CLINICAL TRIALS
Experimental tests in general
- An experimental group is always compared to a control group.
E.g., animals or patients given a drug or receiving a treatment must be compared to other animals
or patients that do not receive the drug or treatment.
- Experimental and control groups must be comparable. They must be treated the same except for the one (independent) variable being tested.
- The experimental conditons should be repeatable in other places, by other experimenters.
- Sample sizes should be large enough to get statistically reliable results.
- Testing of results should be objective (unbiased). This means that the experiment should be conducted so that
observations that would agree with the hypothesis are possible, that observations that would falsify the hypothesis
are also possible, and that neither outcome is favored in the testing process or data collection.
- The study should be peer-reviewed and published.
Clinical tests specifically
- Placebo control: In order to ensure unbiased results, patients not receiving a drug being tested
should instead receive a placebo (inert substance) that looks and tastes like the substance being tested
but lacks the essential or active ingredient being tested.
The reason: patient conditions often improve if they think they are getting a medicine (the "placebo effect").
- Blind testing: Patients must not know if they are in the experimental treatment group or in the control group
(again, because of the placebo effect).
- Double-blind testing: The person giving the treatment to the patient should not know either whether the patient is receiving
the ingredient being tested or just a placebo, to avoid the possibility of emotionally encouraging some patients but not others.
|