Correlation of Anticancer Drug Prices with Outcomes of Overall Survival and Progression-Free Survival in Clinical Trials in Japan

Round 1

Reviewer 1 Report

 

The article entitled “Correlation of anticancer drug prices with outcomes of overall 2 survival and progression free survival in clinical trials in Japan” has been evaluated. This study investigated the connection between drug prices and their clinical efficacy and 9 usefulness using public information on anticancer drugs reimbursed by the Japanese National 10 Health Insurance price listing between January 2009 and March 2020. From this study authors found a correlation between drug prices and their clinical 20 usefulness in terms of their overall survival rate but not their progression-free survival rate. The manuscript needs some major improvements; there are a few suggestions that authors may consider to improve it further:

1.      The use of English language is reasonable, however, there are a number of punctuation and grammatical errors; that should be corrected and rephrased using academic English for a better flow of text for reader. Even the objective statement in the abstract has a few errors.

2.      Introduction is very brief and lacking the background information about various aspects related to this article.

3.      Authors should add some appropriate reference relevant to this study.

(a)   Anticancer drug prices and clinical outcomes: a cross-sectional study in Italy.

Trotta F, Mayer F, Barone-Adesi F, Esposito I, Punreddy R, Da Cas R, Traversa G, Perrone F, Martini N, Gyawali B, Addis A. BMJ Open. 2019 Dec 10;9(12):e033728.

(b)   Prices and Clinical Benefit of National Price-Negotiated Anticancer Medicines in China, Pharmacoeconomics. 2022 Jul;40(7):715-724. doi: 10.1007/s40273-022-01161-7.

4.      The manuscript was based on statistical analysis. The applied statistical method was not sufficient to explain the outcome model. Authors have to apply Regression coefficients β, their levels of significance p and the coefficients of determination R2 should be reported for each model (Fig 1 and 2).

5.      Authors mentioned some limitations of this study; is it possible to overcome any of these by taking other parameters and/or different regression analysis method?

6.      In Supplemental Table 1 authors should List of oncology drugs approved from 2009 to 2020. It was mentioned from 2020 to 2009. Authors should add one more column to mention about compound type.

 

Author Response

Response to Editors’ and Referees’ comments

 

[Referee 1]

The article entitled “Correlation of anticancer drug prices with outcomes of overall 2 survival and progression free survival in clinical trials in Japan” has been evaluated. This study investigated the connection between drug prices and their clinical efficacy and 9 usefulness using public information on anticancer drugs reimbursed by the Japanese National 10Health Insurance price listing between January 2009 and March2020. From this study authors found a correlation between drug prices and their clinical 20 usefulness in terms of their overall survival rate but not their progression-free survival rate. The manuscript needs some major improvements; there are a few suggestions that authors may consider to improve it further:

 

[Response]

Thank you for taking the time to review our manuscript. We have made major revisions to our manuscript, based on your suggestions and comments. I am convinced that our manuscript has become more logical and scientific. Thank you very much. Below, I respond to each of your comments.

 

1. The use of English language is reasonable, however, there are a number of punctuation and grammatical errors; that should be corrected and rephrased using academic English for a better flow of text for reader. Even the objective statement in the abstract has a few errors.

 

[Response]

Thank you for pointing this out. In response to your suggestion, we have reviewed the English language use in the paper again. We made the necessary changes with the help of Editage (a company that provides English editing services https://www.editage.com/). I added following acknowledgement in the back matter of our manuscript.

Back matter, Line 243:

Acknowledgment: We would like to thank Editage (www.editage.com) for English language editing.

 

2. Introduction is very brief and lacking the background information about various aspects related to this article.

 

[Response]

Thank you for your suggestion. We have reviewed the background information comprehensively and have added additional information to introduction section.

 

Introduction

Before revision (Lines 44–54)

Particularly, the extension of OS reflects the actual clinical usefulness for cancer patients. However, PFS extension is used as a primary endpoint in Phase-III trials. PFS is a surrogate endpoint under all circumstances, and it is not necessarily limited to reflecting the clinical usefulness for cancer patients. Furthermore, the price difference between drugs approved based on OS and those based on PFS remains unclear. Specifically, Japan prefers drug pricing based on clinical usefulness, but the details regarding this are ambiguous.

 The principal objective of this study was to clarify whether or not OS (the true endpoint) and PFS (the surrogate endpoint) affect anticancer drug prices. The differences between OS and PFS and the relationship between them and drug prices at approval in Japan were investigated.

 

After revision (Lines 45–70)

Specifically, the extension of OS reflects the actual clinical benefit to cancer patients. However, prolonged PFS is used as a primary endpoint in Phase-III trials. PFS is a surrogate endpoint under all circumstances, and it is not necessarily limited to reflecting the clinical usefulness of a drug for cancer patients. Furthermore, the price difference between drugs approved based on OS and those based on PFS remains un-clear. Satoh et al. reported that, in a study of 45 anticancer drugs approved in Japan between 2006 and 2015, there were no correlations between drug prices and PFS, or between drug prices and OS [11]. Considering China, a study involving 58 Chinese anticancer drugs also reported no correlation between clinical benefit and costs of China's price-negotiated cancer therapies [12]. Moreover, in a study of 30 anticancer drugs in Italy, the price of anticancer drugs did not reflect their clinical therapeutic effect [13]. Another study on anticancer drugs approved by the FDA between 2006 and 2015 reported that the cost of new anticancer drugs has increased rapidly during the past decade; however, the clinical benefit of those drugs has not been proportionate [14]. The results of existing studies generally indicate that there is no correlation between clinical benefit, such as OS or PFS, and drug price. As more and more new anticancer drugs are expected to be launched in the near future, more transparent and reliable pricing methods must be established. Specifically, Japan prefers drug pricing based on clinical usefulness, but the details regarding this are unclear. Therefore, this study investigated the relationship between drug prices and clinical benefits. We also decided to determine whether there is a correlation between clinical benefit, such as OS and PFS, and drug price for recent Japanese anticancer drugs.

 The principal objective of this study was to clarify whether OS (the true endpoint) and PFS (the surrogate endpoint) affect anticancer drug prices. The differences between OS and PFS and the relationship between them and drug prices at approval in Japan were investigated.

 

3. Authors should add some appropriate reference relevant to this study. (a) Anticancer drug prices and clinical outcomes: across-sectional study in Italy. Trotta F, Mayer F, Barone-Adesi F, Esposito I, Punreddy R, DaCas R, Traversa G, Perrone F, Martini N, Gyawali B, Addis A.BMJ Open. 2019 Dec 10;9(12):e033728. (b) Prices and Clinical Benefit of National Price-Negotiated Anticancer Medicines in China, Pharmacoeconomics. 2022 Jul;40(7):715-724. doi:10.1007/s40273-022-01161-7.

 

[Response]

Thank you for your extremely valuable suggestion. We added the two references you suggested, as well as other papers on the topics of correlation between drug price and clinical benefit in anticancer drugs in other countries to the references. Their text citations were added to the Introduction section and Discussion section (please refer to the response of Referee 1, #2). The addition of references also led to a revision of the numbering of subsequent references.

 

4. The manuscript was based on statistical analysis. The applied statistical method was not sufficient to explain the outcome model. Authors have to apply Regression coefficients β, their levels of significance p and the coefficients of determination R2 should be reported for each model (Fig 1 and 2).

 

[Response]

Thank you for your suggestion. We added regression coefficients β, their levels of significance (in p-values), and the coefficients of determination R2 to Fig. 1 and Fig. 2.

 

5. Authors mentioned some limitations of this study; is it possible to overcome any of these by taking other parameters and/or different regression analysis method?

 

[Response]

Thank you for your comment. The issues we mentioned in our limitations are based on the survey methodology. It is not possible to overcome these limitations by changing the analysis method (using other parameters or different regression analysis methods). Based on your suggestions, we added the following to the limitations.

 

Discussion, Line 234-235

Additional sentence

It is difficult to avoid these limitations by using analytical methods or other parameters in this study.

 

6. In Supplemental Table 1 authors should List of oncology drugs approved from 2009 to 2020. It was mentioned from 2020 to2009. Authors should add one more column to mention about compound type.

 

[Response]

Thank you for your suggestion. In Table S1 (supplemental Table 1), we sorted the compounds in order of approval date—from oldest to newest—and added a column for type of compound. Lastly, I uploaded Table S1 to the MDPI system.

 

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments to the Author

We have reviewed your manuscript entitled “Correlation of anticancer drug prices with outcomes of overall survival and progression free survival in clinical trials in Japan”. This is interesting topic, but you need to improve the paper quality.

a. You emphasized OS and PFS terminology on introduction part.  However, I didn’t see any single word of those in abstract. Sorry, readers may not familiar to these terms.

b.  What is your objective of study? In Abstract section, You write …….. the connection between drug prices and their clinical efficacy and usefulness…..(line 9-10) While in introduction, You write to clarify whether or not OS (the true end- 50 point) and PFS (the surrogate endpoint) affect anticancer drug prices ( line 50-51). I feel difficulty to follow your manuscript, as In result part you explain about Rates of increase in OS and PFS extension and incremental drug costs (line 115)

c. How do you collect the publicity available (in which year). Is it available online and can be accessed? Does your data represent national data? How do you perform statistical analysis, why you choose linear regression?

d. Please revise the graph appropriately. I think Drug price should be in Y axis. I don’t see any explanation about red line on the graph (is that regression line ?). While mentioning correlation in the title, but you write association in Figure 1 and 2. That is different term, in addition neither correlation value nor regression equation written. Adding a table legend is suggested

Please also revise Table 1 following the MDPI pattern. You need to describe well this table.

e. I cannot find what statistical technique do you conduct and how you interpret the statistical result (displaying on table 2). How you relate this table to your objective.

f. No conclusion written in manuscript. It is little bit strange. As you write there is a correlation in abstract, but not supported in result part clearly.

I think you need to rewrite accordingly and resubmit again.

Thank you

 

 

Author Response

Response to Editors’ and Referees’ comments

[Referee 2]

We have reviewed your manuscript entitled “Correlation of anticancer drug prices with outcomes of overall survival and progression free survival in clinical trials in Japan”. This is interesting topic, but you need to improve the paper quality.

 

[Response]

Thank you for taking the time to review our manuscript. We appreciate your comments and suggestions for our manuscript. Below, I respond to your comments and suggestions individually.

 

a. You emphasized OS and PFS terminology on introduction part. However, I didn’t see any single word of those in abstract. Sorry, readers may not familiar to these terms.

 

[Response]

Thank you for pointing this out. Based on your observation, we added the words overall survival (OS) and progression-free survival (PFS) to the abstract and reviewed and revised the abstract.

 

Before revision (Lines 8-21)

Japan calculates drug prices using cost accounting and comparing efficacy of similar drugs. This study investigated the connection between drug prices and their clinical efficacy and usefulness using public information on anticancer drugs reimbursed by the Japanese National Health Insurance price listing between January 2009 and March 2020. It investigated the relationship between drug prices and their clinical benefits based on the overall survival and progression-free survival rates. Eighty anticancer drugs were approved in Japan during the study period. The largest number (28 drugs, 35.0%) were approved based on progression-free survival, 18 (22.5%) were approved based on overall survival, and 13 (16.3%) based on the response rate. The mean (±SD) drug price was ¥88,416.2 ± 148,974.7, and the median drug price (with quartiles) was ¥21,694 [¥4,855.0 to ¥93,396.8]. Daily drug price increments had a significant relationship only with their associated overall survival rate. The approved 80 anticancer drugs were evaluated. The relationship between the 46 drugs approved based on their associated overall survival or progression-free survival rates and their prices were examined. A correlation was found between drug prices and their clinical usefulness in terms of their overall survival rate but not their progression-free survival rate.

 

After revision (Lines 8–21)

Drug pricing methods vary extensively across countries. Japan calculates drug prices using cost accounting and based on the efficacy of similar drugs. This study investigated the correlation between drug prices and their clinical efficacy and usefulness, using public information on anticancer drugs reimbursed by the National Health Insurance price listing, between January 2009 and March 2020. We investigated drug characteristics, prices, and clinical benefits based on overall survival (OS) and progression-free survival (PFS). Eighty anticancer drugs were approved in Japan during the study period. The largest number (28 drugs, 35.0%) were approved based on PFS, 18 (22.5%) were approved based on OS, and 13 (16.3%) based on the response rate. The mean (±SD) drug price was ¥88,416.2 (±148,974.7), while the median drug price (with quartiles) was ¥21,694 (¥4,855.0–¥93,396.8). Drug prices were significantly higher for PFS than for OS, while cost index —the drug price to extend PFS or OS by one day— did not differ significantly between PFS and OS. The relationship between the 46 drugs approved based on OS or PFS and their prices were examined. A correlation was found between drug prices and their clinical usefulness in terms of OS, but not PFS.

 

b. What is your objective of study? In Abstract section, You write…….. the connection between drug prices and their clinical efficacy and usefulness…..(line 9-10) While in introduction, You write to clarify whether or not OS (the true end- 50 point) and PFS (the surrogate endpoint) affect anticancer drug prices ( line50-51). I feel difficulty to follow your manuscript, as In result part you explain about Rates of increase in OS and PFS extension and incremental drug costs (line 115)

 

[Response]

Thank you, you made reasonable points. The main objective of this study was to examine the relationship between drug prices in terms of OS and PFS as the end points. To fulfil the objective, we investigated three items in this study: 1) how many and how often OS and PFS were adopted as the endpoints during the 11-year study period, and what were the characteristics of the approved anticancer drugs; 2) whether a difference in drug price was required when extending OS and PFS by one day; and 3) whether the rates of increase in OS and PFS correlated with the daily drug price.

The following sentences were added for clarify the objective of this study.

 

• Abstract was changed (please refer to response of Referee 2, #a).

 

• The following sentence was added to the last part of the Introduction section.

Introduction, Lines 63–66

Additional sentence:

Therefore, this study investigated the relationship between drug prices and clinical benefits. We also decided to determine whether there is a correlation between clinical benefit, such as OS and PFS, and drug price for recent Japanese anticancer drugs.

 

• We added item “3.3 Correlation Between Improvement Rate of Overall Survival and Incremental Drug Costs” to the Results section. Furthermore, Table 2 and the explanatory text were moved to section “3.2 Rates of Increase in OS and PFS Extension and Incremental Drug Costs.”

 

Before revision (Lines 123-145)

3.2. Rates of increase in OS and PFS extension and incremental drug costs

For 46 drugs, the rates of increase in OS or PFS extension and the associated incremental drug costs (daily drug prices) were compared between drugs approved on the basis of OS and PFS, and a simple calculation was performed for incremental cost effectiveness (Figures 1, 2). The daily drug price showed a significant correlation with incremental effects on OS but not with incremental effects on PFS.

Fig. 1 and Fig. 2

In addition, drug prices and CI of drugs approved by OS and PFS were compared. Results showed Table 2 and drug prices were significantly higher for PFS than for OS, but CIs were not significantly different between OS and PFS.

Table 2

 

After revision (Lines 145–162)

3.2. Rates of Increase in OS and PFS Extension and Incremental Drug Costs

To determine whether there was a difference in drug price, OS and PFS had to be extended by one day, and drug prices and CI of drugs approved for OS and PFS were compared. The results showed that the drug prices were significantly higher for PFS than for OS, while CIs were not significantly different between OS and PFS.

Table 2

3.3. Correlation Between Improvement Rate of OS and Incremental Drug Costs

For 46 drugs, the rates of increase in OS or PFS extension and the associated incremental drug costs (daily drug prices) were compared between drugs approved based on OS and PFS, and a simple calculation was performed for incremental cost effectiveness (Figures 1, 2). The daily drug price showed a significant correlation with incremental effects on OS, but not with incremental effects on PFS.

Fig. 1 and Fig. 2

 

• The following sentences were added to the Discussion section; it is similar to the last part of the Introduction section.

Discussion, Lines 194–200

Additional sentence

The objectives of this study were to examine the correlation between OS and PFS and cancer drug prices. To fulfill this objective, we incorporated three items in this study: 1) determining how many times and how often OS and PFS were adopted as the endpoints during the study period of 11 years, and what the characteristics were of the cancer drugs that were approved during these 11 years; 2) investigating whether a difference in drug price was required to extend OS and PFS by one day, and 3) deter-mine whether the rates of increase in OS and PFS correlated with the daily drug price.

 

• I revised the captions of Fig. 1 and Fig 2.

Before revision

Association between improvement rate of overall survival (progression-free survival) and incremental daily drug price.

After revision

Correlation between improvement rate of overall survival (progression-free survival) and incremental daily drug price.

 

c. How do you collect the publicity available (in which year). Is it available online and can be accessed? Does your data represent national data? How do you perform statistical analysis, why you choose linear regression?

 

[Response]

The data collection was described in the Materials and Methods section: “2.2 Data Collection.” Basically, data for every year were obtained by the method shown in references numbers 12–15. In addition, while almost all data were available online, some of the data on Japanese drug prices were not available online and, therefore, we obtained it from the book “NHI Drug Price Standards, version: October 2019.” We revised the text and added a note to “2.2 Data Collection” to that effect. All data regarding drug prices were national data of official drug prices in Japan. A statistical analysis was performed using JMP Pro15, as described in “2.4 Statistical Methods.” Since “linear regression” had already been used in a previous report [25], it was selected for comparison purposes, and therefore, other models were not used in the study.

 

2.2 Data Collection

Before revision (Lines 73-75)

…Council’s website [13], Ministry of Health, Labour and Welfare’s medical insurance website [14], NHI Drug Price Standards [15], and NHI Drug Price Standards Quick Reference Tables [16]. The pharmacological and regulatory characteristics of each anticancer drug were…

 

After revision (Lines 89–95)

…Council’s website [17], the Ministry of Health, La-bour and Welfare’s medical insurance website [18], the NHI Drug Price Standards [19], and the NHI Drug Price Standards Quick Reference Tables [20].

Virtually all data were obtained online; however, some information on drug prices were not available online and was procured from a book: the NHI Drug Price Standards [19] and the NHI Drug Price Standards Quick Reference Tables [20]. The pharmacological and regulatory characteristics of each anticancer drug were assessed according to our previous work…

 

d. Please revise the graph appropriately. I think Drug price should be in Y axis. I don’t see any explanation about red line on the graph (is that regression line ?). While mentioning correlation in the title, but you write association in Figure 1 and 2. That is different term, in addition neither correlation value nor regression equation written. Adding a table legend is suggested Please also revise Table 1 following the MDPI pattern. You need to describe well this table.

 

[Response]

I revised the Y-axis to be the drug price and the X-axis to show the improvement. The red line is the regression line (added figure captions), and “Association” has been corrected to “Correlation.” I added the correlation value and regression equation. Furthermore, Table 1 was changed to adhere to the MDPI format.

 

e. I cannot find what statistical technique do you conduct and how you interpret the statistical result (displaying on table 2). How you relate this table to your objective.

 

[Response]

Thank you for this observation; the description was difficult to understand. We changed both the Abstract and the Materials and Methods sections (2. Materials and Methods: 2.3 Data Analysis) as follows:

 

Abstract

Before revision (Lines 17–18)

Daily drug price increments had a significant relationship only with their associated overall survival rate. The approved 80 anticancer drugs were evaluated. The relationship between…

 

After revision (Lines 17–19)

Abstract

Drug prices were significantly higher for PFS than for OS, while cost index—the drug price to ex-tend PFS or OS by one day—did not differ significantly between PFS and OS. The relationship between…

 

2.3. Data Analysis

Before revision (Lines 80-93)

The relationships between clinical usefulness and drug prices were analyzed using the following method. First, in relation to clinical usefulness, the pivotal clinical trial that formed the basis for approval of each anticancer drug was identified. The differences between the OS and PFS results in the experimental treatment arm and control treatment arm of that clinical trial were then obtained and calculated as follows:

• Improvement OS (%) = [median OS (days) in experimental treatment arm] / [median OS (days) in control treatment arm]
• Improvement PFS (%) = [median PFS (days)in experimental treatment arm] / [median PFS (days) in control treatment arm]

The drug prices of all anticancer drugs until disease progression in the experimental arm were defined as the Cost Index (CI) as follows.

• CIpfs (Yen) = drug price / ΔPFS
• CIos (Yen)= drug price / ΔOS

 

 

After revision (Lines 99–117)

The relationships between clinical usefulness and drug prices were analyzed using the following method. First, in relation to clinical usefulness, the pivotal clinical trial that formed the basis for approval of each anticancer drug was identified. The differences between the OS and PFS results in the experimental treatment arm and control treatment arm of the trial were then obtained and calculated as follows:

• Improvement OS (%) = (median OS [days] in experimental treatment arm) / (median OS [days] in control treatment arm)
• Improvement PFS (%) = (median PFS [days] in experimental treatment arm) / (median PFS [days] in control treatment arm)
• ΔOS (days) = (median OS [days] in experimental treatment arm)-(median OS [days] in control treatment arm)
• ΔPFS (days) = (median PFS [days] in experimental treatment arm)-(median PFS [days] in control treatment arm)

The incremental drug costs of all anticancer drugs until disease progression or death in the experimental treatment arm were defined as the Cost Index (CI), as follows.

• CIpfs (Yen) = drug price / ΔPFS
• CIos (Yen)= drug price / ΔOS

We also added the item “3.3 Correlation Between Improvement Rate of Overall Survival and Incremental Drug Costs” to the Results section. Furthermore, Table 2 and the explanatory text were moved to “3.2 Rates of Increase in OS and PFS Extension and Incremental Drug Costs.” Please refer to the response of Referee 2, #b.

 

f. No conclusion written in manuscript. It is little bit strange. As you write there is a correlation in abstract, but not supported in result part clearly. I think you need to rewrite accordingly and resubmit again.

 

[Response]

Thank you for your observation. We added following conclusion to the end of the Discussion section. Thank you for pointing this out.

Discussion, Lines 225-229

In conclusion, we investigated the relationship between drug prices and clinical benefits, such as OS and PFS, with anticancer drugs approved from 2009 to 2020 in Japan. We found drug prices were significantly higher for PFS than for OS, while the drug price to prolong one day for PFS or OS did not differ significantly between PFS and OS. We also found a correlation between drug prices and OS, but not for PFS.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors made significant revisions based on the reviewer’s comments. The manuscript can be accepted for publication.