Drink Choice is Important: Beverages Make a Substantial Contribution to Energy, Sugar, Calcium and Vitamin C Intake among Australians
Round 1
Reviewer 1 Report
Line 93 Explain in the Methods why only data from the first 24h recall were used and not both days. Include this in the Discussion as a limitation because two days would provide a better representation of usual intake and allows for adjustment for within-person variation.
Discussion Line 240-241: It is unlikely that many Australians are struggling to achieve adequate energy intakes, rather the opposite. Please discuss the issue of excessive dietary energy intake in the context of the findings of the Australian dietary guidelines panel
“It is probable that consumption of sugar-sweetened beverages is associated with weight gain in children and adults (Grade B; Evidence Report, Section 15.1)”.
“The US review found that, for most children, there was limited evidence that intake of fruit juice is associated with increased adiposity when consumed in amounts that are appropriate for the age and energy needs of the child. However, increased intake of fruit juice was found to be associated with increased adiposity in children who were already overweight or obese”.
Line 268 In the Discussion expand on the reason for assessing vitamin C intake given that vitamin C is not a nutrient at risk of deficiency in the Australian population despite the lower than recommended intake of fruit and vegetables. Are you suggesting that vitamin C intake could become limiting if beverage sources changed?
The page numbering needs attention
Author Response
Response to Reviewer 1 comments:
Thank you very much for the relevant points raised in review. We have addressed the reviewers’ comments in the text and provided a response to each point below.
Point 1: Line 93 Explain in the Methods why only data from the first 24h recall were used and not both days. Include this in the Discussion as a limitation because two days would provide a better representation of usual intake and allows for adjustment for within-person variation.
Response 1: The second 24 hour recall was only completed by 64% of respondents who completed the first (N=7,735). While still a large sample, the reduction of sample size would compromise our estimates for sub-groups and potentially introduce a new response bias. The use of two days of intake does enable a better representation of individual intake (for fewer individuals), but not for group intake which is the point of interest in the manuscript. The adjustment for within-person variation is relevant when variation of usual intake across the population/group is the issue of interest. By using one day of intake we are able to estimate the mean of usual intake (equal to the mean of one day intake), and also a measure of the variation of intake in a single day in the population groups (which we have described as the 10th and 90th percentile for consumers in Supplementary Table 1). We think this is a measure of interest to readers.
We have included a sentence to outline this in our methodology (lines 83-86), and have added relevant detail as a limitation (lines 375-379).
Point 2: Discussion Line 240-241: It is unlikely that many Australians are struggling to achieve adequate energy intakes, rather the opposite. Please discuss the issue of excessive dietary energy intake in the context of the findings of the Australian dietary guidelines panel
“It is probable that consumption of sugar-sweetened beverages is associated with weight gain in children and adults (Grade B; Evidence Report, Section 15.1)”.
“The US review found that, for most children, there was limited evidence that intake of fruit juice is associated with increased adiposity when consumed in amounts that are appropriate for the age and energy needs of the child. However, increased intake of fruit juice was found to be associated with increased adiposity in children who were already overweight or obese”.
Response 2: To address these issues raised by the reviewer, we have included additional text in the discussion to highlight the importance of these results in the context of excessive dietary energy intake (Lines 236-238). This analysis provides evidence that beverages such as coffee and alcoholic drinks are also important contributors to total energy intake, as well as the more generally recognised sugar-sweetened beverages and fruit juice. We have tried to make this point, which we think is policy relevant, at lines 302-308.
Point 3: Line 268 In the Discussion expand on the reason for assessing vitamin C intake given that vitamin C is not a nutrient at risk of deficiency in the Australian population despite the lower than recommended intake of fruit and vegetables. Are you suggesting that vitamin C intake could become limiting if beverage sources changed?
Response 3: The reviewer raises a good point. We were surprised at the contribution to vitamin C intake arising from fruit juice, particularly in children. However the population is probably not meaningfully at risk of deficiency if fruit juice were no longer to be consumed. However, subgroups within the population may be vulnerable to sub-optimal vitamin C intake and we have now highlighted this in lines 268-272.
Point 4: The page numbering needs attention
Response 4: I have tried but failed to fix the errors in page numbering (including starting again from the template). The problem seems to be related to the included figure 1. I have requested that the managing editor please reformat for me.
Thank you once again for your review.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Dear Authors,
It is a very interesting manuscript. Please find several comments/suggestions below:
Abstract
Line 13 - please add 'daily' after beverage
Introduction - I suggest to rewrite some parts of the introduction to make it more focused on the aim of this study (and more linked to the title of this manuscript) at the moment the introduction is discussing a pattern of beverage intake rather than the contribution of beverages to the overall dietary intake
Methods
Line 93 "All analyses were conducted using the first day of dietary recall (day 1 data).." - it is not clear why the second 24-hour recall was excluded
Line 127 - "Estimates were calculated for consumers of each beverage category and for the total population within each demographic group." - I am rather confused with this statement; yes it is correct to calculate for consumers but what do you mean by estimates for the total population? have you included both people who consumed and who did not? If so i would say remove these analysis and continue in the results with 'a real' intake only
It is not clear how the authors differentiated between sugar and non sugar-sweetened beverages when estimating sugar intake for example soft drinks can be sugar and non-sugar sweetened
Statistical analysis - you mentioned to use a t-test and Kruskal-Wallis; one of them is a parametric and other a non-parametric; why a mixture of parametric and non-parametric methods were used for the same set of data?; has a data distribution been checked before the analysis?
Also, some of data were categorical - please mention a test which has been used to analyse this type of data.
Results
Line 144 - you said 'the median' next line 147 - is 'the mean'; please make sure that the statistics are reported correctly so depends on the distribution of the data it should be either mean or median for all results
I would suggest to add some additional information/analysis. You reported the contribution of different drinks to energy, Ca, Vit. C and sugar intake but it would be worth to show how it impacts participants overall dietary intake for example if higher intake of some beverages means higher total energy intake or higher intake of milk means higher overall Ca intake etc.
Supplementary Table 2 - the meaning of letters 'a,b,c' should be explained
Author Response
Response to Reviewer 2 Comments:
Thank you very much for your review of our manuscript. We have provided a response to each of the points made in italics below.
Abstract
Point 1: Line 13 - please add 'daily' after beverage
Response 1: Amended to be day of the survey, as ‘daily’ refers more to frequency and this was not measured.
Line 13: ‘on the day of the survey’ has been added after beverage ()
Point 2: Introduction - I suggest to rewrite some parts of the introduction to make it more focused on the aim of this study (and more linked to the title of this manuscript) at the moment the introduction is discussing a pattern of beverage intake rather than the contribution of beverages to the overall dietary intake
Response 2: Thank you for this suggestion – we have now re-written the introduction and removed a section. See Lines 54-67.
Methods
Point 3: Line 93 "All analyses were conducted using the first day of dietary recall (day 1 data).." - it is not clear why the second 24-hour recall was excluded
Response 3: Only 64% of subjects provided a second day of dietary recall, which was done by telephone interview rather than face-to-face interview. Our main estimates of interest were the means of intake for groups, and we felt that restriction of the sample would compromise our capacity for sub-group analysis and potentially introduce bias into the population estimates. We have provided a brief reason for not incorporating the second day at lines 83-86, and included in the study strengths and weaknesses in the discussion at lines 375-379.
Point 4: Line 127 - "Estimates were calculated for consumers of each beverage category and for the total population within each demographic group." - I am rather confused with this statement; yes it is correct to calculate for consumers but what do you mean by estimates for the total population? have you included both people who consumed and who did not? If so i would say remove these analysis and continue in the results with 'a real' intake only
Response 4: For this analysis we are estimating the contribution of beverages to the population intake of energy and nutrients. We believe it is important to include people who have not consumed the beverage on the day of the survey because in this case the beverage contribution for that person is zero. Beverages that are not often consumed, or are consumed by few people, contribute proportionately less to population intake. However it is also relevant to consider how a beverage contributes to nutrient estimates when it is consumed – therefore both estimates are provided.
Point 5: It is not clear how the authors differentiated between sugar and non sugar-sweetened beverages when estimating sugar intake for example soft drinks can be sugar and non-sugar sweetened
Response 5: Thank you for raising this point, we have not been clear in explaining our methods. There were 448 different beverages in the survey which were categorised into 12 categories. Each beverage retained its unique nutrient profile, or, where additions were identified (i.e. milk in coffee) the nutrient profile for each addition was added on. For each individual, nutrient contributions were then aggregated into 12 beverage categories to assist interpretation. To a greater or less extent, all of the categories included beverages which varied in composition – for example, in the coffee category, coffee could include milk or not, and include sugar or not; plain milk included regular milk, reduced fat milk and skim milk. Additions to the methods have been made at lines 103-04, and lines 107-108.
Point 6: Statistical analysis - you mentioned to use a t-test and Kruskal-Wallis; one of them is a parametric and other a non-parametric; why a mixture of parametric and non-parametric methods were used for the same set of data?; has a data distribution been checked before the analysis?
Response 6: The data are predominantly beverage intake data therefore skewed (i.e. not normal). Non-normal data is particularly the case for population data where there may be a large percentage with zero intake. We agree that non-parametric testing is appropriate and have now retested male-female differences using the Mann-Whitney test (Figure 2 and supplementary table 1). While the p-values changed, there was little difference to those contrasts assessed to be significantly different. The statistical analysis section has been modified to provide more clarity.
Point 7: Also, some of data were categorical - please mention a test which has been used to analyse this type of data.
Response 7: Thank you for pointing out this omission – the chi-squared test was used and this is now included at lines 133-134.
Results
Point 8: Line 144 - you said 'the median' next line 147 - is 'the mean'; please make sure that the statistics are reported correctly so depends on the distribution of the data it should be either mean or median for all results
Response 8: Thank you – the median is now used for both statements referred to (lines 139-144).
Point 9: I would suggest to add some additional information/analysis. You reported the contribution of different drinks to energy, Ca, Vit. C and sugar intake but it would be worth to show how it impacts participants overall dietary intake for example if higher intake of some beverages means higher total energy intake or higher intake of milk means higher overall Ca intake etc.
Response 9: Thank you for your suggestion for additional presentation of the data. Our purpose for this manuscript was to describe the overall contribution of beverages and beverage categories to selected nutrient intakes for the Australian population, and to various population sub-groups. We could achieve this using the population based sample providing recall data for one day of dietary intake. We think the question posed is interesting, with some technical issues to address because nutrient intake and beverage intake may also be linked by confounding factors such as age, gender, and physical activity. We think the best way to examine this is using the smaller sample with 2 days of dietary intake to compare whether there is evidence within individuals of consistent differences for nutrient intake on a day when intake of the beverage of interest differs. We could also examine the extent to which drink intake is replaced by other drinks (which may represent a form of nutrient substitution). We consider that this question is of sufficient complexity to be the subject of another manuscript.
Point 10: Supplementary Table 2 - the meaning of letters 'a,b,c' should be explained
Response 10: The superscript letters are explained in the footnote to Supplementary Table 2, however they had not been explained for Supplementary Table 1. We have now included this information in the footnote for that table.
Thank you once again for your review.
Author Response File: Author Response.docx
Reviewer 3 Report
The authors in Introduction should better describe the variability of the juice on the market and describe their variability in nutrient and bioactive components composition. This should be better described also in the Discussion of results.
The aim of work should be rewritten. Some introductive lines before the subdivision in subparagraphs should be inserted in Results. In conclusion the authors should remark the advantages and limits of this research.
Author Response
Response to Reviewer 3 Comments:
Thank you for your review of our manuscript
Point 1: The authors in Introduction should better describe the variability of the juice on the market and describe their variability in nutrient and bioactive components composition. This should be better described also in the Discussion of results.
Response 2: The introduction has now been revised to include brief discussion of the variability of beverages.
Point 2: The aim of work should be rewritten. Some introductive lines before the subdivision in subparagraphs should be inserted in Results. In conclusion the authors should remark the advantages and limits of this research.
Response 2: The aim of the work has been re-written (lines 53-55), the strengths and limitations of the work are outlined in lines 358-384 and have been revised so they can be more clearly identified. We are not sure about what is meant by adding introductive lines before the subdivision in subparagraphs in Results. We have modified the headings (3.1 and 3.2) in the results section for better clarity.
Thank you once again for your review.
Round 2
Reviewer 1 Report
Thank you for addressing the comments
