How to understand medical decisions

List, chase and check - hypothetico deductive reasoning

There seems to be two steps in the 'transparent' diagnostic process.  The first step is to consider a list of diagnostic possibilities.  This is often done by identifying a finding with a short list of diagnostic possibilities and considering this list, e.g. the list of causes of central chest pain.  The next step is to choose one of these diagnoses to chase.  This might be the diagnosis that occurs most often in patients with chest pain or a diagnosis that is dangerous and should be confirmed and treated without delay.  


Symptoms, examination findings or test results are then looked for that occur commonly in the diagnosis being chased and uncommonly or never if possible in at least one of the other diagnoses.  In this way an attempt is made to assemble a combination of findings that occurs commonly in the diagnosis being chased and rarely in the other possibilities.  This can be tested for by reasoning with by elimination.  


If it is not possible to show in this way that the probability of the diagnosis being 'chased' is high (or that it cannot be confirmed because some of the findings were not definitive or not sufficient), then another diagnosis is chosen and chased.  If it is not possible to confirm any of the diagnoses in the list of possibilities, then the most probable based on the findings and reasoning so far can be chosen and this chased be looking for further information, often by performing tests.  The tactics of this process and the knowledge required is described in the Oxford Handbook of Clinical Diagnosis.  The philosopher Karl Popper drew attention to the way that this reasoning is used in science.


Intersecting lists

It is often possible to get the same result by assembling a combination of findings that point to diagnosis in a passive way without 'guessing and chasing'.  For example, the list of possibilities associated with recent chest pain includes a recent coronary artery thrombosis.  A particular appearance on an ECG (called 'raised ST segments) can be caused by a recent coronary artery thrombosis and a weakness in the wall of the heart (called a ventricular aneurysm).  The only diagnosis that the chest pain and ECG change have in common is a coronary artery thrombosis.  If a doctor in the emergency room heard the story and saw the ECG, the combination would therefore point to the probable diagnosis of a recent coronary artery thrombosis.  The mathematical probability theory of this is explained in Chapter 13 of the 3rd edition of the Oxford Handbook of Clinical Diagnosis.  This may possibly explain the way that the brain performs pattern recognition. 


© Huw Llewelyn 2016