Randomized controlled trials are widely accepted as the best method for determining whether prescription drugs work. Typically, randomized controlled trials have two-arms, where one group of patients is randomly assigned to receive the study drug and the other group receives a placebo (a substance with no therapeutic effect). However, there are certain situations where these conventional two-arm trials cannot be used, such as in rare diseases where there not enough individuals with the condition. In this context, single-arm trials (i.e., one group that receives the new drug) are used to determine drug efficacy and safety, although these trials may be difficult to interpret without a placebo arm for comparison. As a solution, external controls have been proposed to emulate the placebo arm in a conventional two-arm trial. External controls are patients taken from a different setting, and can be historical (i.e., individuals with past treatment), or contemporaneous (i.e., individuals with alternate current treatments). To date, however, there is limited guidance on how to best select external controls. Thus, this study will use real-word data to emulate a randomized controlled trial to provide insight on how to best select external controls.
Randomized controlled trials are the gold standard for assessing the efficacy and safety of new drugs. However, there are certain situations where conventional two-arm trials are not conducted. This is common in the field of oncology, whereby single-arm trials have been used to investigate the efficacy and safety of new drugs in rare cancers. However, there is growing interest in the use of external controls (i.e., from a different setting) that can be used as comparator arms for such trials. To date, studies that have investigated the utility of external controls have focused on methods to control the confounding that is introduced when using real-world data. However, important knowledge gaps remain, like the temporality of selected controls in relation to the treatment arm, and how to select the control group itself. Thus, using an empirical example among patients with type 2 diabetes, we will investigate whether it is possible to use observational data to emulate results from a clinical trial. Specifically, we will investigate the effectiveness and safety of liraglutide, a second-to-third line treatment in the management of type 2 diabetes. The efficacy and safety of liraglutide was demonstrated in the LEADER trial, which will be used as a point of comparison. Using a cohort of patients with type 2 diabetes, we will emulate the LEADER trial by applying the same inclusion and exclusion criteria, selecting a treated group (i.e., users of liraglutide) and different comparator groups (i.e. users of other antidiabetic drugs). To address existing knowledge gaps in control selection, we will alter the time period for selection of controls, use different control groups and investigate various approaches to deal with confounding. This study will provide guidance on best practices for comparator group selection, which may help inform the design and analysis of randomized controlled trials using external controls.
This study has two primary outcomes. The effectiveness outcome will be defined as a composite of death from cardiovascular causes, hospital admission for a nonfatal myocardial infarction, or nonfatal stroke (ICD-10 codes listed in Appendix I).
The safety outcome will be defined by a hospital admission for a bile duct- or gallbladder-related condition (e.g., cholelithiasis, cholecystitis, cholangitis, gallstone pancreatitis, and other bile duct, and gallbladder disorders; ICD-10 codes listed in Appendix II).
Samy Suissa - Chief Investigator - Sir Mortimer B Davis Jewish General Hospital
Laurent Azoulay - Corresponding Applicant - McGill University
Devin Abrahami - Collaborator - McGill University
Elodie Baumfeld Andre - Collaborator - Roche
Hui Yin - Collaborator - Sir Mortimer B Davis Jewish General Hospital
Peter Honig - Collaborator - Pfizer Inc - US Headquarters
Richeek Pradhan - Collaborator - Sir Mortimer B Davis Jewish General Hospital
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation