3 ms·
It's not like we need a special control group if that control group is supposed to receive no treatment - we know rather well how untreated Ebola works and the
by PeterisP 12y ago
It's not like we need a special control group if that control group is supposed to receive no treatment - we know rather well how untreated Ebola works and the mortality rate.
If we try out drug X and only 50% of those people die, then we know that drug X is good for Ebola. Once there is drug Y to try, then we can use drug X for a control group.
- gus_massa 12y agoYou must be sure that you are measuring the same mix of virus strains, that the patients are receiving the same medical care and food, that the hygienic conditions are the same (the patient was going to survive, but a diarrhea just killed him), room conditions (is the room temperature important?). You must be sure that you are using the same criteria to count the patients. Are you going to count the almost dead patients as a drug failure, or drop it from the list? Are you going to give the drug to suspected cases or only to confirmed cases? Do you count small children (they usually have a smaller survival rate)? Are you counting the people that die before going to the hospital? Are you counting the (few) people that survive without going to the hospital? If you get a 30% cure rate, (and suppose that the standard cure rate is 20%), it's very difficult to be sure that the improvement is in the mew drug and not in a change in all the other factors or some unwanted statistical trick. Most of the rules for double blind studies are there because in the past there have been many "successful" test of new drugs, that were later retracted. It's very easy to measure incorrectly, especially when you really want the experiment to be a success. (Another rules are there to prevent the consequences of previous "unsuccessful" test, were the "cure" was much worse than the illness.)
- gfodor 12y agoPresumably the idea wouldn't be to prove the effectiveness of the 'cure' but an example of compassionate use.