FIGURE LEGENDS
Figure 5a: Observed (dots) and model-fitted (line) breast cancer incidence.
Figure 6a: Relative stage-specific mortality from breast cancer and fitted negative exponential functions with 95% confidence limits.
Figure 7a: The distributions of 10,000 random sets of QALY weights (Beta distributions) used in the model.
Figure 8a: The distributions of 10,000 random sets of costs (log-normal distributions) used in the model
(Note the different scale for the costs of chemotherapy)
Figure 9a: The distributions of 10,000 random sets of sojourn times (log-normal distributions) and probabilities (Beta distributions) used in the model.
Figure 10a: Results of probabilistic sensitivity analysis (small dots) for screening policies 4, 16 and 33 on the cost-effectiveness plane with results of base case analysis (large dots) incrementally to no-screening.
Figure 11a: Graphical presentation of the breast cancer cumulative mortality and the portion of women who were diagnosed as false positive during their life-course in populations that are screened from 40 to 80 years of age by 1, 2 or 3 year intervals and in the population that is not screened. The cumulative mortality for the 2-year interval policy is lower than cumulative mortality of 3 year interval policy and higher than cumulative mortality of 1 year interval policy. The percentage of women refers to the total number of women alive at the age of 40.
Figure 5a: Observed (dots) and model-fitted (line) breast cancer incidence.
Figure 6a: Relative stage-specific mortality from breast cancer and fitted negative exponential functions with 95% confidence limits.
Figure 7a: The distributions of 10,000 random sets of QALY weights (Beta distributions) used in the model.
Figure 8a: The distributions of 10,000 random sets of costs (log-normal distributions) used in the model
(Note the different scale for the costs of chemotherapy)
Figure 9a: The distributions of 10,000 random sets of sojourn times (log-normal distributions) and probabilities (Beta distributions) used in the model.
Figure 10a: Results of probabilistic sensitivity analysis (small dots) for screening policies 4, 16 and 33 on the cost-effectiveness plane with results of base case analysis (large dots) incrementally to no-screening.
Figure 11a: Graphical presentation of the breast cancer cumulative mortality and the portion of women who were diagnosed as false positive during their life-course in populations that are screened from 40 to 80 years of age by 1, 2 or 3 year intervals and in the population that is not screened. The cumulative mortality for the 2-year interval policy is lower than cumulative mortality of 3 year interval policy and higher than cumulative mortality of 1 year interval policy. The percentage of women refers to the total number of women alive at the age of 40.
TABLES
Table 2a: Treatment distribution by breast cancer stage [1] indicates the portion of women in each cancer stage that were treated with specific intervention.
Chemotherapy / Hormonal therapy / Surgery / RadiotherapyLocal stage / 29% / 43% / 88% / 39%
Regional stage / 66% / 57% / 89% / 52%
Distant stage / 58% / 61% / 20% / 60%
Table 3a: Costs and QALYs used in each Markov State of the model.
Markov State / QALY / CostsNo breast cancer / 1 / None**
Preclinical screen detectable breast cancer / -DCIS / 1 / None**
-Local stage
-Regional stage
-Distant stage
Clinically detected breast cancer / -Local stage / According to treatment / Costs of diagnostics for clinically detected cancers and costs of treatment
-Regional stage
-Distant stage
Screen detected breast cancer / -DCIS / According to treatment / Costs of mammography examination and costs of treatment
-Local stage
-Regional stage
-Distant stage
False positive / QALY (Diagnostic phase) / Costs of mammography examination and costs of invasive diagnostics
Death / from breast cancer / QALY (Terminal illness)* / None
from other causes / 0
* The QALY (Terminal illness) was used in the month before the death.
**If the women are screened (according to screening policy and attendance), the costs of mammography examination are included. If the women are recalled for non-invasive diagnostics, the cost of non-invasive diagnostics is also included.
Table 4a: Baseline incremental costs and effects (in terms of LYS and QALY) relative to no-screening, which has cost of €231 and effects of 23.0 QALYs or 23.1 LYS and efficiency frontier.
From year / To
year / By years / LYS / QALY / Cost (€) / QALY / Cost (€) / ICER (€/QALY)
33 / 50 / 65 / 3 / 0.0403 / 0.0359 / 172.8 / Efficiency frontier / 0.0359 / 172.8 / 4,813
29 / 45 / 65 / 3 / 0.0518 / 0.0465 / 230.6 / 0.0106 / 57.8 / 5,457
30 / 45 / 70 / 3 / 0.0583 / 0.0521 / 268.2 / 0.0056 / 37.6 / 6,751
26 / 40 / 70 / 3 / 0.0701 / 0.0626 / 358.5 / 0.0105 / 90.3 / 8,568
27 / 40 / 75 / 3 / 0.0718 / 0.0640 / 372.7 / 0.0014 / 14.2 / 9,872
28 / 40 / 80 / 3 / 0.0737 / 0.0654 / 394.3 / 0.0014 / 21.6 / 15,516
16 / 40 / 80 / 2 / 0.0797 / 0.0697 / 585.0 / 0.0042 / 190.7 / 45,101
1 / 40 / 65 / 1 / 0.0745 / 0.0587 / 929.6 / Dominated policies
2 / 40 / 70 / 1 / 0.0769 / 0.0626 / 1,020.7
3 / 40 / 75 / 1 / 0.0809 / 0.0646 / 1,090.3
4 / 40 / 80 / 1 / 0.0812 / 0.0654 / 1,140.1
5 / 45 / 65 / 1 / 0.0595 / 0.0483 / 684.0
6 / 45 / 70 / 1 / 0.0663 / 0.0522 / 776.2
7 / 45 / 75 / 1 / 0.0676 / 0.0542 / 846.4
8 / 45 / 80 / 1 / 0.0694 / 0.0549 / 896.8
9 / 50 / 65 / 1 / 0.0458 / 0.0350 / 474.9
10 / 50 / 70 / 1 / 0.0485 / 0.0388 / 568.0
11 / 50 / 75 / 1 / 0.0524 / 0.0409 / 639.1
12 / 50 / 80 / 1 / 0.0528 / 0.0416 / 689.9
13 / 40 / 65 / 2 / 0.0691 / 0.0611 / 470.3
14 / 40 / 70 / 2 / 0.0756 / 0.0665 / 525.6
15 / 40 / 75 / 2 / 0.0780 / 0.0684 / 554.0
17 / 45 / 65 / 2 / 0.0579 / 0.0511 / 354.9
18 / 45 / 70 / 2 / 0.0623 / 0.0547 / 391.8
19 / 45 / 75 / 2 / 0.0659 / 0.0576 / 434.9
20 / 45 / 80 / 2 / 0.0671 / 0.0584 / 455.6
21 / 50 / 65 / 2 / 0.0412 / 0.0362 / 238.2
22 / 50 / 70 / 2 / 0.0477 / 0.0415 / 294.2
23 / 50 / 75 / 2 / 0.0501 / 0.0435 / 322.9
24 / 50 / 80 / 2 / 0.0518 / 0.0447 / 354.3
25 / 40 / 65 / 3 / 0.0642 / 0.0576 / 322.7
31 / 45 / 75 / 3 / 0.0617 / 0.0547 / 296.0
32 / 45 / 80 / 3 / 0.0625 / 0.0553 / 306.4
34 / 50 / 70 / 3 / 0.0434 / 0.0386 / 191.2
35 / 50 / 75 / 3 / 0.0471 / 0.0416 / 220.7
36 / 50 / 80 / 3 / 0.0487 / 0.0428 / 240.9
RESULTS OF THE UNIVARIATE SENSITIVITY ANALYSIS
Table 5a: Structure of the efficiency frontiers in 82 different univariate sensitivity analyses
Screening policy / Frequency on the efficiency frontiers / Cases in which the screening policy is on the efficiency frontier33 / 81 / In all cases except when sensitivity for 50-59 years is at lower limit
29 / 81 / In all cases except when sensitivity for 40-49 years is at lower limit
30 / 80 / In all cases except when discounting is 1%, and when sensitivity for 60-69 years is at lower limit
25 / 3 / Only when discounting is 1%, when sensitivity for 40-49 years is at upper limit and when Portion of DCIS that progress to preclinical Local stage is at lower limit
26 / 82 / In all cases
27 / 82 / In all cases
15 / 1 / Only when the sojourn time in local stage is at lower limit value
28 / 81 / In all cases except when the sojourn time in local stage is at lower limit value
16 / 81 / In all cases except when the sojourn time in local stage is at upper limit value
Table 6a: Average impact of input parameters at their lower/upper range values on the ICER values for screening policies 33, 29, 30, 26, 27, 28 and 16. Note that for parameters marked with *, the lower limits of parameters increased the ICER and the upper limits decreased the ICER.
Discounting / 1% / 5% / 68% / 147%
Portion of DCIS that progresses to preclinical Local stage* / 90.4% / 33.4% / 81% / 141%
Recall rate / 1.4% / 16.6% / 79% / 131%
Relative mortality in regional stage (parameter B)* / 0.162 / 0.094 / 86% / 127%
Relative mortality in regional stage (parameter A)* / 0.862 / 0.592 / 84% / 126%
Percent of invasive diagnostics at recall / 4.7% / 42.8% / 87% / 117%
Cost of mammography exam / €44 / €66 / 84% / 116%
Portion of invasive cancers. that are not preceded by DCIS / 13.3% / 70.6% / 96% / 115%
Portion of clinically detected cancers in local stage / 38.1% / 44.0% / 90% / 113%
Sojourn time in DCIS stage / 3 / 7 / 89% / 113%
Cancer incidence* / +10% / -10% / 90% / 112%
Cost of invasive diagnostics / €320 / €1560 / 88% / 112%
Relative mortality in local stage (parameter A) / 0.221 / 0.293 / 93% / 109%
Portion of clinically detected cancers in regional stage / 44.1% / 49.9% / 93% / 108%
Relative mortality in local stage (parameter B) / 0.125 / 0.193 / 93% / 107%
Sojourn time in regional stage / 0.36 / 1.08 / 95% / 105%
Cost of non-invasive diagnostics / €47 / €154 / 97% / 105%
Sensitivity above 70 years* / 98.3% / 79.2% / 98% / 105%
QALY (Diagnostic phase)* / 0.997 / 0.653 / 98% / 104%
Cost of chemotherapy* / €13500 / €605 / 84% / 104%
Sojourn time in distant stage / 0.35 / 1.04 / 98% / 102%
Sensitivity for 50-59 years of age* / 99.5% / 80.9% / 99% / 102%
QALY (Hormonal therapy) / 0.472 / 0.991 / 96% / 102%
Sensitivity for 40-49 years* / 94.8% / 75.5% / 98% / 102%
Attendance / 60.1% / 87.3% / 99% / 102%
Cost of diagnostics for clinically detected cancers* / €957 / €55 / 96% / 102%
Sensitivity for 60-69 years of age* / 99.7% / 81.1% / 99% / 101%
QALY (Chemotherapy) / 0.309 / 0.975 / 98% / 101%
Cost of radiotherapy* / €4630 / €2650 / 99% / 101%
Cost of surgery / €2320 / €2900 / 99% / 101%
QALY (Distant stage breast cancer ) / 0.410 / 0.620 / 99% / 101%
QALY (Terminal illness) / 0.004 / 0.833 / 100% / 101%
Cost of hormonal therapy* / €7180 / €395 / 93% / 100%
QALY (Radiotherapy) / 0.486 / 0.981 / 100% / 100%
QALY (2m-1year after surgery) / 0.518 / 0.994 / 100% / 100%
Portion of clinically detected cancers in distant stage / 8.1% / 16.1% / 100% / 100%
QALY (Surgery) / 0.578 / 0.995 / 100% / 100%
QALY (Disease free>1year after operation) / 0.777 / 0.999 / 100% / 100%
Relative mortality in distant stage (parameter B) / 0.680 / 0.762 / 100% / 100%
Relative mortality in distant stage (parameter A) / 0.814 / 0.838 / 100% / 100%
Sojourn time in local stage / 3 / 2 / 98% / 96%
SECTION A:
POSSIBLE LIFE-COURSES OF WOMEN IN POPULATION WITHOUT SCREENING AND IN POPULATIONS WITH SCREENING
Population without screening
Most common life course of women in population without screening is healthy life until death. Some women develop a breast cancer, according to breast cancer incidence. A portion of invasive breast cancers is preceded by DCIS, which is non-invasive and can spontaneously regress. The remaining breast cancers start as invasive from the beginning. When the local stage invasive breast cancer has developed, it may progress to regional stage breast cancer or it may be clinically detected and treated. Likewise, the regional stage breast cancer can progress to distant stage breast cancer or it may be clinically detected and treated. Women with distant stage breast cancer are then also diagnosed clinically and treated accordingly.
Death from breast cancer can occur only when women are diagnosed with breast cancer; that is when they are in clinical local, regional or distant stage. At any given time of their life course, women can die because of other causes than breast cancer.
Population with screening
The life courses of women in population that is screened is related to life course of population without screening, except that the DCIS and the preclinical stages of invasive breast cancer can be detected with screening, thus diagnosing the women with breast cancer earlier in their lifetime. New life course is also the path of false positive women, which had positive screening examination but no cancer is found at further invasive diagnostic assessment. Even though that the majority of breast cancers are diagnosed with screening, some breast cancers are also diagnosed clinically. The breast cancers are diagnosed clinically in women who develop breast cancer and do not attend the screening, when they have false negative results at their screening or when the period between screening intervals is long enough to permit the development of the breast cancer. The treatment and the survival of the clinically detected and screen detected breast cancers is the same.
MODEL PARAMETERS AND TRANSITIONS