[8/27, 7:49 AM] AW : Someone posted a pic of shoes on side table of hospital ward and how his prof. (I would call Teacher) made diagnosis with one more more item on the table without even seeing the patient.
The art of Diagnosis!
[8/27, 8:58 AM] Thanga Prabhu: What was diagnosis? I would say dangerous Prof teaching and setting role model the wrong way. We should ever be humble and accept: katrathu kadugalavu, kalathathu kadal alavu. Tamil proverb
[8/27, 9:08 AM] Vijayasimha Ajarananda: OMG! Totally, over my head.
[8/27, 9:16 AM] +974 6642 7543: Remembering Dr. House
[8/27, 9:20 AM] Rakesh Biswas: Very famously taught by Sherlock Holmes and I quote,
"SHERLOCK HOLMES:
"Let him, on meeting a fellow mortal, learn at a glance to distinguish the
history of the man and the trade or profession to
which he belongs. By a man's finger-nails, by his
coat-sleeve, by his *boot*, by his trouser knees, by
the callosities of his forefinger and thumb, by his
expressions, by his shirt cuffs-by each of these
things a man's calling is plainly
revealed"
JOHN H. WATSON, MD: "What ineffable twaddle!"
A. CONAN DOYLE
A Study in Scarlet
[8/27, 9:22 AM] Rakesh Biswas: "Inspection of the shoes is among the most re-
vealing and least used of all aspects of diagnostic
clothing analysis. As with other articles of cloth-
ing, shoes tell a story about their owner: work
shoes on an accident victim, for example, suggest
a different circumstance for an acute event than
slippers, dress shoes,- hunting boots or running
shoes. The presence of one slipper and one shoe
or of an open-toed shoe implies gout, trauma,
other arthritis or bunions on the unshod foot.
Shoes without laces, or laces undone, are more
comfortable to an edematous or inflamed foot.
Patients with Parkinson's disease or other motor
limitation may simply lack the dexterity and
flexibility to tie their shoelaces. A prosthetic shoe
lift is a more obvious clue to a chronically short-
ened leg than simple inspection of a supine pa-
tient. The pattern of wear on the soles of shoes
testifies to gait. For example, a rapid distinction
between old and new hemiparesis in a patient in
an emergency room can be achieved by examining
the shoes, in which the differential wear of a
long-standing limp is clear. Does a patient have
a backache? A glance at new or very high-heeled
shoes may solve the diagnostic mystery. In a
patient with diabetes, the source of sepsis may
be clarified by blood and serum stains seen on
socks or by the presence of ill-fitting shoes."
[8/27, 10:05 AM] Thanga Prabhu: Sir Dr Arthur Conan Doyle was a physician. Sherlock Holmes is his alter ego
[8/27, 10:05 AM] Thanga Prabhu: Elementary my dear Watson 😎
[8/27, 10:09 AM] AW : https://twitter.com/
[8/27, 10:11 AM] AW : Will HCIT be able to pickup these contexts and make sense of it?
That's y medicine is fuzzy and difficult to simulate.
[8/27, 10:14 AM] Sundar IAMI: Sure, these are good to arrive at a starting guess. Surely you cannot treat the patient based on such a diagnosis! So, how valuable is it when there are more deterministic methods?
[8/27, 10:16 AM] Rakesh Biswas: Left looks more worn out than the right?
Now read the Sherlock Holmes quote shared above!
[8/27, 10:17 AM] AW : With such methods one could come pretty close to 2-3 Diff. Diags. And investigate only for those reducing need for compressive workup every time
[8/27, 10:20 AM] Sundar IAMI: And evenly worn out. So probably the left leg is shorter than the right. Chronic. Should be complaining of knee problems in the left leg
[8/27, 10:20 AM] Sundar IAMI: Would affect his gait too. So the hip joint would also wear out unevenly
[8/27, 10:21 AM] Sundar IAMI: Laceless shoes - pedal oedema
[8/27, 10:22 AM] Sundar IAMI: Difficulty in bending down
[8/27, 11:20 AM] Rakesh Biswas: The most important and easiest way perhaps one can make a diagnosis is to capture all the event data points in a timeline that led to the current problem events that require a diagnosis.
Now one may need to look at the fact that "diagnosis" itself is nothing but a compressed version of all the data points that have been brought together and tied to form an "ontology."
The diagnosis could be something known in the past (often labeled static ontology but a more appropriate term could be current dominant ontology).
Again need to be careful with this word "ontology" as it unfortunately has many usages across the spectrum of knowledge! In simple terms an ontology depicts objects and it's ties and if objects are information-data, ontology are packets of bound knowledge (while epistemology is unbound although in it's popular usage it's often confused with ontology).
If no one gets the diagnosis, it's something new that changes past static ontology (that's the problem with this term, it can't be static if it can be changed @Metapsychist Number 1 Kims 2015) and once one is able to grow the diagnosis utilizing newer data points that till then may have been to quote, "from the static ontologies perspective, simply variations labeled statistical noise; from those discovering new ground, phenomena of central interest!"
More here: https://www.researchgate.net/
@Vijayasimha Ajarananda the author is a physicist at TIFR who is working on computational thinking aka problem formulation, abstraction, complex-systems thinking, a subset of asynchronous intelligence (father of today's AI)!
Coming to the main driver question from Sundar to the above Sunday stream of consciousness, our holy grail is reaching a singular diagnosis to the point where the treatment automatically becomes to the point but current medical ontologies are limited in terms of that as most of our diagnostic and therapeutic armamentarium have marginal efficacies at best and hidden harms and costs at worst!
[8/27, 11:32 AM] Sundar IAMI: A great post, Doc. But isn't diagnosis more than a relational set of various datapoints? It is a set that embeds inferences and even biases the author towards a few directions. In that sense it is a set of vectors, derived from the related set of datapoints. It also builds on the past inferences. I agree - ontology doesn't describe it well. The inference vectors are also characterised by likelihood estimations.
[8/27, 11:45 AM] Sundar IAMI: Of course it can. It has demonstrated far more complex tasks. The question is - is it worth it? Will a doctor be willing to take in such observations or inferences from another person and proceed with differential diagnosis? Whenever I diagnosed a complex system (in system engineering), I could trust someone else's input and diagnose the fault only when I was away from the site. I couldn't bring myself to do that when I was present at the site. So I am able to relate to the discomfort of doctors in diagnosing remotely.
No comments:
Post a Comment