General concepts in biostatistics and clinical epidemiology: Experimental studies with randomized clinical trial design

• Main Authors: Sebastián Estrada, Marcelo Arancibia, Jana Stojanova, Cristian Papuzinski
• Format: Article
• Language: English
• Published: Medwave Estudios Limitada 2020-04-01
• Series: Medwave
• Subjects: clinical trial; therapeutics; relative risk; bias; epidemiology; biostatistics
• Online Access: https://www.medwave.cl/link.cgi/Medwave/Revisiones/MetodInvestReport/7869.act

In experimental studies, researchers apply an intervention to a group of study participants and analyze the effects over a future or “prospective” timeline. The prospective nature of these types of studies allows for the determination of causal relationships, but the interventions they are based on require rigorous bioethical evaluation, approval from an ethics committee, and registration of the study protocol prior to implementation. Experimental research includes clinical and preclinical testing of a novel intervention or therapy at different phases of development. The main objective of clinical trials is to evaluate an intervention’s efficacy and safety. Conventional clinical trials are blinded, randomized, and controlled, meaning that participants are randomly assigned to either the study intervention group or a comparator (a “control” group exposed to a placebo intervention or another non-placebo or “active” intervention—or not exposed to any intervention) to reduce selection and confounding biases, and researchers are also unaware of the type of intervention being applied. Intention-to-treat analysis (inclusion of all originally randomized subjects) should be done to avoid the effects of attrition (dropout) and crossover (variance in the exposure or treatment over time). A quasi-experimental design and external controls may also be used. Metrics used to measure the magnitude of effects include relative risk, absolute and relative risk reductions, and numbers needed to treat and harm. Confounding factors are controlled by randomization. Other types of bias to consider are selection, performance, detection, and reporting. This review is the fifth of a methodological series on general concepts in biostatistics and clinical epidemiology developed by the Chair of Scientific Research Methodology at the School of Medicine, University of Valparaíso, Chile. It describes general theoretical concepts related to randomized clinical trials and other experimental studies in humans, including fundamental elements, historical development, bioethical issues, structure, design, association measures, biases, and reporting guidelines. Factors that should be considered in the execution and evaluation of a clinical trial are also covered.