Vol 4 n° 4 - Drug Development
Past issues Contributors How to publish Contributions and comments Home
 
he treatment of schizophrenia represents one of the most difficult areas of medicine for carrying out reliable and informative clinical trials of new medicinal products. The methodological issues that affect studies of neu- roleptic agents are not unique. The evaluation of treat- ments for negative symptoms probably represents the most unusual methodological problem (not covered in this paper), but from a statistical perspective this prob- lem has parallels elsewhere in medicine.The real reason why clinical trials in schizophrenia are so difficult is the fact that a number of methodological issues are present together and in a severe form. This paper is concerned largely with trials that provide the confirmatory evidence of the efficacy of new medic- inal agents, that is those carried out during their phase 3 development or perhaps during the development of a new indication in phase 4. Hence, it is concerned only with controlled trials that provide the most reliable and informative evidence of efficacy for licensing decisions. General guidance on the statistical issues that arise in confirmatory trials and that relate to regulatory decisions can be found in ICH E9 (ICH, International Conference on   Harmonization   of  Technical   Requirements   for Registration  of  Pharmaceuticals  for  Human  Use).1 Guidance on the design, conduct, analysis, and interpre- tation of these trials in the field of schizophrenia can be found  in  the  Committee  for  Proprietary  Medicinal Products (CPMP) Note for guidance.2 Two issues require a broader introduction before dis- cussing their impact on trials in schizophrenia. The first is the use of placebo. The most recent revision of the Declaration of Helsinki3 in October 2000 caused alarm among those conducting and carrying out controlled clin- ical trials by appearing to limit the future role of placebo to a serious extent. The use of placebo in schizophrenia trials was already a problematic matter. Hence in the C l i n i c a l   r e s e a r c h 4 6 3 Study designs, duration, and choice of comparators including the use of placebo John A. Lewis, DSc Keywords:  neuroleptic;  controlled  trial;  methodology;  comparator;  placebo; duration; Declaration of Helsinki; relapse; recurrence Author  affiliations:  Statistical  and  Regulatory  Consultant,  Canterbury,  UK; Visiting  Professor  of  Medical  Statistics,  Department  of  Epidemiology  and Public Health, University of Leicester, Leicester, UK Address for correspondence: Prof J. A. Lewis, Hatch Green Cottage, Hatch Lane, Chartham Hatch, Canterbury, Kent, CT4 7LP, UK
(e-mail: johnlewis@clinitrial.co.uk) T This paper discusses some methodological issues that are relevant to the design of controlled trials of new medici- nal products for use in the treatment of schizophrenia. Two issues are covered more generally and at greater length. The first is the use of placebo. Recent debate of this topic was stimulated by changes to the Declaration of Helsinki. The second is the design of studies to evaluate maintenance treatment in the prevention of relapse and recurrence. With respect to both of these issues, specific implications for trials in schizophrenia are considered. Additional design topics addressed briefly are noninferi- ority designs, add-on designs, withdrawal designs, run-in periods on placebo, loss to follow-up, and short-term and long-term trials. Dialogues Clin Neurosci. 2002;4:463-469.