Categorization (diagnosis), prediction (prognosis).
  

• 1. Large individual differences in doctors' strategies for integrating information are often found, and are related to differences in their judgments.
  
  2. Sources of inaccuracy are revealed and explained
  
  
  • a. by differences between doctor's strategy for using information and the prescribed way to use the information,
    
    b. by the effort/accuracy tradeoff,
    
    c. by inconsistencies or inefficiences in use of information.
    
    

• C. Problem solving, when what to do is not immediately known.
  

• 1. Heuristic search.
  

• a. Work backward if not familiar with problem domain. Example: word problem.
  
  b. Work forward if familiar. Example: word problem.
  
  c. Complex example from surgery: (Abernathy and Hamm, 1995): surgical resident, lung leaking air
  
  

• 2. Choice of strategies for solving medical problems.
  

• a. Effort/accuracy tradeoff in the search for information. Example: Curley et al, choice of everyday information sources.
  
  b. Choice of strategy can be misguided by associations. Example: Wolf et al, choice of information sources in abstract task.
  
  c. Choice to use analytic strategy compared with familiar script does not guarantee result will be correct. Example: Dawes, error in application of test information.
  
  d. Contrast between strategy use in novel and familiar situations.
  

IV. Decision Making (Gretchen Chapman, PhD)

• A. Normative Principles (see handout)
  

• 1. Expected Utility Theory (EUT) for decisions under uncertainty.
  
  2. Multiattribute Utility Theory (MAUT) for decisions with multiple attributes/goals.
  
  3. Discounted Utility (DT) for decisions about outcomes at different points in time.
  
  4. Additional general principles.
  
  

• B. Psychophysical errors
  

• 1. reflection effect
  
  
  • a. risk averse for gains, risk seeking for losses
    
    
    • (1) Asian disease (Kahneman & Tversky, 1984)
      
      (2) mortality/survival (McNeil, et al., 1982)
      
      
  • b. integration/segregation
    
  
  
  • (1) mental accounting/integrate losses, segregate gains (Thaler, 1985)
    
    (2) group vs. individual perspective (Redelmeier & Tversky, 1990; 1992)
    
    
  • c. sunk cost effect (Arkes & Blumer, 1985)
    
    
• 2. loss aversion/endowment effect/status quo bias
  

• a. willingness to pay versus willingness to accept (Thaler, 1980)
  
  b. patient's self assessment (Wells, et al., 1993)
  
  c. adding alternatives increases preference for the status quo (Redelmeier & Shafir, 1995)
  
  

• 3. decision weights are not a linear function of probability: Allais paradox ( Allais, 1953)
  
  

• C. Other decision-making errors
  

• 1. hindsight bias/outcome bias (Arkes, et al., 1981; Dawson, et al., 1988; Baron & Hershey, 1988;)
  
  2. omission bias/naturalism (Ritov & Baron, 1990)
  
  3. failure to integrate probabilities and outcomes appropriately:
  
  
  • a. Estrogen Replacement: Physicians don't base their decisions on their own probability estimates (Elstein, et al. 1986).
    
    b. duration neglect (Redelmeier & Kahneman, 1993)
    
    
• 4. temporal discounting biases, e.g. dynamic inconsistency (Kirby & Herrnstein, 1995)
  

V. How to improve decision making (Gretchen Chapman, PhD)

Debiasing needs to be focused on the source of the error:

• A. strategy-based: corrected by
  
  
  • 1. increase incentives
    
    2. provide tools for effortful operations, e.g., computerized decision support
    
    3. make improved strategies available
    
    4. invest the effort upfront in decision analyses, practice guidelines, and prediction rules rather than in individual cases
    
    
• B. association-based: corrected by
  

• 1. cuing decision maker for neglected information
  
  2. formal instruction about what information is relevant.
  
  

• C. psychophysically-based: corrected by
  

• 1. re-framing; try multiple perspectives
  
  2. change reference point or try multiple reference points