Chapter 4: Randomized-Controlled Trials

CHAPTER OBJECTIVES

• Describe the characteristics of randomized-controlled trials

• Discuss validity issues associated with randomized-controlled trials

• Describe common randomized-controlled designs (parallel, crossover, adaptive)

• Briefly explain key analytical aspects of randomized-controlled trials

• Discuss the strengths and weaknesses of randomized-controlled trials

KEY TERMINOLOGY

• Absolute risk difference

• Active control

• Adaptive designs

• Allocation concealment

• Ascertainment bias

• Attrition bias

• Bias

• Blinding

• Block randomization

• Carryover effect

• Clinical research protocol

• Cluster randomization design

• Crossover design

• Data and Safety Monitoring Board

• Detection bias

• Double-blind trial

• Drug effectiveness

• Drug efficacy

• Effect size

• Exclusion criteria

• External validity

• Factorial randomized trials

• Hawthorne effect

• Historical control

• History bias

• Inclusion criteria

• Instrumentation bias

• Intent-to-treat analysis

• Interim analysis

• Internal validity

• Investigator bias

• Maturation bias

• Noninferiority trial

• Number needed to harm

• Number needed to treat

• Open label

• Parallel study design

• Per-protocol analysis

• Placebo

• Power

• Primary outcome

• Randomization

• Regression to the mean

• Relative risk

• Relative risk difference

• Risk difference

• Sample size

• Selection bias

• Simple randomization

• Single blind trial

• Stratified randomization

• Study sample

• Subgroup analysis

• Surrogate endpoints

• Target population

• Testing

• To the mean

• Validity

• Washout period

INTRODUCTION

Randomized-controlled trials (RCTs) are widely regarded as the strongest type of primary study design to support evidence-based medicine and practices related to the beneficial and detrimental effects of novel drug therapies.^1–3 The goal of these trials is to measure a primary outcome in a highly selected group of individuals, or study participants, that are given an equal chance of being assigned to receive one or more clinical interventions. In medicine, interventions may include drug therapies, prevention strategies, or medical procedures, and may occur in a variety of settings in which healthcare is provided, including academic settings. Most RCTs are designed to determine the effect of a specific intervention on health-related outcomes, including disease prevention and progression.

The RCT is the most common type of trial that is used to determine the efficacy of an experimental intervention compared to a standard therapy or placebo. A key requirement by the Food and Drug Administration (FDA) in the clinical drug development cycle is the RCT as part of a new drug application (NDA).⁴ They are often referred to as Phase III studies or “pivotal trials,” since they are used to establish the relative efficacy and safety of an investigational drug compared to a control group. Adherence to very strict selection criteria, proper trial design, and minimization of bias are all critically important to establish the relative efficacy and safety of a drug. This chapter will review the most common characteristics of RCTs, including common variations on study design, sampling strategies, biases, and errors. This chapter will also address different ways to analyze the study data, including intention to treat and subgroup analysis.

MAXIMIZING VALIDITY AND MINIMIZING BIAS

Few, if any, clinical studies, including RCTs, are designed perfectly, despite having a highly qualified and dedicated research team assigned to conduct them. Therefore, ...