[Statistical standards for clinical long-term studies].

Interest in the implementation of long-term clinical studies has grown considerably in Austria in recent years, partly encouraged by studies of this nature performed in West Germany, Great Britain and the United States. However, such studies have to be based on biometric standards that must be adhered to in the planning, implementation and reporting phases of long-term studies if they are to be regarded as authoritative and to find international acceptance. The present paper deals with such standards, based on the current level of methodological development as well as the authors' experience gathered over many years of participation in such projects. Special emphasis is placed on planning, random sampling and analysis, randomization, documentation, organization, termination of studies, reporting, multicentre investigations and cooperation between doctors and statisticians. The authors present several alternatives with and without the use of computers, attempt to explain to clinicians the role of the biostatistician in the supervision of long-term studies and offer their services to those interested in their cooperation.