Develop and extend Iglu for analyzing circadian patterns.

[steckhan]

Dear Authors,

would it be possible to contribute to Iglu? I have some ideas for integrating more metrics. Best,
Nico

[irinagain]

Thank you for your interest! In general, it's definitely possible to contribute and you can see that there is quite a large number of contributors (it works through fork/pull requests). However, in case of new metrics (rather than fixes/adjustments to existing ones), we want to be careful to only add those that are either of direct need to our users or have been proven relevant in clinical literature. The rationale is that the already existing number of CGM metrics makes can make it very confusing as to which ones should be used over others in the absence of additional justification, a criticism we also got from some reviewers. Thus, if you have some new metrics in mind, it would be good to have references and justification to their need before integrating them as part of the iglu. Would it be possible for you to elaborate/provide references as to which features you have in mind? I am also happy to discuss offline. As with any updates, any additions would need to have complete documentation and pass the default R package checks as necessary steps of being merged to the main branch.

[steckhan]

Dear Irina, Thanks for your mail and the explanation. Sure you need to have only relevant functions for metrics included otherwise this is pretty flooded. To start this my idea would be to enable users to calculate their circadian rhythm and for this there is a publication about a metric that does this based as a non parametric approach To activity data recorded with e.g. activity trackers (such as actiwatch, Fitbit, AppleWatch, ... or IMUs in smartphones). Please have a look: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4890079/ The nparACT has metrics that estimate, the circadian patterns based on measures of interdaily stability (IS), intradaily variability (IV) and relative amplitude (RA) of activity and gives the start times and average activity values of M10 (i.e. the ten hours with maximal activity) and L5 (i.e. the five hours with least activity). I would include a documentation in iglu about circadian patterns and health aspects or how studies might extract and interprete those metrics. We have seen the importance not only in our own research but recently others about the case of gestational diabetes have seen this as a good predictive candidate. Would you like to discuss this online ? https://www.thieme-connect.de/products/ejournals/abstract/10.1055/s-0042-1746365 Analysis of the link between chronotype and glucose dynamics by machine learning algorithms thieme-connect.de Best, Nico

…

Am 10.06.2024 um 20:37 schrieb Irina Gaynanova ***@***.***>: Thank you for your interest! In general, it's definitely possible to contribute and you can see that there is quite a large number of contributors (it works through fork/pull requests). However, in case of new metrics (rather than fixes/adjustments to existing ones), we want to be careful to only add those that are either of direct need to our users or have been proven relevant in clinical literature. The rationale is that the already existing number of CGM metrics makes can make it very confusing as to which ones should be used over others in the absence of additional justification, a criticism we also got from some reviewers. Thus, if you have some new metrics in mind, it would be good to have references and justification to their need before integrating them as part of the iglu. Would it be possible for you to elaborate/provide references as to which features you have in mind? I am also happy to discuss offline. As with any updates, any additions would need to have complete documentation and pass the default R package checks as necessary steps of being merged to the main branch. — Reply to this email directly, view it on GitHub <#156 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEUC5BYZVEWB7VL5K2ZDEP3ZGXW6BAVCNFSM6AAAAABIYGHYWWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNJZGA2DINRRGI>. You are receiving this because you authored the thread.