Blends of statistical designs for dose-response studies
Titel:
Blends of statistical designs for dose-response studies
Auteur:
Hedayat, A.S. George, Chi Y.H. Mahjoob, Kooros
Verschenen in:
Communications in statistics
Paginering:
Jaargang 23 (1994) nr. 2 pagina's 341-360
Jaar:
1994
Inhoud:
It is becoming increasingly apparent that for reasons of safety, efficacy and cost, accurate and useful dose-response/dose-titration information should be provided by clinical trials and included in the drug product labeling. In this exposition, we discuss some design related issues and focus our attention on trial designs primarily meant to gather dose-response or dose-titration information. We consider blends of repeated measurements, parallel-group, and titration designs, a titration design with variable treatment duration, and factorial designs for studying combination drugs. Our goal is to search for optimal and efficient clinical trial designs that will address as many of these clinical questions as possible and at the same time provide supportive evidence on the efficacy of the drug. Although our discussion is very general, we are cognizant of the limitations of these designs and the complexity of their associated analyses in practice, and hope that they might prove to be applicable and useful in some situations.