Let's throw bioethics out the window for a few weeks and conduct a clinical trial. Forget blinding as well. All we keep is randomization and controls. Recruit 4000 juveniles with insulin-dependent type 1 diabetes. Randomize 1000 to receive normal medical treatment and care for such a disease (this is group A). Assign another 1000 to receive only intercessory prayer and faith healing (Group B). Assign the next 1000 to both medical treatment and faith/spiritual healing (Group C). The last 1000 kids don't get any kind of treatment at all (Group D).
Now what are the chances that after a month:
1. Children in group A will fare worse than B or D?
2. Those in B will be better off than in D?
3. There will be a significant difference in the conditions of those in A and C?