I am estimating power for an experiment using AUC of blood glucose response over 120 minutes as the primary outcome. I'm struggling to find good quality reliability data for this, in particular I need an estimate of likely within-person variation. There's lots on GI but that isn't directly helpful.