Optimizing Ramadan Fasting
A Randomised Controlled Trial for People with Type 2 Diabetes during Ramadan
Applying the principles of the ADA/EASD Consensus

by Ibrahim et al, is a comprehensive study looking at the effects of a complex intervention of -1. Patient education and 2. Adjustment of medications in a randomised controlled trial. Ideally such a complex intervention study should be done in a cluster randomised manner-the reason being the effect of spill over in terms of education to others within the same centre. For e.g. clinicians all will still be adjusting the medications in both arms as is standard or care and were provided similar education. Although a pre-Ramadan visit may not have been arranged they would still adjust the medications at a previous visit. The patients themselves may be seeing the clinicians may be related as family or friends and they may communicate with each other on the education provided. This is a limitation that may need to be described as a limitation.
The discussion does not have strengths and limitations outlined.
The discussion needs to be re-written.

Other comments:
1. Was the trial registered at a public site-like clinical trials.com
Is a pre-study protocol and pre-study statistical analysis plan available.

1. Tables..2 , 3 and supplementary tables
   Please add in the P values.

2. For supplementary tables and main tables -they should be separated as it is confusing currently -the way they are arranged.

3. Discussion
   The discussion is too long...it can be shortened.
   It needs to focus on pertinent findings and comparability with other relevant evidence
   Specifically - Generalisability and Limitations need to be stated clearly.

Information about the intervention need not be repeated and should be described briefly in the intervention section.
Liraglutide details not needed as minority used Liraglutide in the study.
A comprehensive review on the subject is not needed...

The manuscript is much improved.
However consider - sharpening the statistical analysis plan and sample size writeup - with a statistician.
Follow the consort guidelines for reporting of clinical trials.

Limitations that this is not a cluster randomised trial and that there is a possibility of
spill over in terms of education to others within the same centre. For e.g. clinicians all will still be adjusting the medications in both arms as is standard or care and were provided similar education. Although a pre-Ramadan visit may not have been arranged they would still adjust the medications at a previous visit. The patients themselves may be seeing the clinicians may be related as family or friends and they may communicate with each other on the education provided. This is a limitation that may need to be described as a limitation.

1. Sample size calculation : This needs an evaluation by a statistician. Exactly how much difference between the 2 groups (control and intervention) is expected in terms of reduction in hypoglycaemia . Since they report OR ...with this sample size...what is the odds in the intervention group vs control group taken into consideration when sample size is calculated.

2. Statistical analysis : Logistic regression analysis is used for outcome variables (categorical) - not count variables...it could be used for any occurrence of hypoglycaemia and /or weight gain as a composite outcome as dependent variable but not number of hypoglycaemia events...Count data is analysed by : ordinary least squares (OLS), OLS with a transformed dependent variable, Tobit, Poisson, overdispersed Poisson, negative binomial, ordinal logistic, and ordinal probit regressions. The statistical analysis writeup needs an evaluation by a statistician and needs to be sharpened.