One mental tool I've found very useful is the 2x2 table (or 3x3, or 4x4, for that matter) to work out possible permutations and combinations of clinical signs and symptoms. I remember one from my undergraduate surgery days in approaching RHC pain for instance:
RHC pain is of hepatobiliary or pulmonary origin. And the hepatobiliary causes come either with or without jaundice, or with or without fever. So given these two dichotomous parameters - we have a 2x2 - and its quadrantic (4) subgroups of differentials.
Fever No Fever Jaundice (1)Cholangitis (2)Choledocholithiasis CholangioCA, CA head of pancreas No Jaundice (3)Cholecystitis (4)Biliary stasis Empyema Mucocele
We can add on more variables as we go along, like palpable lump or no palpable lump. Which should make 2x2x2 = 8 sub-groups. Where L=lump, NL=No lump, we now have:
Fever No Fever Jandice (1a)NL: Cholangitis (2a)NL: Choledocholithiasis (1b)L: ? (2b)L: CholangioCA, CA head of pancreas No Jaundice (3a)NL: Cholecystitis (4a)NL: Biliary stasis (3b)L: Empyema (4b)L: Mucocele
Looking at the above, there is yet no obvious answer for combination 1b - fever+jaundice+lump. Unless, of course, when we factor in the confounding 'dual-pathology'. In that case we can have four possible combinations that result in fever, jaundice and a lump: 1a + 4b (cholangitis with a mucocele), 2a + 3b (choledocholithiatis with an empyema), 2b + 3a (CA + cholecystitis), 1a + 2b (cholangitis with CA).
Confusing? Let's not even get into triple pathology!
One can do this for everything in medicine; for instance in lung examination , taking into consideration dull/resonant percussion, increased/reduced fremitus, tracheal deviation away/towards lesion, increased/reduced breath sounds, and presence/absence of crepitations. That gives us 2x2x2x2x2 = 32 possible permutations.
In ENT, an easy one is the Rinne and Weber's test. Rinne +ve/-ve, Weber towards/away from the bad ear gives us a nice 2x2 for unilateral ear pathology. Drawing a two by two for the BAD EAR, we get: