One of the puzzles of medicine is, to what extent can we apply the scientific method ?
In trying to answer this question, we are faced with a puzzle. Why is it that the basics of the scientific method are often subject to confusion by even those who day to day work inlcudes scientific research ?
The answer, I believe, lies with the true conceptual difficulty of understanding the scientific method.
For instance, if one is to look for a philosophical basis for the scientifc method, we discover that even the most skilled philosophers of the past struggle. Sartre had little to say on the subject. Neitzche and Camus, next to nothing. Even Kant was only able to pursue his approach to the scientifc method haltingly.
Instead, most of what we can deduce about the scientific method can be found in those whose work is either non - philosophical, or whose philosphical efforts were halted by premature death. The former, we have as an example Pasteur, and the latter, Freud.
For the moment, then, beginning with Pasteur and Freud, we can examine how they proceded, implicitly.
Both were concerned with developing a mechanistic model that provided a means of not only testing hypotheses, but determining what hypotheses to test.
A scientifc hypothesis, in other words, before testing, must meet certain criteria. Simply arguing for a given point of view, and then proposing a means of testing it, does not constitute the scientific method.
Einstein showed how important this approach is. Einstein based his hypotheses upon the notion that not only should a hypothesis be testable, but it should be internally consistent, logically, and it should be consitent with other, better described, proven, testable, and accepted hypotheses (which then at that point, become accepted fact.
For instance, the laws of gravitation meet Einstein's hypotheses, as being valid full reporesentations under limited conditions (certain variables having certain values).
So, the scientific method is not simply the generaization of an empiricism that is devoted to an isolated case.
(1) The hypothesis must be consistent with current scientific knowledge.
For medicine, that means that any hypothesis must not only be consistent with current biochemistry and organic chemistry, but that the consistency must have mathematical consistency.
(2) The hypothesis must be testable.
(3) The hypothesis must have implications that lead to other hypotheses that are testable - in other words, the predictive power of the hypothesis must be expansive rather than restrictive.
(3a) The hypothesis must in other words, be not only testable in its simple form, but it must become implicitly testable in those matters which depend upon its veracity.
(3b) And, findings on testing must clarify questions related to both its internal structure (fine detail hypotheses) and findings must indicate what and how broader predictions are testable (either by implying the larger issue, or directing how to properly test the larger issue).
(4) The hypothesis must have an internal logical cosnsitency.
An example of this is Special Relativity. It must be testable and verifiable in all frames, including intertial and accelerating frames.
An alternative is to demonstrate logical consistency using numerical methods. In other words, numerical proofs are logical proofs. One small subset of this argument is that we use statistical methods to demonstate mathemathical consistency in the testing of hypotheses that have no posited internal logical structure (most typically, this is seen in medicine testing).
(5) Flaws in a hypothesis must be present in such a way that they are evident before testing, and simplifying assumptions clearly stated, so that refinements can be trasnparently pursued, with internal mathematical and logical consistency.
(6) The variables or parameters must be clearly stated, and testing must include the domain of all of the variables. An example of success in such a a hypothesis is QCD.