Talk:Differential diagnosis

External links modified
Hello fellow Wikipedians,

I have just modified 2 one external links on Differential diagnosis. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20090210153952/http://gateway.nlm.nih.gov:80/MeetingAbstracts/ma?f=102248792.html to http://gateway.nlm.nih.gov/MeetingAbstracts/ma?f=102248792.html
 * Added archive https://web.archive.org/web/20081002094931/http://en.diagnosispro.com:80/ to http://en.diagnosispro.com/

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 00:06, 13 December 2016 (UTC)

In Popular Culture
If someone more skilled than me can find the sources, it might be nice to add a section for pop culture where DDx has come up. The American television show House comes to mind, for example. 148.87.23.18 (talk) 04:59, 1 March 2019 (UTC)

one top hat, many rabbits in dense cohabitation
This method may employ algorithms, akin to the process of elimination, or at least a process of obtaining information that shrinks the "probabilities" of candidate conditions to negligible levels, by using evidence such as symptoms, patient history, and medical knowledge to adjust epistemic confidences in the mind of the diagnostician (or, for computerized or computer-assisted diagnosis, the software of the system).

One simply can't skim that sentence in hasty review, so for my own notes, I attempted to expand the sentence into note form:

This method may employ that shrink the "probabilities" of candidate conditions to negligible levels, by using evidence such as: to adjust epistemic confidences
 * algorithms akin to the process of elimination
 * (or at least) process of obtaining information
 * symptoms
 * patient history
 * medical knowledge
 * in the mind of the diagnostician
 * software of the system — per computer-assisted diagnosis

Even in this format, "computerized or computer-assisted diagnosis" made my eyes so rheumy, one or the other had to go.

In doing this, I managed to reunite subject and predicate (bolded in first quotation), only to discover a numeric disagreement. Not a good sign that this was previously invisible.

With air on the matter, why is "probabilities" rendered in scare quotes?

And how, precisely, is "shrinking 'probabilities' " related to "adjust epistemic confidences"?

The only reason I see to scare quote "probabilities" is that we have an undeclared denominator (except in larger studies over controlled subject populations, you're more likely in the domain of rough likelihoods).

But I guess no matter what, you can always have an "epistemic confidence" — denominator schmominator.

Finally Penultimately, I don't know what "negligible" is doing in this sentence.

If you get things narrowed down until you think one thing is the majority of the outstanding probability, and you have a cost effective treatment option for that one thing, and a reasonable time horizon to judge whether the treatment is succeeding, and none of the not-entirely-negligible alternatives are going to blow up irreversibly or kill the patient in the meanwhile, then you've got a workable plan, and you're good to go.

The doctor might also continue to actively monitor the not-entirely-negligible alternatives if any risk remains that what's not being treated (the non-diagnosis) takes a sharp or turn for the worse (or a conspicuous turn in any direction).

Having just said that, I might add "treatment history" to the list, right after "patient history" (which might formally be subsumed, but it's not necessary called to mind without a more explicit handle).

I'm not myself in this field, but I did once have a college roommate who went on to complete a PhD at Stanford in expert system medical diagnosis in the ER setting, where timeliness of the diagnostic process sometimes mattered as much or more as correctness.


 * Is the patient going to bleed out in under 60 s?
 * You take a 5 s gander at pertinent factors and the answer comes back "possibly".
 * Get on it stat, no further diagnosis required—nor narrowed probabilities, nor epistemic confidences—not until medical hands are free to contemplate an additional intervention.

Differential diagnosis modulo extreme alacrity.

That's another thing. This text reads a bit like Sherlock Holmes smoking his pipe in his tweed jacket in his big comfy chair in front of the fireplace. Or a bit like Awakenings with a big facility crammed with medical embarrassments in long-term storage rather than treatment. Well, you've got months and months to ponder every last detail; most of those patients aren't going anywhere in the present decade. Or a medical student reading a fat medical tome fleshed out more with theory than practice.

The one setting where CSI-levels of chess tournament scrutiny are systematically applied are the ones that clearly violate: first, do no harm. You're going to treat a suspected cancer with a form of radiation or chemotherapy that really could kill the patient all by itself. Bad, bad time to make a major diagnostic oopsy.

It's an interesting set of algorithms, because there's speed chess of the thing, and there's also correspondence chess of the thing. This kind of overstuffed sentence is 100% tweed and does little justice to the former. &mdash; MaxEnt 01:49, 24 July 2022 (UTC)