STATISTICAL THEORY AND LIFE CONSEQUENCES.

STATISTICAL THEORY AND LIFE CONSEQUENCES. This article illustrates the split between Bayesians and frequentists. A great deal of attention has been paid during the last forty years to measuring and trying to reduce bad cholesterol” (LDL), which is thought to contribute to heart disease. A new drug has been shown to be effective in reducing LDL, but has not been shown to meet the “gold standard” of reducing mortality. Is this enough for the Federal Drug Administration to permit the use of the drug? Can the FDA assume that a reduction in LDL will reduce mortality? The article also gives the example of a hypothetical new cancer drug which reduces the size of tumors, but has not been shown to reduce mortality. Should the FDA allow this drug to be used? Because evidence of the effect of a drug in extending human life requires studies of considerable length, adopting the frequentist view means that new drugs will be introduced with a considerable time lag. When the life of a loved one is at stake, perhaps you don’t want to wait for a gold standard review.