Page 15/22
How would you allocate the drug amongst the thousands of patients with tuberculosis?
Mycobacterium tuberculosis (purple) in the lung |
- Option 1: treat the next one hundred patients and comparing them with historical controls.
- Option 2: give every second patient the drug and compare the outcome between treated and treated.
- Option 3: randomise the allocation of the drug and compare the outcome between treated and untreated.
|
Option 1 suffers from problems that the control patients may not be good matches for severity of disease and the historic treatment might have been so bad that anything would be better (remember bloodletting for fever).
Option 2 suffers from the problem that the doctors can bias the trial by controlling when they enter the patients. The most sick then get the new wonder drug so you do not have a good control group.
The MRC choose to do two trials:
