Progress in therapeutics has not always arisen from controlled clinical trials. Chance observations have historically led to huge advances. Today's three most commonly used cardiovascular drugs are good examples: digoxin is a component of digitalis (famously reported by Withering after observing the treatment of a dropsical lady by a gypsy); aspirin is derived from the willow tree bark, first reported by the Reverend Edmund Brown to treat his own malarious fevers; and warfarin is the result of a University of Wisconsin investigation into a hemorrhagic disease of cattle. Lest we forget, Jen-ner's experiments would be ethically impossible today: they included deliberate exposure to smallpox, and aspirin is a drug that would probably fail in a modern preclinical toxicology program due to chromosomal breaks and gastrointestinal adverse effects due to systemic exposures in rodents. Modern clinical trials are therefore not necessarily the Holy Grail of therapeutic progress.
Statistical theory must also be held not only with respect but also with healthy skepticism (although this is really the subject of Chapter 19). It should be remembered that the development of statistics, as they have come to be applied to clinical trials, has arisen from a variety of non-mammalian biological sources. Experimental agriculture stimulated the early giants (Drs Fisher, Yates) to explore probability density functions. While epidemiological studies have confirmed much that is similar in human populations, it is unknown whether these probability density functions apply uniformly to all disease states. Any statistical test that we employ makes assumptions that are usually not stated.
Was this article helpful?