Design and analysis of trials of combination therapies

Combination therapies can have significant advantages over monotherapies. Combinations of therapies can provide additive (or even synergistic) effects on efficacy. They may permit use of lower doses of each component to achieve a given level of efficacy, improving tolerability and reducing adverse effects. A multicomponent treatment may facilitate tailoring of therapy to the needs of individual patients (e.g., treatment augmentation in nonresponders). These characteristics seem highly attractive in developing treatment strategies for alcohol abuse and dependence, because existing monotherapies have shown modest efficacy, at best.

However, trials of combination therapies present challenges in design, execution and interpretation, including: (1) choice of the treatment combinations to be compared; (2) definition of primary and secondary hypotheses; (3) differences between interventions in the duration of treatment, the time lag from the start of treatment to an observable effect on outcomes and interval for assessment of efficacy; (4) study power/sample size; (5) logistics of treatment delivery, masking and outcome assessment; and (6) attribution of adverse events.

Most of these issues arose in the COMBINE project, a sequence of trials intended to explore the use of combinations of behavioral and pharmacological approaches in the treatment of alcohol dependence. The resolution and impact of the challenges above for the COMBINE trial will be described.

Trials of combination therapies address many important clinical questions; however, their level of complexity requires considerable forethought, pilot investigations and organization.