Decision Analytic Approaches for Evidence-

Based Practice

Decision Analytic Approaches for Evidence-

Based PracticeM8120

Fall 2001

Suzanne Bakken, RN, DNSc, FAANSchool of Nursing & Department of Medical

InformaticsColumbia University

OutlineOutline• Health care decision making• Expected value decision making• Building a decision tree with Data

What is a Decision?What is a Decision?• A decision is an irreversible choice

among alternative ways to allocate valuable resources

What makes a decision What makes a decision hard?hard?

•Complexity•Uncertainty including limited

information•Dynamic effects•High stakes•Unclear alternatives•Unclear preferences

Are These Decisions?Are These Decisions?• A California utility faces likely electrical power shortages and is

considering constructing a power plant using either coal or nuclear energy.

• You are concerned whether you could have found a better deal on the CD player you just bought.

• A person with appendicitis is uncertain whether there will be unpleasant side effects from the appendectomy he is about to have.

• A graduate student is considering whether to pay a mechanic to fix her 8 year old car or trade it in on a newer model.

• An oil company is attempting to estimate oil prices one year from now.

• You are trying to decide if you should have the birthday party you’re planning for a friend outside.

Anatomy of a Health Practice Anatomy of a Health Practice DecisionDecision

• Goal - to choose the action that is most likely to deliver the outcomes that patients find desirable

• Outcomes of alternative practices must be estimated (primarily analytical)– Collection and analysis of evidence regarding benefits, harms,

and costs of each option– Subjective judgment

• Desirability of outcomes of each option must be compared (patient preferences/utilities) – Benefits compared with harms– Outcomes versus cost– Resources consumed (Eddy, 1990)

Decision Analysis: Expected Decision Analysis: Expected Value Decision MakingValue Decision Making

• Prescriptive• Analytic• Explicit

Basic ConceptsBasic Concepts

• Biological events random• Outcomes of illness uncertain• Outcomes of treatments uncertain• Must choose between treatments - a

gamble• Utility - a measure of preference• Expected value - result expected on

average

Quantifying UncertaintyQuantifying Uncertainty

• Probability as a language for expressing uncertainty

• Bayes’ theorum for probability revision

Probability FundamentalsProbability Fundamentals

• Strength of belief• A number between 0 and 1 that expresses an

opinion about the likelihood of an event• Probability of an event that is certain to occur

is 1• Probability of an event that is certain to NOT

occur is 0

DefinitionsDefinitions• Prior probability - the probability of an

event before new information (finding) is acquired; pretest probability or risk

• Posterior probability - the probability of an event after new information (finding) is acquired; posttest probability or risk

• Probability revision - taking new information into account by converting prior probability to posterior probability

Role of Probability Revision Role of Probability Revision TechniquesTechniques

Abnormal Finding

Diagnosis

BeforeFinding

AfterFinding

0 1Probability of Disease

Prior Probability Posterior Probability

Role of Probability Revision Role of Probability Revision TechniquesTechniques

Negative Finding

Diagnosis

AfterFinding

BeforeFinding

0 1Probability of Disease

Posterior Probability Prior Probability

Steps in Decision AnalysisSteps in Decision Analysis• Create a decision tree

– Identify and bound problem– Structure the problem– Characterize information needed

• Calculate the expected value of each decision alternative

• Choose the decision alternative with the highest expected value (payoff, utility)

• Use sensitivity analysis to test the conclusions of the analysis

Whose View?Whose View?

• Individual patient• Physician• Society• Government• Healthcare institutions

Create the Decision TreeCreate the Decision Tree

• Define the decision problem• Identify the decision alternatives• List the possible clinical outcomes of each of the

decision alternatives• Represent the sequence of events leading to the

clinical outcomes by a series of chance nodes and decision nodes

• Choose a time horizon for the problem• Determine the probability of each chance outcome• Assign a value (preference, utility, payoff) to each

clinical outcome

Simple Decision TreeSimple Decision Tree

Operate

Do not operate

Disease present

Disease present

Disease absent

Disease absent

Outcome; Treatment with disease

Outcome; Treatment without disease

Outcome; Treatment with disease

Outcome; Treatment without disease

Represent Sequence of Represent Sequence of EventsEvents

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative Death

Palliate

Operative DeathOperative Death

Survive

Survive

No cure

Cure

Cure

No Cure

No cure

Cure Try for the cure

Determine Probability of Each Chance Determine Probability of Each Chance OutcomeOutcome

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative deathOperative death

Survive

Survive

No cure

Cure

Cure

No Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.02

p=.98 p=.10

p=.90

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

Assign ValuesAssign Values

• Utilities, preferences, payoffs– Mortality– Length of survival– Cost– Quality of life– Quality of life years

Standard gamble

Assigning Values to the Decision Assigning Values to the Decision AlternativesAlternatives

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative death U=0Operative death U=0

Survive

Survive

No cure

Cure

Cure

No Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.02

p=.98 p=.10

p=.90

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

U=2

U=20

U=2

U=20

U=20

U=2

U=20

U=0

Path ProbabilityPath Probability

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative deathOperative death

Survive

Survive

No cure

Cure

Cure

No Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.02

p=.98 p=.10

p=.90

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

Path probability of a sequence of chance events is the product of all probabilities along that sequence (summation principle)

Folding Back the TreeFolding Back the Tree

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative death U=0Operative death U=0

Survive

Survive

No cure

Cure

Cure

No Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.02

p=.98 p=.10

p=.90

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

U=2

U=20

U=2

U=20

U=20

U=2

U=20

U=0

Folding Back the TreeFolding Back the Tree

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative death U=0Operative death U=0

Survive U=20

Survive

No cure

Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.02

p=.98

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

U=2

U=20

U=2

U=20

U=20

U=0

.1 X 20 + .90 X 2 = 3.8

Fold It AgainFold It Again

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative death U=0Operative death U=0

Survive U=20

Survive

No cure

Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.02

p=.98

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

U=2

U=20

U=2

U=20

U=20

U=0

.1 X 20 + .90 X 2 = 3.8

Fold It AgainFold It Again

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

Operative death U=0

Survive

No cure

Cure

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.10

p=.90p=.90

p=.10

p=.01

p=.99

Try for the cure

U=2

U=20

U=2

U=20

U=20

U=0

U = .98 X 3.8 + .02 X 0 = 3.72

Try for Cure Vs. PalliativeTry for Cure Vs. Palliative

Operate

Do not operate

Disease present

Disease absent

Disease present

Disease absent

Survive

Operative death

Palliate

No cure

Cure

p=.10

p=.90

p=.10

p=.90

p=.90

p=.10

p=.01

p=.99

Try for the cure

U=2

U+20

U=20

U=0

U = .98 X 3.8 + .02 X 0 = 3.72

U = .90 X 18.2 + .10 X 0 = 16.38

Final Fold - Operate Vs. Do Not Final Fold - Operate Vs. Do Not OperateOperate

Do not operate

Operate

U=18.38

U=19.46

Sensitivity Analysis: What Happens if Sensitivity Analysis: What Happens if Probability of Disease Changes?Probability of Disease Changes?

p(disease) Surgery Medical

0 19.80 20.00

.10 19.46 18.38

.20 19.11 16.76

.30 18.77 15.14

.40 18.43 13.52

.50 18.09 11.90

Should the decision still be operate?

Sensitivity Analysis: What Happens if Sensitivity Analysis: What Happens if Operative Mortality Changes?Operative Mortality Changes?

p(operative death) Surgery Medical

0 19.64 18.38

.05 19.55 18.38

.10 19.46 18.38

.15 19.37 18.38

.20 19.28 18.38

.25 19.19 18.38

Should the decision still be operate?

Expected Value Decision Making

• Data for Healthcare• Exercise #1