Coming from a PhD in Physical Activity, Nutrition, and Wellness, fasted/non-fasted would be a serious research design issue. Usually human feeding studies have a fasting period prior to the blood draw or prior to a particular food item is consumed so other intake is not confounding results.
------------------------------
Jonathan Kurka, PhD
Senior Research Analyst
College of Health Solutions
Arizona State University
------------------------------
Original Message:
Sent: 05-01-2018 14:36
From: Edzard van Santen
Subject: Fasting and non-fasting cholesterol
Depends on the objective of the study. If it includes a comparison between fasting an no-fasting I would analyze it as a multi-source regression problem, i.e., add fasting status as a class variable to the analysis. Comparing regression coefficients via a contrast is easy
In God we trust, all others need to bring data.
=====================================
Edzard van Santen, PhD
Professor, Agronomy Department
Director, Statistical Consulting Unit
Institute of Food and Agricultural Sciences
PO Box 110500
Gainesville, FL 32611-0500
Phone: 352-392-3067
http://agronomy.ifas.ufl.edu/ =====================================
Original Message------
Dear ASA Colleagues:
Does anyone have suggestions for references on analyzing longitudinal cholesterol data, in which some people fasted and others did not? I am wondering whether including an indicator for fasting (1=yes, 0=no) would be considered acceptable or whether fasting and non-fasting need to analyzed separately. To make matters more complicated, some participants fasted at some times, but not at others.
The data are LDL, HDL, and Total cholesterol, and triglycerides. The time range is two years.
--
Regards,
Brandy R. Sinco
Statistician and Programmer/Analyst, UM School of Social Work
Current Projects:
Mon, Wed, Fri CHW Integration/REACH;
Tues, Thurs RISE/WCBT