Relative efficiencies of alternative preference-based designs for randomised trials

Area of Science:

• Clinical trial methodology
• Biostatistics
• Health services research

Background:

• Clinical trial outcomes can be influenced by patient treatment selection (selection effects) and receipt (preference effects).
• Conventional parallel group trials cannot evaluate these patient preference effects as preferences are not measured.
• Patient preferences are crucial for understanding treatment effectiveness and adherence.

Purpose of the Study:

• To discuss preference-based clinical trial designs for measuring selection and preference effects.
• To compare the potential of preference-based designs against conventional trials.
• To analyze factors influencing the efficiency of different preference-based designs.

Main Methods:

• Discussion of three preference-based designs: two-stage, fully randomized, and partially randomized.
• Comparison of estimable effects in conventional versus preference-based trials.
• Analysis of factors affecting the relative efficiency of preference-based designs.

Main Results:

• Preference-based designs offer the potential to estimate selection and preference effects, unlike conventional trials.
• The efficiency of these designs depends on participant indecision, preference rates, and randomization ratios.
• Different scenarios impact the advantages and disadvantages of each design.

Conclusions:

• Preference-based designs provide a more comprehensive evaluation of treatment effects by incorporating patient preferences.
• Understanding selection and preference effects is vital for improving clinical trial validity and generalizability.
• The choice of design depends on specific trial characteristics and research questions.