Indicate valid data days

[Julia]

Hi all,

I have collected geneactiv data from patients that worn the watch for 7 days.

In the GGIR script (version 3.1.0) part 2 and 4, I used  includedaycrit = 16, to define the number of minimum valid hours in a day to attempt physical activity (part 2) and sleep (part 4) analysis.

In the daysum file, it still contains patients with days that consists of 100% nonwear (nonwear_perc_day_spt). These days therefore have less than the defined 16 hours / day to attempt PA and sleep analysis, I would guess.  

Question 1: How is it possible that these days were not filtered out of the dataset by the ' includedaycrit = 16' criteria? 

Question 2: does this mean that for patients with less than 16 hours a day, the physical activity and sleep characteristics are based on imputed data? 

Question 3: are there any recommendation to set a filter for minimal % wear to indicate a valid day for physical activity measures? I can imagine that if nonwear_perc_day = >90% it might be preferable to not include this day in your analysis(for example to calculate the mean parameters per week) for physical activity, but if nonwear_perc_day is ~20% is would be preferable to include this day and make use of the imputations to indicate PA levels during that day. 

For non-wear during the nights, I would guess that filtering on 'cleaningcode < 2 ' would be sufficient. 

Thank you in advance!

Kind regards,

Julia 

[Vincent van Hees]

Hi Julia,

Yes, I am afraid this is a bug in part 5 report not correctly applying the inclusion criteria (bug was introduced February this year), which I fixed but have not merged into a new release yet. Will do it in the upcoming days.

Thanks,

Vincent

--
You received this message because you are subscribed to the Google Groups "R package GGIR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to RpackageGGIR...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/5a0ac1ae-4af3-4753-862f-631f12aa7a96n%40googlegroups.com.