An Evaluation of Point-of-Care HbA1c, HbA1c Home Kits, and Glucose Management Indicator: Potential Solutions for Telehealth Glycemic Assessments
Round 1
Reviewer 1 Report
This study, about comparing alternative glycemic measures is quite relevant given the dramatic increase in telemedicine visits and has both clinical and research implications.
Introduction:
- I encourage the authors to consider revising to improve the flow and logical progression for setting the stage of the study.
- Given that GMI is a lesser known indicator of HbA1c, please enhance the description here or in the methods section.
Methods:
- Why wasn’t a laboratory A1c test performed for onsite clinical visits? Consider explaining here or in the discussion section (in relationship to current lines 159-164) why POC was used as the reference value vs laboratory testing.
Discussion:
- Given that 44% of GMI vs POC values deviated by a clinically significant amount, I think additional explanation of the utility of GMI is warranted. The authors highlight the benefits (convenience when already using CGM), but what will it tell the user? And will a clinician trust it and for what purposes?
- Would the authors be able to discuss their study findings within the context of repeat testing over time? Would repeat testing potentially introduce additional variation and to what extent?
Conclusion: Lines 159-164 – move to discussion section.
Author Response
We would like to thank Reviewer 1 for the thoughtful comments. We have addressed all feedback and uploaded our response to all of the questions and comments from Reviewer 1. Thank you!
Author Response File: Author Response.docx
Reviewer 2 Report
The authors compare 3 technologies for the estimation of HbA1c on a cohort of 71 patients:
ARDL (home kits) n=99
GMI (from CGM) n=88
Point Of Care (during in-clinic visits) n=32
The authors built a dataset matching measurements from the different technologies:
ARDL vs POC n=32
ARDL vs GMI n=88
POC vs GMI n=27
Here follows the comments to the manuscript:
x. It is not clear in lines 75-80 how the matching numbers are obtained. The authors should expand on how for the comparison POC vs GMI n=27 (instead of 32) and also expand on why for ARDL vs GMI n=88. In the latter case, is it because the authors included measurements collected both in clinic and at home for a subset of the subjects?
x. The authors applies the Lin's correlation coefficient to pairs of measurements, including repeated measures from the same subject. This is true for ARDL vs GMI with n=88 since the total number of participants is 71. Measures from the same subject are likely to be correlated and not independent, with the net results of inflating the Lin's correlation coefficient. Hence the authors should repeat the analysis using only one single matched measurement per subject.
Are there repeated measurements also for the other 2 comparisons (ARDL vs POC and POC vs GMI)? if yes, then the one measurement per subject rule applies.
x. The Lin's correlation equivalence and the Bland-Altman plots do not provide any conclusion on the equivalency of the different measurement technologies. It would be good if the authors could expand the analysis including a TOST analysis for equivalency, using as equivalency boundaries what they call the "clinically meaningful amount" of 0.5%. This analysis would contextualize the inherent variability in the measurements to what is considered the minimum clinical important difference, a well known concept in assays comparisons. If the variability in the measures is smaller than what is consider clinically important, then I agree with the conclusions of the manuscript. If not (there is a chance the number of matched pairs might too small) I would tune down the conclusion and suggest to collect more data (the work from Beck et al. was much more powered).
I recommend a major review of the manuscript before accepting it for publication.
Author Response
We would like to thank Reviewer 2 for their time and critical review of our manuscript. We have acknowledged and addressed all of Reviewer 2 suggestions and have uploaded a document with a response to Reviewer 2 comments and the changes made. Please see the attachment. Thank you!
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The authors addressed the open points mentioned in the review. Before accepting the paper for publication it would be good for the authors to provide in the manuscript an indication about which statistical test they have used for the TTOST and provide in the results section a table with the p-values and the percentages of points outside the boundaries [-0.5 | 0.5 %]. This would justify their statement about equivalency.
Author Response
We thank the reviewer for this comment. Please see the attachment for the updated changes.
Author Response File: Author Response.docx