|
|
Dose-Response
Studies (Continued)
Because
most responses to drugs vary with the logarithm of the dose,
investigators commonly choose doses in exponential ratios--e.g.
10 mg, 20 mg, and 40 mg. For reasons discussed in Section
3 of this chapter, one may wish to include a placebo
and standard analgesic control.
Rather
than proceeding in the two above steps, one may wish to
get initial efficacy and dose-response information from
the same patient group. One option is to use a prospective
dose-response study as the initial study. Because some patients
will be getting low, ineffective doses, this may be somewhat
less efficient than a "sledgehammer study" in
establishing the principle of drug efficacy. Some of this
power may be salvaged by carrying out a dose-response regression,
where a statistically significant regression coefficient
is generally viewed as evidence of efficacy. Another possibility
is to use a two-stage crossover trial in which patients
are initially titrated to their maximal tolerated dose (MTD)
or placebo, and responders are then randomly assigned to
various proportions (say 0, 25, 50, and 100%) of the MTD.
|