Good Gawd man, they have to. The DSMB is charged with observing the progress of the study and intervening if something untoward and unexpected indicates that the study is harming the participants. It operates on the basis of feedback; the ongoing study produces data which confirms that the study is following the expected path or indicates that it is not. In Reduce It, events are the expected path.There is a projected rate and two groups rate. Events do NOT in and of themselves indicate V is causing harm. Only significantly (non parametric definition) more deaths/events than can be expected or markedly more than the placebo throw a flag. Then they intervene.
"the object of the exercise is proportion of events per group"
Jeeze. This is a mess. Can you imagine that deaths/rashes/strokes/whatever pile up and the DSMB says,"We had no idea. We just looked at the quarterly numbers and never looked for a trend over time. Were we supposed ta?
"It is watching for an anomalous distribution of deaths among the expected deaths."
Oh Well. Start with that the DSMB has the placebo/V group rates. In no way can they be restricted to a composite number