Transcript Cost-effectiveness Analysis

Cost-effectiveness Analysis:
A practical primer
Eran Bendavid
CEA is a comparative
analysis
• First step is to identify your alternatives
• Second step is to identify your alternatives
– Clinical management: medication vs. surgery,
medication A vs. B
– Prevention: program vs. no program, or universal vs.
targeted to high risk individuals, or vs. treatment
• Focus your question
• When you’re done, start over and see if more alternatives
popped up while you were completing your analysis
CEA is a comparative
analysis
• Alternatives can be treatment options, prevention
strategies, or any combination.
• Example: How to address a pandemic influenza threat…
– Do nothing
– Treat everyone
– Treat only confirmed cases
– Close schools
– Home quarantine for suspected cases
– Combinations of pharmacological and nonpharmacological strategies
Steps in conducting a costeffectiveness analysis
(1) Define analysis: explicit and specific statement of the
problem being tackled.
(2) Construct conceptual model: allow for all relevant
alternatives
Steps in conducting a costeffectiveness analysis (cont’d)
(3) Determine input values: identify all costs; decide on
measure of effectiveness.
(4) Collect costs and health outcomes; summarize by
incremental values; plot on graph
(5) Prepare manuscripts. You’re nearly done.
Construct conceptual model
• How do alternatives affect your model?
• “Make things as simple as possible, but not simpler”…AE
• How will you model be affected if you are considering
mass treatment versus treatment of confirmed cases
only?
Treat +s
Mass Rx
Treat -s
Pandemic
threat
True +, Rx
Test +
False +, Rx
Rx
confirmed
True -, no Rx
Test False -, no Rx
S, no
Rx
I, no
Rx
R, no
Rx
S, yes
Rx
I, yes
Rx
R, yes
Rx
S,
true +
(rx)
I, true
+ (rx)
S,
false+
(rx)
I,
false
+ (rx)
S,
true -
I, true
-
S,
false -
I,
false -
R, yes
Rx
R, no
Rx
Costs
• Which costs do you count?
– All direct costs:
• Hospitalizations
• Medications
• Vaccinations
• Diagnostics
Costs
• Which costs do you count?
– Indirect costs:
• Time costs
• Cost of lost productivity
• Opportunity cost
– Careful of double counting:
• Double counting indirect costs
• Indirect costs and quality of life adjustments
Measures of Effectiveness
• Mortality (deaths or deaths averted)
 Morbidity: e.g., episodes of illness, infections, duration of
disability (e.g., years of sight)
 Life years: expected duration of life
 Quality-adjusted life years (QALYs): life years
x utility scores
 Disability-adjusted life years (DALYs):YLL+YLD
Why are DALYs and QALYs best?
Here’s an example
• Aneurysm: clinical situation = woman, aged 50, with
unruptured cerebral aneurysm found incidentally.
Options = no treatment or surgery (clipping).
• Perspective = societal. i.e., economic effects on patients,
providers, insurers, etc not separated. All costs counted,
regardless of who pays.
• Effectiveness measure is QALY gained.
This CEA compares surgical clipping to no treatment
for the management of an asymptomatic cerebral
aneurysm, for a 50 year old woman, estimating the societal
cost per QALY gained.
Cost inputs
Cost input
Value (range)
Source
Clipping
$25,150 (18,000-35,000)
Cohort study –
cost accounting
system
Moderate/severe
disability
$20,000/yr (13,000-30,000)
SAH hospitalization
$47,000 ($33,000-$67,000)
Discount rate
3% (0-5)
Published
estimate
Cohort study –
cost accounting
system
CEA guidelines
Tally costs and
effectiveness
• Each health state in the model is associated with unique
costs and effectiveness
• Sum up the costs and benefits of strategies
• Put it in a table and on a graph
This CEA compares surgical clipping to no treatment
for the management of an asymptomatic cerebral
aneurysm, for a 50 year old woman, estimating the
societal
cost per QALY gained.
QALYs
disc
Normal survival,
worry
No aneurysm rupture
0.9825
No surgery
21.37
$534
Die
Aneurysm rupture
0.45
0.0175
Survive
0.55
Cost
disc
21.4
$
13.3
$ 23,941
21.4
$ 35,912
Normal survival
21.5
$ 25,150
Early death
13.4
$ 49,091
Normal survival
21.5
$ 61,062
Disability,
shorter survival
4.8
$218,700
Immediate death
0.0
$ 25,150
Early death,
worry
Normal survival,
worry
-
Ms. Brooks
No aneurysm rupture
Differences
-1.63
Ä QALYs
$39,132
Ä$
Dominant
$ / QALY
Strategy
Key Inputs
Rupture risk/yr
Expected life span
RR rupture w/ surgery
Surgical mortality
Surg morb (disability)
Cost of aneurysm hosp
1
Survive surg.
Die
0.902
Aneurysm rupture
Clipping
0
19.74
$39,666 Surgery-induced disability
0.0005
35
0
0.023
0.075
$ 47,000
0.075
Surgical death
0.023
0.45
Survive
0.55
The cost per QALY gained is
defined as:
Cost with surgery - cost with no surgery
QALYs with surgery - QALYs with no surgery
I.e.,
Δ Cost
Δ QALYs
Formulation must be incremental: from no intervention to
intervention, or from lower cost to higher cost intervention.
CEA Framework
Costs
Effectiveness
CEA Framework
Costs
CE ratio relevant
CE ratio irrelevant and
not interesting
Effectiveness
CE ratio irrelevant and
interesting
Treat everyone vs.
confirmed cases for H1N1
$12k
Change in costs
Comparator:
Confirmed cases
$1000
per
DALY
$500
per
DALY
$6k
$100
per
DALY
$0
0
5
10
Gain in health benefit (DALYs)
Treat everyone vs.
confirmed cases for H1N1
$12k
Comparator:
Confirmed cases
Change in benefit: 4
DALYs
Incremental CER:
$2,900/DALY
Change in costs
Change in cost:
$11,600
Treat
everyone
$1000
per
DALY
$500
per
DALY
$6k
$100
per
DALY
$0
0
5
10
Gain in health benefit (DALYs)
Treat everyone vs.
confirmed cases for H1N1
$12k
Comparator:
Confirmed cases
Change in cost:
$4,720
Change in benefit: 8
DALYs
Incremental CER:
$560/DALY
Change in costs
Assumption: At
high-risk for
infection
Treat
everyone
$6k
$1000
per
DALY
$500
per
DALY
High risk
$100
per
DALY
$0
0
5
10
Gain in health benefit (DALYs)
Base case graphically
$
$39,666
$534
19.74
21.37
QALYs
Base case graphically
$39,666
$
$534
0
19.74
21.37
QALYs
In manuscript, the results might be
presented as follows.
Scenario
QALYs
Total
Incremental
Costs
Total
Incremental
$ / QALY
No symptoms, <10 mm, no past SAH
 No treatment 21.37
 Clipping
19.74
--1.63
$534
--
$39,666 $39,132
-Dominated
CEA is iterative
 Steps usually in order, more or less.
 Often desirable to refine or redefine the analysis as it
progresses
 Good news: Until published, can revise.
 Bad news: Until published, can revise.
Dominance
$10k
ICERs:
Comparator vs A: Dominated
(strictly)
C vs B: ($9,000-$2,800) / (7-5)
=$3,100/QALY
D vs B: ($6,200-$2,800) / (5.5-5)
=$4,800/QALY
Dominated by extended
dominance
Strategy D
Costs
B vs A: ($2,800-$1,000) / (5-2)
=$600/QALY
Strategy C
$5k
Comparator
Strategy B
Strategy A
$0
0
5
10
Gain in health benefit (QALYs)
CEA of HIV prevention
strategies
Scenario
QALYs
Program Costs
Total Added Total Added $ / QALY
No prevention
20,000 --
$0
--
--
Targeted
20,025 25
$20,000 $20,000
$800
Universal
20,027 2
$200,000 $180,000
$90,000
12000
10000
Cost
8000
6000
4000
2000
0
19.5
20
20.5
Years of life
21
21.5
Drugs
Cost
Years of Life
Absolute Difference Absolute Difference
$0
-20
-$100
$100
20.4
0.4
$1,000
$900
21.2
0.8
$10,000
$9,000
21.3
0.1
No treatment
Drug A - old generic
Drug B - old brand name
Drug C - new brand name
Cost per Year of Life Gained
Incremental
Versus No Rx
(correct)
(misleading)
$250
$250
$1,125
$833
$90,000
$7,692
In this Workbook, there are 4 examples of the imporance of incremental CE analyses. The first two sheets are
examples of non-embedded treatment options, the third is another embedded example, and the fourth is a sequential
algorithm.
12000
10000
In the example in this sheet, each more expensive drug adds some years of life compared with less expensive
alternatives. What drug would you use based on cost-effectiveness?
8000
Cost
Assume that all drugs will work at the level indicated, i.e., that patients have not failed treatment with any of the drugs.
(A sequential drug algorithm is considered later.)
6000
The incremental cost-effectiveness of the generic drug is very attractive, and the incremental CE of the old brand name
drug is also attractive. The incremental CE of the new brand name drug is quite a bit less attractive ($90,000
per year of life gained).
4000
If the CE is calculated for each drug against "No treatment", the new drug looks much more reasonable
($7,692 per year of life gained). But this is misleading, since Drug C doesn't really add much vs. Drug B or A.
It must be compared against them if they are viable choices.
2000
0
19.5
20
20.5
Years of life
21
21.5
Drugs
Cost
Years of Life
Absolute Difference Absolute Difference
$0
-20
-$100
$100
20.4
0.4
$1,000
$900
21.2
0.8
$10,000
$9,000
21.3
0.1
No treatment
Drug A - old generic
Drug B - old brand name
Drug C - new brand name
Cost per Year of Life Gained
Incremental
Versus No Rx
(correct)
(misleading)
$250
$250
$1,125
$833
$90,000
$7,692
In this Workbook, there are 4 examples of the imporance of incremental CE analyses. The first two sheets are
examples of non-embedded treatment options, the third is another embedded example, and the fourth is a sequential
algorithm.
12000
10000
In the example in this sheet, each more expensive drug adds some years of life compared with less expensive
alternatives. What drug would you use based on cost-effectiveness?
8000
Cost
Assume that all drugs will work at the level indicated, i.e., that patients have not failed treatment with any of the drugs.
(A sequential drug algorithm is considered later.)
12000
6000
The incremental cost-effectiveness of the generic drug is very attractive, and the incremental CE of the old brand name
drug is also attractive. The incremental CE of the new brand name drug is quite a bit less attractive ($90,000
per year of life gained).
4000
10000 If the CE is calculated for each drug against "No treatment", the new drug looks much more reasonable
($7,692 per year2000
of life gained). But this is misleading, since Drug C doesn't really add much vs. Drug B or A.
8000
6000
4000
It must be compared against them if they are viable choices.
0
19.5
20
20.5
Years of life
21
21.5
Sensitivity analysis: the last
step
Sensitivity analysis (1-way),
aneurysm management CUA
$ / QALY, clipping
$400,000
$300,000
$200,000
Base case
estimate
$100,000
$0
0
0.005
0.01
0.015
0.02
Annual risk of rupture
0.025
Putting it all together
Putting it all together
Putting it all together
What does CEA say about
value of life?
• A cost-effectiveness threshold is one way to use CEA to
determine which interventions represent good value.
• In the US and OECD countries, that threshold is
somewhere between $50,000-$100,000/QALY.
• What is the threshold in other countries?
– Related to per-capita GDP as a proxy for income
– Less that 1 x pcGDP: very good value
– 1-3 x pcGDP: acceptable
CEA can be misused
 Defend policies deemed unacceptable for other
reasons (depriving of rights, unfair, cruel, etc)
 Methods correct, interpretation skewed
 Methods incorrect or strategies not considered