The architecture of diagnostic research

In my opinion, Bayes Theorm has little to do in shedding light
towards resolving a clinical problem. I do not know why Bayes Theorem was
ever introduced into clinical practice. Bayes originally was interested in
a chance event and its frequency of its occurence. Its application to
biomedical tests to arrive
at a reasonable diagnostic clinical problem is not exactly correct,
because if the physician were to individualize each clinical patient
clinically with a good history and physical exam and certain basic
pertinent lab tests and with patient follow up of his clinical findings
then let us say it is a problem of left ventricular failure, he would
arrive at these conclusions: No left Ventricular failure,
mild left ventricular failure to moderate or severe heart failure. The
base is the clinical diagnosis and the supportive evidence would be
finding a mild to elevated BNP
if it is really indicated to do the test?

Individualation of each patient problem clinically would obviate the use
of Bayes Theorem and meta analysis of series of patients. In conclusion
though I do not agree with the authors, I commend them for giving the
reader a good statistical exercise.