Treatment Benefit and Treatment Harm Rates with Nonignorable Missing Covariate, Endpoint, or Treatment

Round 1

Reviewer 1 Report

I appreciate the authors' effort in presenting this paper and thank the editor for the chance to review it. The paper focuses on a promising and practically significant question. However, I believe there is a pressing need for the paper to be carefully proofread and refined for clarity and readability before considering re-submission.

A few areas of concern were evident from the outset. For instance, in the abstract, certain grammatical errors and unclear phrasings are present. The central sentence that articulates the paper's objective states: "In this article, we address the problem of identifying treatment benefit rate and treatment rate when treatment or endpoints or covariates is missing." It seems the term "harm" in "treatment harm rate" was unintentionally omitted. As another example, there was a super long sentence at the bottom of page 2 describing the trial which was difficult to comprehend. 

While the authors touch upon the limitation regarding the covariate $X$ being constrained as a binary variable in the discussion section, I feel this constraint significantly impacts the broader significance and applicability of the proposed methodology.

The main body of the manuscript could benefit from a more concise structure. For example, the definitions of MAR and MNAR might either be condensed or relocated to supplemental materials, as they are widely recognized and understood in the field.

Upon acceptance of the paper by a journal, I recommend that the authors share their computational code or software on a public repository. This would be beneficial for researchers aiming to utilize or reproduce the work.

 Extensive editing of English language required

Author Response

Manuscript ID Mathematics-2619220:

“Treatment Benefit and Treatment Harm Rates with Nonignorable Missing Covariate, endpoint or treatment”

By Yi He, Lin Zhi Zheng and Peng Luo

 

Reply to Reviewer 1:

We would like to thank you very much for your helpful suggestions and valuable comments on our paper which greatly improved our paper. According to your suggestions, we have revised the paper. Below we give a detailed reply to your comments. In the following responses, your comments are copied in Italic.

 

For your general comments

The paper focuses on a promising and practically significant question. However, I believe there is a pressing need for the paper to be carefully proofread and refined for clarity and readability before considering re-submission. 

 

Thank you very much for your positive recommendation.

 

For your specific comments

1. The central sentence that articulates the paper's objective states: "In this article, we address the problem of identifying treatment benefit rate and treatment rate when treatment or endpoints or covariates is missing." It seems the term "harm" in "treatment harm rate" was unintentionally omitted.

Re: We have added this word. 

 

2. As another example, there was a super long sentence at the bottom of page 2 describing the trial which was difficult to comprehend.

Re: This sentence is indeed too long, I have deleted it.

 

3. While the authors touch upon the limitation regarding the covariate $X$ being constrained as a binary variable in the discussion section, I feel this constraint significantly impacts the broader significance and applicability of the proposed methodology.

Re: For the sake of mathematical simplification, the case where X is a vector is not discussed here. When X is a vector, I feel that there is no essential difference when the outcome variable and the treatment variable has missing data. When the covariate is missing, it may be mathematically more complex.

 

4. For example,the definitions of MAR and MNAR might either be condensed or relocated to supplemental materials, as they are widely recognized and understood in the field.

Re: We referred to reference [3], in which he also detailed five missing mechanisms and discussed them one by one.

 

We hope that the revision of our paper has been revised by following your comments and suggestions. Thank you very much for your valuable suggestions and comments which are very helpful and greatly improved our paper.

Author Response File: Author Response.doc

Reviewer 2 Report

File in attachment.

Comments for author File: Comments.pdf

Author Response

Manuscript ID Mathematics-2619220:

“Treatment Benefit and Treatment Harm Rates with Nonignorable Missing Covariate, endpoint or treatment”

By Yi He, Lin Zhi Zheng and Peng Luo

 

Reply to Reviewer 2:

We would like to thank you very much for your helpful suggestions and valuable comments on our paper which greatly improved our paper. According to your suggestions, we have revised the paper. Below we give a detailed reply to your comments. In the following responses, your comments are copied in Italic.

 

For your general comments

I found the manuscript well written and could not find any flaws in the mathematical sections presented. The proposal is interesting and relevant...The findings should be of interest to data analysts. 

 

Thank you very much for your positive recommendation.

 

For your specific comments

1. Did the authors implement the methods using software R?

 

Re: Yes, we have revised the article. We mentioned it in the last sentence of the first paragraph of section 5. 

 

2. Is there a new package?

 

Re: Writing a new package is a great suggestion, but unfortunately, my students are all master's students and they have all graduated, so we don't have time to write. 

 

We hope that the revision of our paper has been revised by following your comments and suggestions. Thank you very much for your valuable suggestions and comments which are very helpful and greatly improved our paper.

 

Author Response File: Author Response.doc

Reviewer 3 Report

This paper is well-written and interesting. I have some comments that should be clarified before considering for publication.

1- What is the advantage and disadvantage of the proposed procedure?  

2- The authors should express when the proposed procedure does not work well.

3- Commenting based on the obtained result should be analyzed with more caution.

4- Why the authors did not compare the proposed method with the other existing methods?

Finally, the paper can be published after a revision.

Author Response

Manuscript ID Mathematics-2619220:

“Treatment Benefit and Treatment Harm Rates with Nonignorable Missing Covariate, endpoint or treatment”

By Yi He, Lin Zhi Zheng and Peng Luo

 

Reply to Reviewer 3:

We would like to thank you very much for your helpful suggestions and valuable comments on our paper which greatly improved our paper. According to your suggestions, we have revised the paper. Below we give a detailed reply to your comments. In the following responses, your comments are copied in Italic.

 

For your general comments

This paper is well-written and interesting.  

 

Thank you very much for your positive recommendation.

 

For your specific comments

1. What is the advantage and disadvantage of the proposed procedure? 

Re: The advantage is that we give sufficient conditions to make TBR and THR identifiable in the presence of missing data. In addition, we illustrate our method through simulated data, and then apply our method to several actual data. The disadvantage of our article is that assumption 2 and MNAR are too strong together.

 

2. The authors should express when the proposed procedure does not work well.

Re: In the last three paragraphs of section 7, we mentioned the shortcomings of this article. When one of assumption 2 and MNAR does not hold, the proposed procedure does not work well.

 

3. Commenting based on the obtained result should be analyzed with more caution.

Re: This is a very important suggestion. In the article, we first validated the effectiveness of our method through numerical simulation in the 5 section. Then in the 6 section, we applied this method to actual data.

 

4. Why the authors did not compare the proposed method with the other existing methods?

Re: This is a very good question. In this article, our main contribution is to provide sufficient conditions for TBR and THR to be identified in the presence of missing data. Among the existing literature, there is no literature discussing this issue.

 

We hope that the revision of our paper has been revised by following your comments and suggestions. Thank you very much for your valuable suggestions and comments which are very helpful and greatly improved our paper.

Author Response File: Author Response.doc

Round 2

Reviewer 1 Report

The authors have proofread their manuscript, and this version is much clearer compared to the original one. In terms of methodologies, it adequately discusses its limitations in the discussion section.