The most important assumption concerns our ratio's denominator. We need that the consumption of beef occurs in quantifiable servings, and, more importantly, we need that each of these servings is EQUALLY LIKELY to be consumed by a given consumer. Clearly this is a serious assumption. Some consumers eat only steaks cut from organically grown cattle, while others eat only sausage from fast food restaurants. Our estimate seems likely to be overestimating your risk if you consume only organic steaks and underestimating your risk if you consume only sausage from fast food restaurants. We are forced to view our number as the risk associated to some sort of "average consumer".
To estimate the numerator means we need to quantify how many servings were disease-inducing-servings of beef. This requires us to assume that cases of CJD associted to the epidemic were caught from a particular serving. It may be that several exposures are in fact necessary before catching the disease. Or it may be that the dissease is the result of some inherent inabiltiy of certain individuals' bodies to cope with this prion's presense, which was not really the "serving's fault". However, barring any additional medical data, this assumption seems reasonable.