Collective wisdom is an awful way to diagnose medical problems.

The crowd will tell have him running around getting tested for everything from celiac to Grave's disease. When the tests turn up negative, the crowd will claim that the testing threshold is wrong, or that doctors just can't be trusted. He will be told to try everything from vitamin B12 shots, to allergy medication, eat more kale and bananas. Perhaps if someone recently read a weston price website he will be advised to eat pork chops and drink buttermilk. Maybe polyphasic sleep is the answer? Or maybe light therapy and a white noise generator? At least this is Hacker News so nobody is recommending cupping or leeches or magnetic wristbands... yet. Asking the internet is like a schizophrenic consulting with the voices in his head. Every fruitcake idea will be presented with equal weight.

With that said, my contribution to the noise is that if these symptoms have been there for 18 years and haven't killed him, it's probably psychological.

>With that said, my contribution to the noise is that if these symptoms have been there for 18 years and haven't killed him, it's probably psychological.

By that logic, any chronic disease is psychological.

If it is psychological then it is entirely possible that a magnetic leech wristband would be an effective cure.

I think of it somewhat like researching a topic and starting with Wikipedia. It's just a really good start.

There are a lot of very smart people here and I think it was a great idea to come here. If the OP is smart enough to not take everything said as a cure, they could very well gain some very valuable bit of knowledge that could lead them on the right path; just like starting with Wikipedia.

I prefer to think of this as an internet based brainstorming session

It really seems like doctors could start cooperating more; specially with rare diseases.

Do doctors always use computers as part of the diagnosis process? Shouldn't it be a criminal offense not to do this?

In fact, if the diagnosis differs from what the computer says, then this should always be reported. Failure to do so should be considered a criminal offense.

Downvote. You really throw around that phrase "criminal offense" too much.

Penalties are certainly a mistake in the case of your statement, as you haven't really outlined any computerized diagnostic system that is accurate on the entire range of human maladies, is available to all, and that can account for all of the fuzziness of reporting symptoms (just to name a few obstacles).

But I can't see at all where you're coming from when you say criminal offense. This would suggest that doctors should be locked up?

A doctor who refuses to use such a system is putting the lives of his patients at risk. How is that acceptable?

By that logic, a doctor that doesn't try every possible remedy is putting his/her patients at risk, no? So you should throw a doctor in jail for refusing to consider blood-letting as an acceptable cure?

It might be unacceptable (although as I said, you haven't even made that case). But that's a long way away from being criminal behavior.

What do you mean by "use computers"?

I'm referring to using an officially mandated automated diagnosis system.

Does such a thing even exist? That sounds like a very hard problem.

Actually it does. It's fed with data that says: Condition X is usually caused by Y with Z% probability, etc etc.

Then you feed it all the symptoms, sometimes it asks questions, and gives you a list of causes sorted by probability.

It works almost exactly like this game http://www.guessthename.com/

Apparently the biggest problem with it is that doctors won't use it.

The second problem is feeding data - you basically have to parse every medical paper ever published.

Wait a second, if the system does in fact exist, and it works well, and the biggest problem is that doctors won't use it, I'm with amichail.

Guessing a name is a much simpler problem than guessing a disease. Can you point to any such systems for medical diagnoses?

Here's a paper on one (just a random one from google) http://www.med.univ-rennes1.fr/cerf/publi/ADM_index1.html

Here's a patent: http://www.freepatentsonline.com/6607482.html

There are lots of a papers, and a small number of working prototypes, but I have not heard of any that are actually being used.

Something like this? http://www.youtube.com/watch?v=5oXCm4N3J_M