Developing new statistical methodology harnessing nonconcurrent control data in infectious disease platform trials

Lead Investigator: Akihiro Hirakawa, Tokyo Medical and Dental University
Title of Proposal Research: Developing new statistical methodology harnessing nonconcurrent control data in infectious disease platform trials
Vivli Data Request: 9907
Funding Source: This work is partially supported by Japan Agency for Medical Research and Development grant JP24mk0121268.
Potential Conflicts of Interest: None

Summary of the Proposed Research:

COVID-19 is a sickness caused by a virus called coronavirus. So far, more than 200 million people have been affected by COVID-19. Anyone can catch COVID-19, but some people might get sicker than others, especially older adults or those with other health problems. That’s why it’s important to find treatments and ways to stop the virus from spreading.
In recent years, scientists have been working on ways to make clinical trials faster and more efficient. Imagine a clinical trial like a big test to see if a medicine works. Normally, each test focuses on just one question. But now, there’s something called a master protocol. It’s like a super test that can answer lots of questions at once. This is really important during times like the COVID-19 pandemic when we need to find treatments quickly.
There is a form of master protocol known as a platform trial. It’s like a big testing ground where different treatments for COVID-19 are compared to each other using one common group of people who don’t get any treatment (this is called the control group). In the control group, a placebo is often used. A placebo is a fake treatment that looks like the real one but doesn’t actually have any active ingredients. When we add new treatments in platform trials, we need to make sure the control group is fair. Sometimes, we divide this control group into two parts: one part gets picked at the same time as the treatment group (this is called the concurrent control), and the other part was picked before the treatment was added (this is called the non-concurrent control). The use of non-concurrent control data increases the amount of data available and helps to confirm whether a new treatment really works. However, there’s a problem. Sometimes, the features of people in the control group change over time, or the place where the trial is happening changes (this is called the time trend). This can mess up our results and make it hard to tell if a treatment really works. So, we need better ways to incorporate the data of the non-concurrent control group.
We’re going to look at data from the patients receiving a placebo of past trials for COVID-19 and see if there were changes over time in disease symptoms or death rate. We’ll use this information to create new ways to analyze the data from future trials. Our research is very important because it helps doctors find better ways to treat people with COVID-19. Moreover, by creating new ways to analyze the data from platform trials, our research can make future clinical trials even more efficient.

Requested Studies:

Colchicine Coronavirus SARS-CoV2 Trial (COLCORONA)
Data Contributor: Montreal Heart Institute
Study ID: NCT04322682
Sponsor ID: MHIPS-2020-001

A Randomized, Double-Blind, Placebo-Controlled, Parallel-Group Phase 3 Study of Baricitinib in Patients With COVID-19 Infection
Data Contributor: Lilly
Study ID: NCT04421027
Sponsor ID: 17830

Randomized Master Protocol for Immune Modulators for Treating COVID-19
Data Contributor: NIAID (a data-sharing platform funded by the National Institutes of Health)
Study ID: NCT04593940
Sponsor ID: Pro00106301