I am trying to calculate total AUC for pre and post intervention and considering individual AUC before looking at the sleep quality effects of an intervention in salivary melatonin and cortisol changes.
I briefed participants to take samples at set timepoints throughout the day where they gave 6 samples spread throughout the day before the intervention, and then another 6 samples 4-weeks later after the intervention (12 samples in total per participant).
Set standarised timepoints:
1st sample: upon waking
2nd sample: 30min following 1st sample
3rd sample: 3hours following 2nd sample
4th sample: 6hours following 3rd sample
5th sample: 30min before bedtime
6th sample: just before falling asleep
When calculating total AUC the formula for each trapezoid (which would be 5 trapezoids in this case to get total AUC):
area = ((a+b/2)*h)... where a&b are hormone values & h is the distance along the x axis = time diff between samples
trapezoid 1: (1st sample cortisol value + 2nd sample cortisol value/2)*h
However, with participant work/uni schedules participants did not always keep to set timepoints and rather had their own 'individual timepoints' i.e. the 3rd and 4th sample deviated considerably from the set 3h and 6h standardised schedule, where there was up to a 3h deviation, shifting the 4th sample closer to bedtime 5th & 6th.
my Q is .. does it matter that h is standardised or should this be based on individual timepoints (i.e. the pattern participants followed based on their schedule) which would change the AUC value as the 'h' could be 7h instead of 3h for e.g.
Any thoughts would be really appreciated, thankyou! :)
- Badges
- Science topic