The reason i say all that is that reducing early members of a cohort throws the final averages way off when meant to be part of a five year program and would make a positive finding almost impossible unless real cures took place.
The solution for researchers would be to discard data from early deaths as an outlier. In mining it is the nugget effect. Thus we have a protocol that likely kills of the weak sisters allowing the remaining healthier cohort to live longer on average. This provides positive results to support the phantom protocol.