[Roaming through methodology. XXIII. The need for randomization and blinding in therapeutic investigations]

After the start of a treatment patients often get better. This may be due to the treatment or other circumstances. There may be errors in the execution of the study, the disease may disappear spontaneously, effective co-interventions may be present and possibly there are also placebo effects. Most of the time, only sufficiently large randomised, blinded research is useful to estimate the efficacy of drugs and of other treatments. The best comparison is often between the new treatment and the best available one, not the sham treatment.