Doctors use AI to test new coronavirus treatments on patients

A new clinical trial uses AI to quickly identify promising coronavirus treatments — and quickly dismiss ineffective ones.

On April 9, the University of Pittsburgh Medical Center launched a new clinical trial designed to quickly test the efficacy of multiple coronavirus treatments on COVID-19 patients.

At the center of the trial is REMAP (randomized, embedded, multi-factorial, adaptive platform), a system that was originally designed to use artificial intelligence to identify effective treatments for severe pneumonia.

The system aims to allow doctors to treat patients with whatever is the best available option, without sacrificing scientific rigor needed to learn about effectiveness.

This is a unique approach to clinical research — and it’s one that could prove ideal in the setting of a pandemic like the coronavirus.

Testing Coronavirus Treatments

The need for effective coronavirus treatments is clear — less clear is the path we should take to identifying those treatments.

Typically, researchers looking for a new way to treat a disease would launch a clinical trial with a set number of participants.

Each participant would receive one of a small number of therapy options in a certain treatment category.

This is an unprecedented pandemic and we need an unprecedented response.

Derek Angus

Once that trial ended, the researchers might launch another trial to test a different category of treatment, and again, they’d enroll a set number of participants and subject each to one of several therapy options.

In the case of COVID-19, this might play out something like a trial to test two or three promising antiviral drugs, followed by a trial to test a couple of immunomodulators, followed by another trial of something else.

The problem with this traditional approach is that it takes a lot of time. But without clinical trials, we can’t know for sure that a new coronavirus treatment is effective.

The purpose of REMAP, then, is to help doctors identify the most effective coronavirus treatments far more quickly than they could through traditional clinical trials — without sacrificing the benefits of clinical testing.

The REMAP Trial

All 40 of UPMC’s hospital systems are participating in the REMAP trial, which is currently testing the efficacy of several coronavirus treatments, including hydroxychloroquine.

Any COVID-19 patient admitted to one of those hospital systems can join the trial.

Each participant will receive standard COVID-19 care, and all but 12.5% will be randomly assigned to receive between one and three of the experimental coronavirus treatments (within a few weeks, the number of people in that placebo arm will drop to 1%).

Based on the information doctors write in patients’ electronic health records, the REMAP platform will automatically adjust the coronavirus treatments it randomly assigns.

Single therapies or combinations that do well will be assigned to more new patients, while those that don’t will be removed from consideration altogether.

If a promising new coronavirus treatment emerges during the course of the trial, the researchers can simply add it as an amendment to the ongoing study.

“This allows us to always rapidly identify which treatment works best, while keeping the number of patients needed to achieve statistical significance low,” Derek Angus, an intensive care physician at the UPMC and the trial’s senior investigator, said in a news release.

“It also means we get the best treatment to the most patients right out of the gate,” he added.

The team aims to quickly scale up the trial to include more than 200 hospitals worldwide, and as more people join, the REMAP platform should get quicker in delivering its recommendations — and speed is of the utmost importance in the battle against the coronavirus.

“We must throw out old ways of thinking and fuse clinical care and clinical research into one extremely efficient system,” Angus said. “This is an unprecedented pandemic and we need an unprecedented response.”

The placenta may play a role in the genetic risk of schizophrenia
Researchers at Johns Hopkins have found that genes associated with schizophrenia risk may impact the placenta, not just the brain.
AI is changing how Americans find jobs, get promoted, and succeed at work
AI-driven software has the potential to advance quickly and change how companies make strategic decisions about their employees.
What is an AI black box? A computer scientist explains
AI black boxes refer to AI systems with internal workings that are invisible to us. What are the implications of working without transparency?
Gain-of-function research is more than just tweaking risky viruses
Gain-of-function experiments in the lab can help researchers get ahead of viruses naturally gaining the ability to infect people in the wild.
The radical drop in maternal mortality was a public health miracle
In 1758 in Sweden, 1205 mothers died for every 100,000 live births, which was likely representative of the global maternal mortality rate.
Up Next
fighting the coronavirus
Subscribe to Freethink for more great stories