#COVID19 and #Breastcancer: A Qualitative Analysis of Tweets

Round 1

Reviewer 1 Report

 The  research question identifying the effective use of twitter during the pandemic is a useful  and adds to current knowledge on the use of social media in research and practice. 

The results were clearly outlined and the methodology were clear with regards to the use of the software, Just a question of clarity Line 75, Does Symplur routinely collect data from all social media sites? If so was twitter just one if the platforms from which data was extracted?  The flow chart could be labelled so the reader can identify that it was not a part of the results.  ( Should be fig 1? ) 

Line 122 the use of the word may  suggests the activity has not been completed. 

Lines 175 and 201 seems to have  typographical errors.

The figures in lines 321 and 322 need to be clarified, I understand that this is a tweet  but the numbers seem to be incorrectly cited and that could lead to misinformation in the  tweet . Is it 789,800 women or 7,89,800 suggesting millions ? 

line 631 could you comment in the credibility of the data presented especially since was not mentioned that the dissemination involved peer review?   This could lead to misinformation, which was alluded in the discussion  but could also lead to reservations of the tool as a means of communication.

line 701  Would the number of retweets help in this assessment ?  It was mentioned as a limitation of the study that the retweets were included . 

I understand that twitter was popular during the pandemic, were there other social media platforms that were utilized  that may have been as effective as twitter? 

Author Response

Thank you for the review of our article. The following is our response to Reviewer 1's comments and suggestions. 

The research question identifying the effective use of twitter during the pandemic is a useful and adds to current knowledge on the use of social media in research and practice.  

The results were clearly outlined and the methodology were clear with regards to the use of the software, Just a question of clarity Line 75, Does Symplur routinely collect data from all social media sites? If so was twitter just one if the platforms from which data was extracted?  The flow chart could be labelled so the reader can identify that it was not a part of the results.  ( Should be fig 1? )  

• Additional details regarding Symplur and its function are included in the data collection section (page 2-3). 

Line 122 the use of the word may suggests the activity has not been completed.  

• This paragraph was left in the text in error and has been removed. 

Lines 175 and 201 seems to have typographical errors. 

• These lines are extracted directly from tweets. To maintain the integrity of the data, no editing of the tweets was done prior to including them in the paper. 

The figures in lines 321 and 322 need to be clarified, I understand that this is a tweet but the numbers seem to be incorrectly cited and that could lead to misinformation in the tweet . Is it 789,800 women or 7,89,800 suggesting millions ?  

• As in the convention in India, the comma is in the correct place and the number is seven hundred eighty-nine thousand eight hundred. The way the number is written “7,89,800” in the tweet is correct. 

line 631 could you comment in the credibility of the data presented especially since was not mentioned that the dissemination involved peer review?   This could lead to misinformation, which was alluded in the discussion  but could also lead to reservations of the tool as a means of communication. 

• Additional discussion regarding this has been added to the discussion section (page 20) 

line 701  Would the number of retweets help in this assessment ?  It was mentioned as a limitation of the study that the retweets were included .  

• Retweets were included in the total dataset (meaning duplicates of the same tweet). However, the number of total times that an individual tweet was retweeted was not included in the data. Nor was there any connecting information such as which users liked the tweet.  

I understand that twitter was popular during the pandemic, were there other social media platforms that were utilized  that may have been as effective as twitter?  

• This study did not specifically examine other social media platforms. Twitter was selected as it is a common platform used among the stakeholders of interest. 

Reviewer 2 Report

Thanks for the opportunity to review this article. This paper describes a qualitative analysis of tweets about #COVID19 and #Breastcancer. This article shows innovative research that can help researchers acknowledge aspects being discussed in a specific environment of the Internet. I found the research interesting, albeit some minor aspects can be revised.

In design, the authors write, «This qualitative descriptive study utilized content extracted from Twitter in the form of tweets. Tweets shared by clinicians, health researchers, advocacy organizations, breast cancer patients, and support persons were assessed using qualitative thematic content analysis». Albeit I understand that this analysis is the main aim of the study, I would invite the authors to offer a more detailed description of the overall design of the research. It seems confusing to mix qualitative opinions from different actors, like clinicians, health researchers, advocacy organizations, breast cancer patients, and support persons.

The authors write, «Symplur, a data analytics program, was initially used for data extraction.» As this tool, Symplur, is the key element used to retrieve the data analyzed, I would invite the authors to show a better description of the tool and its advantages or disadvantages. I believe this is important to allow the readers to assess better the results and the conclusions obtained. This is important, as the tool can introduce a considerable selection bias.

The authors describe that the tweets included were written between March 11, 2020, to October 31, 2020. I would invite them to describe why they chose this period. Later, the authors write, «The hashtags #BreastCancer, #BreastOncology, #Pandemic, and #COVID19, were used with a time interval from March 11, 2020, to January 31, 2021». Once more, I would invite the authors to explain why they chose this period instead of two full years and why they decided to mix data from different periods.

The authors write, «198 tweets were obtained in this same section. Once retweets were excluded from the dataset, 68 tweets remained. The content of the tweets was then assessed independently by two researchers (GN and IB). Those tweets that were not pertinent to the research question were excluded, leaving a sample size of 42 tweets». This is another essential part of the research, as the initial 198 tweets obtained are filtered, and finally, only 21% of the initial tweets are analyzed. This process is, therefore, essential as, again, it can introduce a considerable bias. Thus, I would invite the authors to describe this process in more detailedly. This applies to the second separate search.

In Table 2, the authors mix countries and one continent, Europe. Is it a mistake? Furthermore, it has no sense that Europe has only five tweets.

Please remove the text «This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn».

I understand this is qualitative research, but the results section seems too long. If possible, I would invite the authors to try to shorten it. This is just a proposal.

In the limitations, I would invite the authors to describe and, if possible, assess the potential selection bias described before.

Author Response

Thank you for the review of our article. The following is our response to Reviewer 2's comments and suggestions. 

Thanks for the opportunity to review this article. This paper describes a qualitative analysis of tweets about #COVID19 and #Breastcancer. This article shows innovative research that can help researchers acknowledge aspects being discussed in a specific environment of the Internet. I found the research interesting, albeit some minor aspects can be revised. 

In design, the authors write, «This qualitative descriptive study utilized content extracted from Twitter in the form of tweets. Tweets shared by clinicians, health researchers, advocacy organizations, breast cancer patients, and support persons were assessed using qualitative thematic content analysis». Albeit I understand that this analysis is the main aim of the study, I would invite the authors to offer a more detailed description of the overall design of the research. It seems confusing to mix qualitative opinions from different actors, like clinicians, health researchers, advocacy organizations, breast cancer patients, and support persons. 

• Greater detail has been added to the data collection section (page 2-3). This includes additional explanation regarding the inclusion of different actors including patients, clinicians, and support persons. 

The authors write, «Symplur, a data analytics program, was initially used for data extraction.» As this tool, Symplur, is the key element used to retrieve the data analyzed, I would invite the authors to show a better description of the tool and its advantages or disadvantages. I believe this is important to allow the readers to assess better the results and the conclusions obtained. This is important, as the tool can introduce a considerable selection bias. 

• Additional details regarding Symplur are included in the data collection section (page 2). 

The authors describe that the tweets included were written between March 11, 2020, to October 31, 2020. I would invite them to describe why they chose this period. Later, the authors write, «The hashtags #BreastCancer, #BreastOncology, #Pandemic, and #COVID19, were used with a time interval from March 11, 2020, to January 31, 2021». Once more, I would invite the authors to explain why they chose this period instead of two full years and why they decided to mix data from different periods. 

• Explanation for the time intervals and mixed datasets used in this analysis are included in the data collection section (page 2) and under limitations (page 22). 

The authors write, «198 tweets were obtained in this same section. Once retweets were excluded from the dataset, 68 tweets remained. The content of the tweets was then assessed independently by two researchers (GN and IB). Those tweets that were not pertinent to the research question were excluded, leaving a sample size of 42 tweets». This is another essential part of the research, as the initial 198 tweets obtained are filtered, and finally, only 21% of the initial tweets are analyzed. This process is, therefore, essential as, again, it can introduce a considerable bias. Thus, I would invite the authors to describe this process in more detailedly. This applies to the second separate search. 

• Further detail regarding the assessment of tweets included in the final dataset is included in the data collection section (page 2). 

In Table 2, the authors mix countries and one continent, Europe. Is it a mistake? Furthermore, it has no sense that Europe has only five tweets. 

• An explanation for this is provided in the demographics section on page 6. 

Please remove the text «This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn». 

• This has been removed. 

I understand this is qualitative research, but the results section seems too long. If possible, I would invite the authors to try to shorten it. This is just a proposal. 

• While this is a very thoughtful suggestion, we, as a team feel that the quotes included in the results section area important to maintain the integrity of the analysis. Moreover, the length of this section is comparable with other qualitative papers. We have not made any changes to this section. 

In the limitations, I would invite the authors to describe and, if possible, assess the potential selection bias described before. 

• Additional discussion of potential bias is included on page 22.