Hi peeps, I need help setting up a control group to evaluate the impact a new diabetes awareness and management course.
Some background info: the hospital offers diabetic patients with a learning program comprised of 5 continuous classes, with 3 months in between each class (pretest data taken at 1st class to establish baseline, 5th class for post-test knowledge). The hope is that patients who have participated in the class will have higher than average knowledge about diabetes and how to manage it, along with better biometric measures (bmi or weight, blood sugar or cholesterol level, etc.) The course is taught in group setting by a health educator and a doctor specializing in diabetes (the purpose of the group-setting is so that when patients see other patients asking questions, they could learn more from that than the usual 10-mins doctor consulatation). The hospital want to know if, independent of the course, would a similar group of diabetic patients achieve the same result?
My problem is that the participants in the experimental group (the group getting the education through the group seminars) isn't randomly selected. All diabetic patients at the hospital get referral to join the course through their primary physician, but only a few (about 15%) actually joined. So how would I set up a control group among those who didn't join the program to have a valid comparison?