Review article: Primer for clinical researchers on innovative trial designs for emergency medicine

Abstract

Randomised trials have long been recognised as the gold standard research tool for evidence-based medicine. The past decade has seen the emergence of several innovative trial designs that are revolutionising how trials are conducted. These innovative designs enable more efficient, pragmatic trials that can address complex research questions which were previously not possible. In this paper, we provide an overview of the key innovative designs that are likely to be useful in the emergency medicine context, namely cluster crossover and stepped wedge designs, sequential multiple assignment randomised trial (SMART) designs, and platform trials. We describe the main features of each design, outline their pros and cons, and describe when they may or may not be useful. We also provide examples of these innovative designs in contexts that are relevant to emergency medicine.

Keywords: adaptive trial; cluster randomised; platform trial; randomised trial.

Figure 1

Schematic of innovative trial designs. (a) Cluster crossover trial with two intervention arms and two periods. (b) Stepped wedge trial with four clusters and five periods. (c) Sequential multiple assignment randomised trial with two stages, where non‐responders are re‐randomised to intervention or control. (d) Platform trial with three domains*. R, randomisation. *The platform starts with two domains. Domain 1 initially randomises participants between three arms. A new intervention (Intervention 3) is added to Domain 1 in Stage 2, and one intervention is stopped (Intervention 1) and another new intervention (Intervention 4) is added in Stage 3. Domain 2 randomises participants between two arms throughout the trial. A new domain (Domain 3) is added in Stage 2 and randomises participants between two arms until the end of the trial.