Transcript Cost-Effectiveness of Testing High-Risk Women for Mutations in the

Cost-Effectiveness of Testing High-Risk Women for Mutations in the BRCA 1 and BRCA 2 Genes:
Does Early Detection Lead to Better Health and Economic Outcomes?
Michelle Sonia Wilkinson, MSPH
Department of Health Policy and Management
University of North Carolina Gillings School of Global Public Health
Probability Parameters
Abstract
Background: Breast cancer is the leading cause of cancer deaths for women aged 40-59. A woman’s lifetime
risk for breast cancer in America is 12%, but women with BRCA mutations are at substantially higher risk.
Objective: To analyze the cost-effectiveness of screening high-risk women for mutations in the BRCA genes.
Design: Cost-effectiveness study using Microsoft Excel to model a decision tree and Oracle Crystal Ball to
run Monte Carlo simulations to test for sensitivity. Costs and outcomes discounted at 3%. Data Sources:
American Cancer Society, National Cancer Institute, peer-reviewed cost-effectiveness studies. Target
Population: Hypothetical cohort of 50,000 high-risk 30-year-old women with family history of breast cancer.
Time Horizon: Cohort followed for duration of natural lifespan or until breast cancer death. Perspective:
Societal Interventions: Cohort of high-risk women screened for BRCA gene; those with the gene proceed to
have one of five interventions – surveillance, mastectomy, oophorectomy, tamoxifen chemopreventive therapy
or both mastectomy and oophorectomy. Outcome Measures: Incremental cost, life-years saved, incremental
quality adjusted life years (QALY), deaths averted, cost per life-year saved, cost per QALY. All costs
measured in 2008 US$. Results of Base-Case Analysis: The Incremental Cost-Effectiveness Ratio (ICER)
is $3,230 per QALY. Base-case is cost-effective at the threshold value of $50,000 per QALY. Results of
Sensitivity Analysis: Probabilities for developing breast cancer with and without BRCA mutation have most
effect on one-way sensitivity. Probabilistic sensitivity shows cost/QALY to be cost-effective 76% of the time.
Limitations: Relapse is not considered in the model. Branches in the decision-tree with three or more options
were not tested for sensitivity. Conclusions: Genetic screening for high-risk women should be undertaken,
provided that the test results are confidential.
Background
• Incidence Rate 117.7 per 100,000 for all races.
• Mortality Rate 24.4 per 100,000 for all races.
• 5-10% of cases are caused by hereditary mutations in BRCA gene.
• About 250,000 American women have mutated BRCA genes.
Pr(Event)
Having BRCA gene, given a family history
Average woman's lifetime probability of breast cancer diagnosis
Developing breast cancer, given the presence of BRCA gene
Localized stage at time of cancer diagnosis
Regional stage at time of cancer diagnosis
Distant stage at time of cancer diagnosis
5-Year Survival if diagnosed with localized disease
5-Year Survival if diagnosed with regional disease
5-Year Survival if diagnosed with distant disease
Cancer diagnosis after surveillance only
Cancer diagnosis after prophylactic mastectomy
Cancer diagnosis after oophorectomy
Cancer diagnosis after 5 years of daily tamoxifen chemo-preventive therapy
Cancer diagnosis after both mastectomy and oophoretcomy
Choosing Mastectomy as intervention
Choosing Oophorectomy as intervention
Choosing Tamoxifen as intervention
Choosing Mastectomy and Oophorectomy
Choosing Surveillance Only
Base-Case
20%
12%
80%
61%
31%
6%
98%
80%
26%
62%
10%
50%
53%
22%
11%
14%
5%
14%
57%
Decision Tree
Min
15%
11%
40%
Max
87%
13%
90%
88%
72%
23%
60%
3%
40%
40%
20%
Source
Sprecher Institute for Cancer Research
American Cancer Society
National Cancer Institute
SEER 17 Incidence (2000 - 2005)
SEER 17 Incidence (2000 - 2005)
SEER 17 Incidence (2000 - 2005)
Foley and Lordell, 2005
Foley and Lordell, 2005
Foley and Lordell, 2005
Anderson et al. 2006
Schrag et al, 2000
Schrag et al. 2000; Grann et al. 1998
Schrag et al. 2000
Anderson et al. 2006
Uyei et al. 2006
Uyei et al. 2006
Uyei et al. 2006
Uyei et al. 2006
Uyei et al. 2006
107%
88%
28%
76%
50%
100%
75%
24%
Survive
Localized**
Survive
Cancer ‡
Base-Case
$2,814
$7,965
$666
$13,305
$1,477
$20,755
$550
$1,271
$1,021
$745
$31,091
$7,986
$19,101
$45,829
$100
Min
$1,021
$4,691
$606
$11,649
$915
$16,340
$501
$733
$919
$331
$27,982
$7,187
$18,338
$41,246
$50
Outcomes (in QALYs)
Perfect Health
Oophorectomy
Mastectomy
Mastectomy & Oophorectomy
Tamoxifen Chemoprevention
Well with Positive BRCA test
Breast Cancer
Metatastic Disease
Death
Decrement for Surgery
• Mutated BRCA genes increase chance of developing breast cancer between 40-90%.
• 2% of U.S. women are “high risk.”
Base-Case
0.90
0.82
0.76
0.73
0.81
0.76
0.77
0.59
0.00
0.25
Max
$4,239
$10,264
$1,410
$25,826
$2,040
$36,090
$1,212
$1,809
$1,123
$3,973
$34,200
$8,784
$19,866
$50,412
$150
Survive
Mastectomy
0.27
No Cancer ‡
BRCA †
Oophorectomy
No BRCA†
No Test*
 A first-degree relative diagnosed with breast cancer in both breasts or a male relative
with breast cancer.
LY†
QALYs‡
No Genetic Test
$490,972,882
48,374
2,547,352
1,012,283
Genetic Test
$969,216,313
48,769
2,554,074
1,160,332
Decision Node
Chance Node
Terminal Node
Sub-tree A
Sub-tree A
* DNA test to determine presence of BRCA gene
† BRCA is the gene mutation known to cause heriditery cancer
‡ Female Invasive Breast Cancer
** As definied by the National Cancer Institute
• The parameters were derived from a literature review.
• Data comes from the National Cancer Institute SEER Project, previous peer-reviewed
journal articles on BRCA testing, the American Cancer Society and the Tufts-New England
Medical Center Cost-Effectiveness Analysis Registry.
Costs*
DA**
$478,243,431
395
QALYs‡
6,722
148,049
DA**
QALY‡
LY Saved†
$1,209,408
$3,230
$71,142
Cost-Effectiveness Acceptability Curve:
Probability that Genetic Testing for BRCA Mutations is
Cost-Effective
$3,000,000,000
$50,000/QALY
$2,500,000,000
$1,500,000,000
Incremental Costs
$1,000,000,000
$500,000,000
$0
• Societal perspective with lifelong time horizon.
$50,000/QALY
• Costs are all presented in 2008 US$ and have been inflated using the Consumer Price
Index (CPI) Medical Care Component.
-550,000
-350,000
-150,000
50,000
-$500,000,000
250,000
450,000
650,000
-$1,000,000,000
-$1,500,000,000
Incremental QALYs
Sensitivity Analysis Methods
• Microsoft Excel 2007 used to model base case; Oracle Crystal Ball v. 11.1 used for
Monte Carlo simulation (1,000 iterations) and one-way sensitivity analysis.
• Under the base-case scenario, the cost per death averted falls below the range used by
the federal government in estimating the value of a statistical life.
• 85.3% of the points fall in the northeast quadrant. In this
quadrant, costs are higher and outcomes are improved.
• High-risk women may develop breast cancer before the median age (61 years) of
diagnosis in general population.
• Costs to treat localized and regional disease assumed to be the same.
• Relapse is not part of the model.
$2,000,000,000
-750,000
• The study cohort of hypothetical women had an incremental increase of nearly 150,000
QALYs.
Limitations
Incremental Cost-Effectiveness Ratio Plane: Cost per QALY
• Costs and QALYs discounted at 3%. Study uses gross-costing.
• Genetic testing for BRCA mutations in high-risk women is cost-effective under basecase assumptions. High-risk 30-year-old women should be tested for BRCA mutations.
• Based on the sensitivity analysis, the genetic testing would be cost-effective at the
threshold of $50,000 per QALY 76% of the time.
Sensitivity Analysis Results
• Cost-effectiveness study with decision tree to show linear progression of disease.
• Branches with 3 or more options were held constant.
LY Saved†
• Sensitivity analysis was conducted on all three outcomes; results for cost per QALY are shown because there is no standard thresholds for cost per death averted or life-year saved.
Methods
• Costs parameterized as triangular distributions. Event probabilities and outcomes
(QALYs) parameterized as beta distributions.
Ratios (Cost per*)
• Genetic testing resulted in an incremental increase of 148,049 quality-adjusted life years and saved 6,722 life years.
Genetic testing for BRCA mutations will be cost-effective for high-risk women. Genetic
testing will lead to an increase in quality-adjusted life years, more deaths averted and
more life-years saved.
Sub-tree A
• The study consists of a hypothetical cohort of 50,000 high-risk 30-year-old women with a
family history of breast cancer.
• Using the Gold et al. (1996) standard of $50,000 per QALY, the base case scenario is cost-effective.
Hypothesis
Surveillance
• Each chance node has an associated probability of occurrence as well as a cost and
outcome parameter.
Notes:
* Cost in 2008 US Dollars
† Life Years gained
‡ Quality-adjusted life years
**Deaths Averted
†† Outcom es and costs discounted to 3%.
 Women of Ashkenazi (Eastern European) Jewish descent.
Sub-tree A
Conclusions
Probability that BRCA Testing is CostEffective
 Both breast and ovarian cancer among first- and second-degree relatives or two or
more first- or second-degree relatives with ovarian cancer.
Survivors
Tamoxifen
Figure Lengend:
Increm ental
Total Costs*
Sub-tree A
Test*
Source
Grann et al. 2000; Anderson et al. 2006
Grann et al. 1998; Grann et al. 2000; Anderson et al. 2006
Grann et al. 2000; Anderson et al. 2006; Norum et al. 1997
Grann et al. 1998; Grann et al. 2000; Anderson et al. 2006
Grann et al. 2000; Anderson et al. 2006
Grann et al. 2000; Anderson et al. 2006
Grann et al. 1998; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1998; Grann et al. 2000
Grann et al. 2000; Anderson et al. 2006
Author Assumption
Determ inistic
 Two first-degree relatives (mother, sisters, daughters) with breast cancer or three or
more first- or second-degree relatives (grandmothers, aunts, cousins) diagnosed with
breast cancer.
Sub-tree A
Oophoretcomy and Mastectomy
Base-Case Results
Who is High Risk?
Distant**
Die
Source
Grann et al. 1999; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1999; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1999; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1999; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1999; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1999; Grann et al. 2000; Anderson et al. 2006
Grann et al. 1998; Grann et al. 1999; Grann et al. 2000
Grann et al. 2000; Anderson et al. 2006
Grann et al. 2000
Grann et al. 2000; Anderson et al. 2006
Anderson et al. 2006
Anderson et al. 2006
Grann et al. 1998; Grann et al. 1999; Grann et al. 2000
Anderson et al. 2006
National Cancer Institute
Standard Deviation
0.05
0.27
0.26
0.25
0.25
0.29
0.22
0.27
Regional**
Die
Cost and Outcome Parameters
Costs (in 2008 US$)
Surveillance
Oophorectomy
Oophorectomy Follow-up
Mastectomy
Mastectomy Follow-up
Oophorectomy and Mastectomy
Oophorectomy and Mastectomy Follow-up
Tamoxifen Chemoprevention
Tamoxifen Follow-up
Genetic test from Myriad
Cancer Treatment, 1st year after Diagnosis
Cancer Treatment, 2nd year onward
Treating Metatastic Disease
End of life care
Mammogram
Die
Sub-tree A
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
• Branches in the decision tree with three or more options were not subject to sensitivity
analysis.
0.76
• It was assumed that high-risk women with unknown status would have the same
decremented utility value as women who knew they had a BRCA mutation.
$50,000
Selected References
• Anderson K, Jacobson J, Heitjan D, Zivin JG, Hershman D, Neugut A and Grann V. Cost-effectiveness of preventive strategies for women with a
BRCA1 or BRCA2 mutation. Annals of Internal Medicine 2006; 164(6): 397-407.
$0
$25,000
$50,000
$75,000
$100,000
Threshold Willingness to Pay ($/QALY Gained)
•One-way sensitivity analysis changing one input at a time.
• 14.5% of the points fall in the northwest quadrant. These
are dominated; costs increase but outcomes are worse.
• Probabilistic sensitivity analysis conducted using 1,000 Monte
Carlo simulations.
•The variable with the biggest effect on cost per QALY is the
utility of being well with a BRCA+ test result.
• 0.2% of the points fall in the southeast quadrant. This is
cost-saving, as costs decrease and QALYs increase.
• At a threshold level of $50,000 per QALY, the probability that
genetic testing for BRCA mutations is cost-effective is 0.76.
•Grann V, Jacobson J, Whang W, Hershman D, Heitjan D, Antman K et. Al. Prevention with tamoxifen or other hormones versus prophylactic surgery
in BRCA1/2 positive women: a decision analysis. Cancer J Sci Am 2000;(6): 13-20.
• Grann V, Panageas K, Whang W, Antman K and Neugut A. Decision analysis of prophylactic mastectomy and oophorectomy in BRCA1-positive or
BRCA2-positive patients. Journal of Clinical Oncology 1998; 16(3): 979-985.
• SEER Fact Sheet on Breast Cancer. Surveillance, Epidemiology and End Results (SEER) Program. National Cancer Institute. [Online]
http://seer.cancer.gov/statfacts/html/breast.html Accessed 10/23/08.
• Sprecher Institute for Cancer Research. Cornell University. Fact Sheet #48: Family History, Inheritance and Breast Cancer Risk. [Online]
http://envirocancer.cornell.edu/FactSheet/General/fs48.inheritance.cfm Accessed 9/25/08.
Acknowledgements
Thank you to my academic advisor, Andrea Biddle, PhD, Associate Professor in the Department of Health Policy and Management, UNC Gillings School of
Global Public Health, for her guidance in completing this research project. Special thanks to the Theta Chapter of the Delta Omega Society and the Office of
Research at the UNC Gillings School of Global Public Health for their support and assistance.