Gravar-mail: Impossibility to eliminate observer effect in the assessment of adherence in clinical trials