Effect of Breast Screening Regimen on Breast Cancer Outcomes: A Modeling Study
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
This is an important and timely study. Based on the scientific facts, the United States Preventive Services Task Force (USPSTF), once again, supports screening starting at the age of 40 (instead of 50), but they have persisted in supporting biennial and not annual mammography. Since there are no Randomized, Controlled Trials to test screening intervals, the best that we can do is through modeling which the authors have done. The CISNET models referenced by the authors all agree that the most lives are saved by annual screening starting at the age of 40. The authors now have shown that the Canadian Partnership Against Cancer’s OncoSim model is also in agreement. This is critical information and should be published as soon as possible so that the USPSTF, which is presently evaluating its guidelines, can have access to these important data.
CORRECTIONS
Page 5 Line 177: “Total number of cancer(s) diagnosed” - “cancer” should be pleural.
Page 6 Line 252 Remove the hyphen in “inde-pendent”
Line 254 Remove the hyphen from “Inter-national”
Page 7 Line 275 Remove the hyphen from “Can-ada”
Line 281 Remove the hyphen from “carci-noma”
Author Response
- This is an important and timely study. Based on the scientific facts, the United States Preventive Services Task Force (USPSTF), once again, supports screening starting at the age of 40 (instead of 50), but they have persisted in supporting biennial and not annual mammography. Since there are no Randomized, Controlled Trials to test screening intervals, the best that we can do is through modeling which the authors have done. The CISNET models referenced by the authors all agree that the most lives are saved by annual screening starting at the age of 40. The authors now have shown that the Canadian Partnership Against Cancer’s OncoSim model is also in agreement. This is critical information and should be published as soon as possible so that the USPSTF, which is presently evaluating its guidelines, can have access to these important data. -(Thanks for these comments - Au)
- - CORRECTIONS
Page 5 Line 177: “Total number of cancer(s) diagnosed” - “cancer” should be pleural. - (done Au)
Page 6 Line 252 Remove the hyphen in “inde-pendent” - (done Au)
Line 254 Remove the hyphen from “Inter-national” (done Au)
Page 7 Line 275 Remove the hyphen from “Can-ada” (done Au)
Line 281 Remove the hyphen from “carci-noma” (done Au)
Reviewer 2 Report
Comments and Suggestions for Authors
Your discussion should be expanded to including results from studies outside of Canada.
Your findings are more promising than many other large scale trials including all those presented by the USPSTF. Please discuss.
Your results seem to require 100% participation.
Please reanalyze your results using real world participation, which is closer to 70%
Author Response
Your discussion should be expanded to including results from studies outside of Canada. (The model specifically describes breast cancer in Canada and because it applies to women who participate in screening (rather than are simply invited to be screened) it lends itself to comparison with a study like The Pan Canadian Study. But these results are in general agreement with those from other observational studies and we have now added a citation to the paper by Broeders et al. which summarizes several such studies. - Au)
Your findings are more promising than many other large scale trials including all those presented by the USPSTF. Please discuss. (Our findings are for women who actually participate in screening (rather than just invited) and therefore, as expected, are more positive than those of "intent to treat" studies. However, our results are in general agreement with those of observational studies where relatively modern screening and treatment methods were used as well as the CISNET modeling estimates, used by the USPSTF. We do point out that the CISNET data combine the results of multiple, quite different, CISNET models. The CISNET estimates actually suggest greater absolute mortality reductions that our model does. This may be partly due to different amount of follow-up used. Our model suggests higher relative mortality reductions because the baseline breast cancer mortality rate in OncoSim for unscreened women appears to be lower than for the US models (different health care systems?, insurance?) - Au).
Your results seem to require 100% participation. (yes, they explicitly describe outcomes for a screened cohort, i.e. all cohort members are screening participants as only those who are screened are subject to the benefits of the intervention - Au)
Please reanalyze your results using real world participation, which is closer to 70% (In the Discussion we have described how the reader can do this using our data. We have also given a specific example where we have used a 70% participation rate. Clearly, mortality reduction will be less under such conditions. Our approach allows the reader to separate the effect on screening effectiveness of two important factors, the efficacy of the intervention in providing reduced mortality through earlier detection and treatment of cancers and the effect of participation on effectiveness - Au.
Reviewer 3 Report
Comments and Suggestions for Authors
An interesting study on the simulation of Breast Cancer. The following minor changes should be made.
1. In the Material and Methods section, insert a link to the simulator web page, so that readers who wish to use it can request it from the organization. https://www.partnershipagainstcancer.ca/tools/oncosim/
2. The first time the acronym CISNET is used, the meaning of the abbreviation should be indicated: Cancer Intervention and Surveillance Modeling Network (CISNET).
3. In the discussion when authors talk about CISNET models authors should explain in a sentence or two what they are so that a reader unfamiliar with the subject can understand.
4 In my opinion, the titles of the tables are too long and should be shortened.
Author Response
An interesting study on the simulation of Breast Cancer. The following minor changes should be made.
- In the Material and Methods section, insert a link to the simulator web page, so that readers who wish to use it can request it from the organization. https://www.partnershipagainstcancer.ca/tools/oncosim/ (We had originally had a link toward the end of the paper, but now we have also added one here. Au)
- The first time the acronym CISNET is used, the meaning of the abbreviation should be indicated: Cancer Intervention and Surveillance Modeling Network (CISNET). (done Au)
- In the discussion when authors talk about CISNET models authors should explain in a sentence or two what they are so that a reader unfamiliar with the subject can understand. (Thanks for this suggestion. We have provided a reference and a bit of detail. It is a long and complex story, so we can't do any justice to explaining the different CISNET models, but this is done clearly in the cited reference.)
4 In my opinion, the titles of the tables are too long and should be shortened. (We have attempted to do this). We realize that they are a bit lengthy, however, in this type of work, precise understanding of the meaning of variables is essential so we have been careful to try to supply required detail.)
Round 2
Reviewer 3 Report
Comments and Suggestions for Authors
The authors have done all the changes asked.
