[ExI] good bexarotene article

[spike]

>... On Behalf Of BillK
Subject: Re: [ExI] good bexarotene article

On Sun, Feb 12, 2012 at 4:12 AM, spike wrote:
>>... There may be a way to reduce diseases to a number or series of
numbers.
> This will be needed for sorting purposes...
> Have we any medics among us who can suggest a data structure?  Or 
> point to one that already exists?  spike
>
>


>...Quick to write down, but you are talking about a huge project. Once you
get into it, it is ferociously complicated.

>...The medical profession have been trying to develop a disease coding
system for over 40 years, with some success.

>...SNOWMED CT seems to be the latest system
<https://en.wikipedia.org/wiki/SNOMED_CT>

>...It covers areas such as diseases, symptoms, operations, treatments,
devices and drugs. Its purpose is to consistently index, store, retrieve,
and aggregate clinical data across specialties and sites of care. It helps
organizing the content of electronic health records systems, reducing the
variability in the way data is captured, encoded and used for clinical care
of patients and research...BillK

_______________________________________________

Thanks BillK!  The more I thought about this yesterday, the more clear it
became that this idea is so obvious it must have already been attempted a
thousand times, and there had to be an existing front runner somewhere.  I
am surprised I had never heard of it.

About a decade ago, I created a matrix for everything that could go wrong
with a particular oddball motorcycle, the Suzuki Cavalcade.  We eventually
collected data from over 600 bikes, along with VIN, so I could get a
manufacture date and so forth.  We saw clusters of a particular failure,
trends, calculated mean time between failure of all the subsystems that
commonly fail and so on.  That tool allowed us to use all the mathematical
tools we developed in the rocket science industry.  I could imagine
something about four orders of magnitude more complicated than that for
humans, which would be an interesting tool in medical research.

We could use something like that Snowmed CT database, and arrange a
volunteer base to do some kind of background computing project to look for
unexpected correlations.  Once we get thousands of people grinding away in
the background, some of those people might study Snowmed CT and come up with
suggested extensions and improvements, or if nothing else, just clean up
their own records, mention all the grass they smoked back in college, and
such as that.  Something roughly analogous happened with the cavalcade
database after we started getting useful data: the others used it and put
their own data in there, such that it snowballed.  If the Snowmed CT
database does likewise, it would be appropriately named.

Way to go BillK!

spike