When I began arranging clinical trials, I was shocked to learn how long the concept had been around.  And, I bet you know the first trial, too.

Way back around 1750.   Remember all those British sailors succumbing to scurvy?  (Spontaneous bleeding, limb pain, and frequently, death.)  That’s when Dr. James Lind divided the sailors into six groups (2 each) and finagled their diets.  Lo and behold, those that were proffered fresh fruit never got scurvy.  Et, voila…

Oh, wait.  He didn’t really have a control group- the one getting a placebo.   Nor was the trial random- the fact is he might have chosen healthier subjects in his trial.

It wasn’t for another 100 years when these criteria were included in the clinical trial protocols.  Those became part and parcel of the study to discern if streptomycin would work to eradicate infections.

Now, we follow those sort of protocols, as we envelop more and more subjects in the trial (called phases I, II, and III) to ensure the safe and effective use of drugs and devices.

Except this system takes time.  And, time is often what we are lacking when things like a pandemic hits.  Moreover, the subjects in a trial have a 50/50 chance of getting no treatment (a placebo)- so it’s often difficult to enroll folks in clinical trials.

It wasn’t until testing of ECMO (ExtraCorporeal Membrane Oxygenators) that clinical trial design took a shift.  Dr. Robert Bartlett incorporated the nascent concept of game theory in this design.  First a baby underwent CPAP (continuous positive airway pressure) treatment.  Should that baby survive, the next child would undergo CPAP.  If the child died, the second test child would receive ECMO.  If that child survived, ECMO would again be used- until there was a problem.  There wasn’t- the ECMO treated babies survived. (Note that other physicians didn’t consider this a valid study- so ECMO didn’t become the norm for another decade, when ECMO was tested using randomized test regimens.)

Now, more and more of us rely on what is called platform trials.  That means we set up a system that can be repeatedly used- with a single control group being examined versus a plethora of test conditions and drugs.

Others try to avoid using a placebo.   Adaptive trials follow a similar protocol to that used by Dr. Bartlett (ECMO at the University of Michigan, described above).   Every test subject follows the rules developed before the trial begins.  That means if a drug is showing promise, subsequent tests continue using the promising drug, instead of splitting the group between drug and placebo.

Others still use a combination of platform and adaptive testing.   Which process had become acceptable by 2017 when Drs. Janet Woodcock (FDA head) and Lisa LaVange encouraged these protocols in an article in the New England Journal of Medicine.

As a matter of fact, it was an adaptive protocol that was used to determine a COVID-19 treatment in the UK.  (This was the RECOVERY trial, headed by Drs. Peter Horby and Martin Landray.)

The problem is that these sorts of trials require staff to be far more competent in statistics and analysis.  Teams have to be very large (along with PhD statisticians, study designers, etc.).   Trial replication is more difficult.  And, if long-term follow-up is required for the subjects, these protocols really do not provide the desired capabilities.

But, for things like vaccines and pandemics- they are probably our best choices.

Share this: