Comparison between Smokers and Smokeless Tobacco Users in Their Past Attempts and Intentions to Quit: Analysis of Two Rounds of a National Survey

Round 1

Reviewer 1 Report

Thank you for the opportunity to review this manuscript. Having reread this manuscript (after reviewing for another journal), I see no major improvements to what was suggested to improve the manuscript. Currently, as this manuscript stands, the authors present past attempts to quit and future intentions to quit among exclusive cigarette smokers, and exclusive smokeless tobacco use in two rounds of GATS surveys done in Bangladesh. The interesting information is about quitting behaviours which is important in LMICs due to lack of perhaps assistance to quit services. In this context attempts to quit in the past and intentions to quit in future and their associations with demographic factors add very little to science. Quit intentions associations must be tested with factors such as dependence, exposure to harms of smoking, being counselled by a healthcare provider etc. Rather than attempting an incorrect Poisson regression (for a categorical variable), authors may consider reporting type of assistance attempted to quit, duration for which they quit, complete cessation. Examine if these have changed between two rounds of surveys. 

Authors provide % of current cigarette smokers and smokeless tobacco users who attempted to quit in the past and their intention to quit in future.  GATS survey reports provide this information on their statcompiler and country reports. So it is not relevant question.

 Authors should address more proximally associated factors such as duration, and number of cigarettes smoked, receipt of smoking cessation advice etc. Poisson regression is not correct approach used by the authors to assess the risk factors for quit attempts and intentions to quit in future. References are appropriate but not enough

 

Author Response

Please see the uploaded file "Response to reviewers' comments.docx"

Author Response File: Author Response.docx

Reviewer 2 Report

This study uses two waves of data (2009 and 2017) from the Global Adult Tobacco Survey (GATS) in Bangladesh to compare quit attempts, intentions, and methods used among smokers, smokeless tobacco (SLT), and dual users. Both smokers and dual users were more likely than SLT users to have attempted quitting during the past year and to intend to quit in the future. Few people received cessation assistance for either smoking or SLT. 

 

The study has several strengths. The topic is of considerable interest given high rates of tobacco use and low quit rates in SE Asia. The data are high quality by virtue of the use of a validated instrument and a nationally representative sample, and the paper is well written. I have some suggestions to improve the paper:

 

Methods, Section 2.2: provide more detail about how a “quit attempt” was defined or worded in GATS. For example, did an attempt only count if it lasted for 24 hours or was it worded as a “serious” quit attempt? 

 

Results, Table 2: Quitting behavior among dual tobacco users is an interesting topic but the analysis in Table 2 makes it difficult to understand the behavior of this group in some important respects. Both outcomes (quit attempts and quit intentions) for dual users apparently can refer to smoking, SLT use, or both products. It would be helpful to provide information on what product(s) dual users have tried to quit and want to quit in the future. Have they tried to quit both products in the past year, or do they prioritize one over the other? And the same for future intentions to quit.

 

Minor issues:

 

The title can exclude the word “in the future” since an intention to quit logically can only refer to the future.

 

In Table 2, the aPR (CI) 1.04 (0.97-1.12) should not be bolded since it is not statistically significant.

 

In Table 2, for the variable “Believe tobacco use is harmful for health,” the referent should be “no” not “yes” to be consistent with how the data are reported in the text (lines 203-206). Similarly, in Table 3 the referent for this variable also should probably be “no” instead of “yes” (the authors do not interpret this association in the text, so I am not absolutely sure whether the referent label is incorrect in Table 3, but the interpretation that smokers who do NOT believe that tobacco use is harmful are more likely to intend to quit is counterintuitive).

 

Discussion: on line 340, state the comparator: “Smokers in Bangladesh were significantly more likely THAN SLT USERS to have made quit attempts in the past 12 months…”

Comments for author File: Comments.pdf

Author Response

Please see the uploaded file "Response to reviewers' comments.docx"

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

I request the authors to admit to all the data limitations to assess the determinants of quit attempts and quit behaviors.

Author Response

Comment: I request the authors to admit to all the data limitations to assess the determinants of quit attempts and quit behaviors.

Response: We thank the reviewer for reading the previously revised version of our manuscript and offering this latest round of comments. We have now provided a point-by-point response and outlined all the limitations, as shown below:

The study has some limitations. Firstly, the GATS collects self-reported data, and thus the previous attempts and intentions to quit tobacco in the future may have been influenced by the recall and social desirability biases. Secondly, attempts and intentions to quit may be associated with the level of tobacco dependence [38,39]; however, our "user type" variable which has two categories – occasional users and daily users – may not have captured the exact degree of dependence. Also, it would have been useful to examine the relationship between quit duration in past attempts and intention to quit in the future. However, since only a small proportion of participants mentioned the quit duration in the form of months, weeks, and days instead of a precise measure, we could not include it. Furthermore, data were collected from Bangladesh, and therefore, the findings may not be generalisable to all settings. However, given that there are similarities among the neighbouring countries in terms of the pattern of tobacco use and related factors, the study findings may be partially generalizable to the sub-continent.