Dietary and Activity Factors Influence Poor Sleep and the Sleep-Obesity Nexus among Children

Comment   From Reviewer 1

Response:

Point 1:

Introduction   >> Line 3: Finnish children instead of finish children 

This   has been corrected.

(See highlighted change on line 72)

Point 2:

Table   1: Under the sleep-wake categories, what do EL, EE, LL and LE stand for?   Please, could you define those acronyms in the footnote.

These   have been added to the legend in Table 1 as below:

EL=   early to bed/late to rise; EE=   early to bed/late to rise; LL=late   to bed/late to rise; LE=late to   bed/late to rise

Point 3:

Discussion   >> Lines 295-297. According to the authors, screen devices and SSB   consumption significantly increase the likelihood of having =3 sleep problems.   I would be a little more cautious presenting these results, saying for   example that those variables were associated with grater odds of having =3   sleep problems. Based on the empirical results, association is all that this   study can prove and, so, the way it’s currently written suggests some kind of   causal-effect relationship.

The   wording has been changed to infer the association to be of increased odds   rather than causation.

(See highlighted changes in line   299 and below)

… “were   shown to be associated with   increased odds of having ≥3 sleep problems”

Point 4:

Discussion   >> Lines 295-297. I’d state as a limitation the fact that the study   sample does not represent the population of 10 years-old in Victoria and   Australia. This brings some issues with the feasibility of extrapolating the   results, which must be admitted.

In   response to the study population limitations:

Text   has been added to further clarify the limitation of generalising results   beyond the current sample (Please see   highlighted changes in lines 412-415) and below.

“As mentioned, the current study   benefits from a large representative sample of children within two rural   regions of Victoria, with high school (GSC=68%, GV60%) and student (GSC=90%,   GV=78%) response rates.  However, although the current   population groups did not differ significantly, it cannot be assumed that the   reported associations can be generalised beyond children aged 8-13 years   within rural Victoria, Australia.”

Point 5:

Discussion   >> I missed a paragraph stating implications for health and education   policymaking.

Thank   you for this comment and our clear omission of this critical implications   piece.  The below text has been   added  (Please see highlighted changes in lines 400-408).

These findings support the importance of adhering to the 24-hour   movement guidelines outlined by the Australian Government Department of   Health, considering the role of multiple behaviours on the healthy   development and obesity risk among children and adolescents [1]. However, with current   results suggesting the importance sugar sweetened beverage consumption on children’s   sleep behaviours and weight status outcomes, there is a need for more   specific guidelines on these behaviours. Current Australian guidelines   recommend the intake of SSB be limited [2], however it would be   recommended the Government work with researchers to provide quantifiable   guidelines to limit this behaviour (similar to the recommended 2 hour screen   time limit).

Comment   From Reviewer 2

Response:

Major   comments:

Point 1:

Authors   think too much of statistical significance (P<0.05 and/or P<0.01).   Recently, it was recommended not to mention statistical significance, but to   present confidence intervals   (https://www.nature.com/magazine-assets/d41586-019-00857-9/d41586-019-00857-9.pdf).

The   authors have removed the over emphasis on statistical significance, firstly   by removing language indicating the existence or lack of a significant   difference or association. Secondly, taking into account your later comment   on reporting exact p-values, these have now been provided.  Where relevant the confidence intervals   (previously provided) are also reported. This will allow readers to interpret   the occurrence of associations for themselves. 

Point 2:

Please   present participation rate of Great South Coast Childhood Obesity Monitoring   Study and the Goulburn Valley Health Behaviours Monitoring Study, especially   in the grade four and grade six. This information is important to know how   the results represent the target populations.

Data   around school and student response rates have been included in the methods   section, including a breakdown per region.

(Please see highlighted changes   in lines 106-114)

Point 3:

Personally,   I have favorable impression on your presenting all the regression analysis.   However, there may be readers who concern multiple analysis. You can present   your main outcomes in tables 3 and 4, and you can move the others to   supplement. And if you present exact p-values, readers can perform Bonferroni   correction to deal with multiple analysis.

The   authors agree with being in favour of presenting all regression analysis   tables, and have maintained these within the manuscript.

As   previously mentioned, the authors have also provided the exact p-values for   the readers benefit.

Point 4:

In   the discussion, please compare your study population with other studies analyzing   the similar age range.

Where   possible studies of similar samples were used to discuss previous finding   with those of the current study (age ranges 8-16) [1-8]. However, due to   limited available studies focussing on the different dimensions of sleep and   the multiple behaviours of physical activity, screen-time and diet, studies   including slightly older adolescent populations (age ranges 6-19) [9-13] and adults [14,15] have been   incorporated to further unpack the findings of the current sample.

Throughout   the discussion text has been added to more clearly highlight the age group of   studies used for comparison.

Point 5:

Physical   activity and sleep are not recorded or monitored with devices. Study   population is in one region in one country. And it is not clear that the   results are also true in other locations. Please add these as study   limitations.

In   response to the study population limitations:

Text   has been added to further clarify the limitation of generalising results   beyond the current sample  as   highlighted by Reviewer 1. (Please see   highlighted changes in lines 412-415).

“As mentioned, the current study   benefits from a large representative sample of children within two rural   regions of Victoria, with high school (GSC=68%, GV60%) and student (GSC=90%,   GV=78%) response rates.  However, although the current   population groups did not differ significantly, it cannot be assumed that the   reported associations can be generalised beyond children aged 8-13 years   within rural Victoria, Australia.”

In   response to the limitations around the measurements used for physical   activity and sleep:

This   had been mentioned in the limitations initially. Text has been added to   further emphasise the need for more objective measures. (Please see highlighted changes in lines 416-422 and below).

“Furthermore, while questionnaire   items were sourced from studies with previous validated question items   [48-50] with these and similar self-report questionnaires validated among   young children, the self-report nature of the sleep and behavioural factors   items could be subject to recall or report bias [79]. While there is evidence supporting the   suitability for use of questionnaires in the examination of the association   with weight status [80], objective measures (such as accelerometry) are   deemed more accurate for assessment of sleep and physical activity behaviours.”  

Minor   comments:

Point 6:

Lines   199-200: “serves”. Please clarify serves per day or per week.

This   has been amended.

(See highlighted changes on lines   199-201)

Point 7:

Line   205: Please explain “takeaway” more. For non-natives, it is not easy to   understand “takeaway”.

This   has been amended as below.

(See below and highlighted change   on line 205)

“(meals purchased from restaurants or fast-food vendors)”

Point 8:

Table   1: Please spell out ICSEA, EL, EE, LL and LE in the legend.

Thank   you for this comment, these have been added to the legend in Table 1 as   below:

ICSEA= Index of Community Socio-Educational   Advantage; EL= early to bed/late to rise; EE= early to bed/late to rise; LL=late to bed/late to rise; LE=late to bed/late to rise

Point 9:

Throughout   the tables: please present exact p-values instead of P<0.01, p<0.05 and   NS (for example, P=0.352, P=0.024, P=0,003 or P<0.001).

Thank  you for this comment, as mentioned above   (Point 1), this has been addressed throughout.

Point 10:

Table   2: please add total number of subjects who had one or less, two, three or   more sleep dimensions scored poor at the top of the table.

This   had initially been left out as these figures are outlined in Table One, as   well as to minimise clutter.

This   has now been amended.

(See highlighted changes at the   top of Table 2)