2. What is Decision Analysis?What is Decision Analysis?
Best course of action underBest course of action under
conditions of uncertaintyconditions of uncertainty
Rationalize decision makingRationalize decision making
Basic conceptsBasic concepts
Expected valueExpected value
Decision trees/IngredientsDecision trees/Ingredients
ExamplesExamples
Audience participationAudience participation
3. Key Concept: Expected ValueKey Concept: Expected Value
Average payoff ofAverage payoff of
random trialrandom trial P x VP x V
P(outcome)P(outcome)
Outcome valueOutcome value
Expected value mayExpected value may
not be a possiblenot be a possible
outcomeoutcome
Gambling exampleGambling example
4. Should I Buy a Ticket?Should I Buy a Ticket?
Wikipedia: California Lottery Mega MillionsWikipedia: California Lottery Mega Millions
Minimum payoff is $12,000,000Minimum payoff is $12,000,000
P(Winning) is 1/175,711,536P(Winning) is 1/175,711,536
Cost is one dollarCost is one dollar
6. Types of Expected ValuesTypes of Expected Values
Expected ValueExpected Value
P(outcome)P(outcome)
Value of outcomeValue of outcome
DollarsDollars
Years of survivalYears of survival
UtilityUtility
QALYQALY
7. Outcome: UtilityOutcome: Utility DesirabilityDesirability
of a Health Stateof a Health State
Expected Value= P(outcome) x ValueExpected Value= P(outcome) x Value
9. Decision Analysis IngredientsDecision Analysis Ingredients
Comparing expectedComparing expected
values of differentvalues of different
options, X vs Yoptions, X vs Y
Decision treeDecision tree
OutcomesOutcomes
ProbabilitiesProbabilities
Expected Value= P(outcome) x ValueExpected Value= P(outcome) x Value
19. Shortcomings of Decision AnalysisShortcomings of Decision Analysis
Risk AverseRisk Averse
QALY (Y x U)QALY (Y x U)
Years survivalYears survival
UtilityUtility
SubjectiveSubjective
Doctor/patient/familyDoctor/patient/family
Other NumbersOther Numbers
LiteratureLiterature
UncertaintyUncertainty
Test accuracyTest accuracy
CostsCosts
20. Sensitivity AnalysisSensitivity Analysis
Technique that recognizes range of valuesTechnique that recognizes range of values
ProbabilitiesProbabilities
OutcomesOutcomes
TypesTypes
One-WayOne-Way
Two-WayTwo-Way
Three-WayThree-Way
26. One Way Sensitivity AnalysisOne Way Sensitivity Analysis
Range
.000001 to .1
27. Two Way Sensitivity AnalysisTwo Way Sensitivity Analysis
Range Cath
Mortality
.000001 to .1
Range CAD
Probability
.05 to .95
28. Three Way Sensitivity AnalysisThree Way Sensitivity Analysis
Sensitivity Analysis on
Mortality of Cath and Sensitivity of Cath and Sensitivity of CT
Mortality of Cath
SensitivityofCath
0.00 0.03 0.05 0.08 0.10
0.990
0.943
0.895
0.848
0.800
Cardiac CT
Coronary Cath
Sensitivity of CT = 0.800
Sensitivity Analysis on
Mortality of Cath and Sensitivity of Cath and Sensitivity of CT
Mortality of Cath
SensitivityofCath
0.00 0.03 0.05 0.08 0.10
0.990
0.943
0.895
0.848
0.800
Cardiac CT
Coronary Cath
Sensitivity of CT = 0.848
Sensitivity Analysis on
Mortality of Cath and Sensitivity of Cath and Sensitivity of CT
Mortality of Cath
SensitivityofCath
0.00 0.03 0.05 0.08 0.10
0.990
0.943
0.895
0.848
0.800
Cardiac CT
Coronary Cath
Sensitivity of CT = 0.895
Sensitivity Analysis on
Mortality of Cath and Sensitivity of Cath and Sensitivity of CT
Mortality of Cath
SensitivityofCath
0.00 0.03 0.05 0.08 0.10
0.990
0.943
0.895
0.848
0.800
Cardiac CT
Coronary Cath
Sensitivity of CT = 0.943
30. Do it Yourself Decision AnalysisDo it Yourself Decision Analysis
Problem: You chooseProblem: You choose
IngredientsIngredients
Decision tree: time flows left to rightDecision tree: time flows left to right
ChoicesChoices
OutcomesOutcomes
ProbabilitiesProbabilities
ValuesValues