Decision Analysis

1
Decision Analysis

• Dr M G Dawes
• Centre for Evidence Based Medicine

2
Contents

• What is Decision Analysis?
• Decision Trees
• How they work
• Critical Appraisal of Decision Analysis
• A bit on economics
• An example of some software

3
What is Decision Analysis?

• Urinary Tract Infection
• Patient presents with symptoms
• What would happen if you opted for one path in
  preference to another
• On what basis would you assess outcome

4
How does one start?

• List all the options and display
• Commonly accepted format is a tree diagram.

Treat
UTI
Dont Treat
Decision Node
5
Decision data
Chance Node
better
0.9
Treat
Not better
0.1
UTI
better
0.2
Dont Treat
Not better
0.8
6
Result 0.9 is better than 0.2
7
Add reality to probability scores

• Cost of tests
• Cost of treatment
• Days lost from work
• Cost of re attending clinician
• Multiply the probabilities by the costs

8
Cost Data
Costs
2
better
0.9
Treat
9
Not better
0.1
UTI
better
2
0.2
Dont Treat
Not better
9
0.8
9
Rollback Costs
2
(20.9) (90.1)2.7
better
0.9
Treat
9
Not better
0.1
UTI
better
2
0.2
Dont Treat
Not better
9
0.8
(0.22)(0.89)7.6
10
Results

• More people get better (90 vs 20)
• It is cheaper (2.70 vs 7.60)

11
What does the patient think?

• Utilities
• How would you feel?
• QALYs
• Quality adjusted life year

12
QALYs
From Alastair Gray
13
Time Trade Off Method (to assess utility of a
health state)

• You have arthritis (severe - unable to walk to
  shops need a buggy in pain most of the time)
  and are aged 48
• Choose between living with arthritis until 80 or
  living in perfect health for a shorter length of
  time
• eg 50 60 70
• If 70 65 67 69 71 73 75
• Etc until chosen a year

14
Time Trade Off

• Utility is 1- (number of years willing to give
  up/(80-current age)
• If age selected was 75
• Utility 1-(80-75)/(80-48)
• 0.84
• The better your health the less the years you
  give up

15
Trade Off vs Age
16
Utilities
Utilities
9
better
0.9
Treat
2
Not better
0.1
UTI
better
9
0.2
Dont Treat
Not better
2
0.8
17
Rollback
9
(90.9) (20.1)8.3
better
0.9
Treat
2
Not better
0.1
UTI
better
9
0.2
Dont Treat
Not better
2
0.8
(0.29)(0.82)3.4
18
Results

• More people get better (90 vs 20)
• It is cheaper (2.70 vs 7.60)
• The utilities are better (8.3 vs 3.4)
• Probably should treat??
• Sensitivity analysis

19
UTI What are the options?

• Treat on symptoms alone
• Treat after doing a test
• Exercise what are the options
• Diagnosis
• Therapy
• Outcome

20
How does one develop this?

• Need to know all the baseline data.
• For UTI
• What proportion of patients with typical symptoms
  have UTI
• What is the sensitivity and specificity of a
  dipstix.
• What is the success rate of antibiotic treatment?

21
On what basis would you assess outcome?

• Bacterial eradication
• Symptoms
• Diary
• Questionnaire
• Interview
• Time off work
• What do patients think?

22
(No Transcript)
23
(No Transcript)
24
How are these studies appraised

• Are the results valid
• What are the results
• Can these be applied to my patients

25
Are the results valid?

• Were all important strategies and outcomes
  included?
• Were all of the realistic clinical strategies
  compared?
• Were all clinically relevant outcomes considered?
• Was an explicit and sensible process used to
  identify, select and combine the evidence into
  probabilities?

26
Validity Check (2)

• Were the utilities obtained in an explicit and
  sensible way from credible sources?
• Was the potential impact of any uncertainty in
  the evidence determined? 

27
What are the Results?

• In the baseline analysis, does one strategy
  result in a clinically important gain for
  patients? If not, is the result a toss-up?
•  How strong is the evidence used in the analysis?
  
•  Could the uncertainty in the evidence change the
  result?

28
Can I apply the results to my patient?

• Do the probability estimates fit my patients'
  clinical features?
• Do the utilities reflect how my patients would
  value the outcomes of the decision?

29
Bottom Line

• The power of decision analysis is not in the
  numbers at the decision node
• It is the ability to change the utilities and
  probabilities
• Watching how this affects the decision node
• Thus it should be seen as a dynamic tool
• Software for this is available

30
EXCEL or WWW.TREEAGE.COM

• Nice system
• Quite friendly
• Free download to try
• US 295 academic price