Thread by @EricTopol

A little primer on independent Data and Safety Monitoring Boards (DSMB), who will be reviewing the #SARSCoV2 vaccine trials to adjudicate safety and efficacy, making recommendations to the clinical trialists 1/

Consists of a multidisciplinary group of several members: clinical trialists, biostatistician, bioethicist, and experts in topic (virologist, immunologist, epidemiologists, vaccinologists)
The timing/schedule of data review usually based on enrollment 2/

The data are teed up by the biostatisticians. The DSMB is blinded to vaccine or placebo groups. There is ordinarily a time lag/gap from the "cleaned" dataset reviewed by DSMB that is behind ongoing events 3/

Review of safety is relatively straightforward with a thorough discussion of any serious adverse event. Efficacy is more complicated and there are pre-determined "stopping rules" for futility or overwhelming efficacy. 4/

The latter "overwhelming efficacy" is an extreme call that is being discussed now for an October vaccine approval. There are many issues and DSMBs *rarely* recommend to stop an ongoing clinical trial, a natural experiment 5/

Statistical penalty; difference in (P value) threshold for stopping early: an early look at data by DSMB is a "cost" to statistical significance. That's difficult to meet.
The reason: mid-stream interruption of a trial might, by chance, be at a moment of exaggerated benefit 6/

I hope it's clear that the stakes for stopping a Phase 3 #SARSCoV2 vaccine trial prematurely, based upon a limited number of events (e.g. 100, as has been suggested), extrapolated to vaccinating hundreds of millions of people, are exceptionally high 7/

There's also the safety issue that will likely take longer to resolve, because the question is about a low incidence event—untoward immune response to infections. At what %?
Preventing infections is the goal, but that this is achieved w/o or w/ very low serious adverse effects 8/