decision analysis
DESCRIPTION
dTRANSCRIPT
1
CHAPTER 9
DECISION ANALYSIS
2
Decision Analysis
- The application of an analytical method for systemically
comparing different decision options
- Graphically displays choices and facilitates calculations of
values needed to compare options
- Assists in selecting the most cost effective alternative
- Assists in making decisions when the decision is complex and
there is uncertainty about some info
3
Steps in Decision Analysis1. Identify the specific decision
2. Specify alternatives
3. Draw the decision analysis structure
4. Specify possible costs, outcomes & probabilities
5. Perform calculations
6. Conduct a sensitivity analysis
4
1. Identify the specific decision- Specific decision to be evaluated should be clearly defined by
answering the following questions:
What’s the objective of the study?
Over what period of time will the analysis be conducted?
What perspective will be stated? (patient, organization, society…)
- For our example:
The objective is whether to add a new antibiotic to an institution
formulary (essential drugs basket)
The perspective is that of institution
Time period: 2 weeks
5
2. Specify alternatives- Ideally, the most effective alternative should be compared
- New products may be compared against a standard (older, well
established) therapy
- Decision analysis may compare more than 2 treatment options
(comparison of 5 statins)
- May compare intervention with no intervention
- For our example Antibiotic A (new medication) to be compared
with standard Antibiotic B
6
3. Draw the decision analysis structure- Branches (arms) of decision tree :
• Choice nodes
• Chance nodes
• Terminal (final outcome nodes)
7
8
4. Specify possible costs, outcomes &
probabilities
- For each option, information should be obtained for the
probability of occurrence & consequences of occurrence
- Probabilities are assigned for each branch of chance nodes
- The some of probabilities for each branch must be 1
- Consequences are reported as monetary outcomes, health
related outcomes or both
9
5. Perform calculations- The probability at the terminal node is calculated by multiplying
the probability of each arm from choice node to terminal node
- Total costs are calculated by adding all the costs over all the
branches from choice node to terminal node
- Multiply (C x P) for each node
- Sum all results for each choice to get the average cost per
patient
10
11
12
- Antibiotic B is less expensive even when including the costs of
treating adverse effects
- Antibiotic A is better clinical option (higher probability of
success and lower probability of adverse effects)
- Decision maker may use ICER or INB to determine whether to
add A to the formulary.
ICER= $700-$650 = $500 per extra success
0.9-0.8
13
- Using INB: if the suggested range is between 1000 & 2000 $:
- INB λ=1000 = +$50
- INB λ=2000 = +$150 then A is cost effective
6. Conduct a sensitivity analysis