This JAMA Guide to Statistics and Methods discusses the use of dose-response models in phase 2 studies for identifying promising doses to use in further development of a new drug.
Clinical trials in drug development are commonly divided into 3 categories or phases. The first phase aims to find the range of doses of potential clinical use, usually by identifying the maximum tolerated dose. The second phase aims to find doses that demonstrate promising efficacy with acceptable safety. The third phase aims to confirm the benefit previously found in the second phase using clinically meaningful end points and to demonstrate safety more definitively.
Dose-finding trials—studies conducted to identify the most promising doses or doses to use in later studies—are a key part of the second phase and are intended to answer the dual questions of whether future development is warranted and what dose or doses should be used. If too high a dose is chosen, adverse effects in later confirmatory phase 3 trials may threaten the development program. If too low a dose is chosen, the treatment effect may be too small to yield a positive confirmatory trial and gain approval from a regulatory agency. A well-designed dose-finding trial is able to establish the optimal dose of a medication and facilitate the decision to proceed with a phase 3 trial.
Selection of a dose for further testing requires an understanding of the relationships between dose and both efficacy and safety. These relationships can be assessed by comparing the data from each dose group with placebo, or with the other doses, in a series of pairwise comparisons. This approach is prone to both false-negative and false-positive results because of the large number of statistical comparisons and the relatively small number of patients receiving each dose. These risks can be mitigated by combining data from patients receiving multiple active doses into a single treatment group for comparison with placebo (“pooling”), but only if it is possible to reliably predict which doses are likely to be effective.
In general, dose-response relationships are best examined through dose-response models that make flexible, justifiable assumptions about the potential dose-response relationships and allow the integration of information from all doses used in the trial. This can reduce the risk of both false-negative and false-positive results; incorporating all data into the estimates of efficacy and safety for each dose produces more accurate estimates than evaluating the response to each dose separately.
In an example published in JAMA, Gheorghiade et al1 reported the results of the SOCRATES-REDUCED trial, a randomized placebo-controlled dose-finding clinical trial investigating 4 different target doses of vericiguat for patients with worsening chronic heart failure, with the primary outcome being a reduction in log-transformed level of N-terminal pro-B-type natriuretic peptide. The primary approach to analyzing the dose response, combining the data from patients allocated to the 3 highest target doses (pooling) ...