Claude Bernard and Experimental Medicine
Claude Bernard is regarded as one of the first physicians, surgeons, to embrace scientific experimentation as a means of defining medicine. He believed that people who conducted statistical experimentation and stated statistically derived numbers without a definite purpose were in error. His belief was that there should always be a definite article that the experiment was looking for. He gave several examples of his belief in experimentation toward a goal in which he was able to relate his idea of the ridiculousness of just spouting numbers for their sakes. He relates the study of spinal root nerves that found that sometimes they were sensitive and other times they were not. He argued that this experimentation yielded nothing of value because it specified nothing.[footnoteRef:1] He next put forth the example of an individual who conducted a series of operations for the same condition and said that they resulted in a 40% mortality rate.[footnoteRef:2] Bernard again contended that this result meant nothing because it did not say under what conditions the person would be expected to live or die upon receiving the operation. Thus, his biggest complaint was that the so-called medical scientists did not provide true scientific findings that could either be verified or refuted. Their method was lacking. [1: Claude Bernard, An Introduction to the Study of Experimental Medicine, translated by Henry Copley Green (New York: Dover Publications, 1957), 137.] [2: Ibid]
It is the tendency of people to believe in their own observations and trust the statistics that they have compiled. For example, the doctor who gave a 40% fatality rate for his particular surgery may have understood what made one patient live and one die, but he did not share this with the rest of the world, at least not according to Bernard. This misuse of statistics occurs currently just as much as it did in the past. One researcher gave the example of a recent United States census that found that there was a "high correlation between location of churches and violent crime."[footnoteRef:3] Whereas it is a surety in most people's minds that this fact does not exist in reality, the correlation can still be made. The reason for...
It is poor science not to look farther and discover exactly what is going on to make the statistic seem to work. The reality is that there are a great number of churches, and that many different occurrences, such as violent crime location can be related to the location of churches because they are prolific. Bernard was saying much the same thing. He looked at the so-called scientific experiments that were being done with clinical cases and realized that because doctors were not applying the correct rules of scientific inquiry they were reaching false conclusion, or, at the very least, conclusions that meant exactly nothing when applied. [3: Rahul Dodhia, "Misuse of Statistics," Raven Analytics (accessed Nov 13, 2012) ]
His first argument is that statistics cannot be applied to medicine because it should be what he calls a "determinate" rather than an "indeterminate" science.[footnoteRef:4] He says that he realizes that much of medicine is conjectural, but that physicians should be working toward making it less so. The way that they can do that is with scientific experimentation rather than by believing that medical practice is somehow an art form.[footnoteRef:5] He gives another argument which is the use of the law of large numbers.[footnoteRef:6] His statement regarding this practice is that it is fine in an indeterminate science because they need large numbers to make any sense of what they are looking at many times. It is the argument against a small sample size that they are making, and this applies to most types of science where the numbers are large and can be accessed. However, he says that this does not apply, to any great degree, to medicine. Most of the time a physician is not dealing with a large number of cases. He or she must examine what is available to them and not be afraid to use this experimentation just because there are not a large number of patients with a particular disease of disorder. [4: Claude Bernard, An Introduction to the Study of Experimental Medicine, translated by Henry Copley Green (New York: Dover Publications, 1957), 139.] [5: Ibid.] [6: Ibid, 138.]
Bernard makes the…
