vmelkon - 29-1-2016 at 20:33
How are blood tests performed?
Such as for measuring bilirubin.
glucose
sodium
potassium
chloride (possibly with AgNO3)
amylase
etc.
Do you learn this by studying medicine?
Metacelsus - 29-1-2016 at 21:32
Serum electrolytes (sodium, potassium, chloride, bicarbonate, etc.) can be measured by ion-selective electrodes. Glucose can also be measured
electrochemically. I'm not sure about the other ones.
ziqquratu - 31-1-2016 at 16:26
You would certainly learn about this by studying medicine - although I'm not sure you'd learn how to do it, since I think it's fairly rare for doctors
to run their own tests (of this sort) these days.
The analysis of blood or other tissues is referred to as pathology. A reasonable starting point, as always, is the appropriate article on Wikipedia. Have a read there - they list some of the techniques used for various types of analysis, with links to the relevant pages for each.
A problem when searching for specific information on protocols is that much of this testing is automated - over 15 years ago, I spent a few weeks
over summer working in a pathology lab in a small town hospital, and even then most non-histological tests (where someone actually has to look down a
microscope at a sample) were performed simply by loading the vials of blood or urine (perhaps after centrifuging, depending on the tests) into a rack
and checking the appropriate boxes on a computer. I can only imagine that things have gone further down that road in the interim!
However, if you're looking for a specific test, you might have more luck than looking for "handbooks" which outline lots of methods. For example, a
quick search for "amylase measurement" (without quotes) gave, as hit 3, this document, which appears to give a detailed outline of how the test is done (on page 3). I'm sure similar information can be found for other
specific tests, if you search them by name.
phlogiston - 1-2-2016 at 12:28
Bilirubin can be measured by the Jendrasik-Grof method. This assay relies on its reaction with Ehrlich's diazo reagent (diazotized sulfanilic acid) is
added to the sample, which yields an blue colored product (lambda max=530 nm) which is quantified spectrofotometrically. This measures conjugated
bilirubin. For total bilirubin, caffeine benzoate is added too.
Amylase assays typically work by incubating the sample with ethylidene-pNP-G7 (abbreviated 'EPS'). This substrate yields a highly colored product when
it is cleaved (lambda max=405 nm)
In general a number of different analytical techniques are commonly used for blood tests, depending on the desired speed, accuracy, specificity,
precision, etc.
Most tests detect enzymatic activity, or are based on (radio)immunological methods and colorimetry. I have used LC-MSMS and GCMS extensively to
quantify metabolites in blood and other fluids of patients suffering from metabolic disorders.
Most medical doctors have only a very superficial understanding of the methodology. They only need to know enough to be able to interpret the result
(which test can I trust, and under what conditions is it unreliable, etc).
In the lab, you'll see chemists, biochemists, analytical chemists, clinical chemists, biologists, designing and doing the tests.
The assays you mention are so standard and optimized, they will be done by robots. All the lab crew has to do is put a barcode on the sample tube and
put it on a conveyor belt. The robots are fed big bottles of standard reagents, bought ready-made. When it breaks, they call someone to fix the thing.
The real fun is in developing new assays.
[Edited on 1-2-2016 by phlogiston]
Tsjerk - 1-2-2016 at 13:07
I agree with Phlogiston,
Medical doctors rarely know what is going on behind the screens, or what is the real mechanism behind a decease. Because they don't have to know.
On the other hand, I would definitely not want a scientist/researcher at my bed... Yes sir, this medicine could possibly work... or not. but anyway,
the outcome is the most important! See the work of Edward Jenner 
Ozone - 1-2-2016 at 13:11
NMR is gaining a lot of momentum for high-throughput clinical diagnostics. For example, cholesterol/TAGs: http://www.mlo-online.com/increasing-lab-efficiency-using-nm...
Typically, they are using (rooms full of) 400 MHz spectrometers (I can imagine ULF relaxometry for lipoprotein particle sizing, also). See Vantera: http://www.sciencedirect.com/science/article/pii/S0009912014...
In reality, I expect many more assays to be adapted for one-shot panels for bilirubin and glucose (which would be easy by NMR), and others. I also
expect to see reliable lipoprotein subfractionation, fatty acid profiling, blood carbohydrate profiles (not just glucose), and many others.
I can easily imagine several panels run non-destructively and with minimal preparation/handling being run at once (especially if chemometrics are
used).
O3
Tsjerk - 1-2-2016 at 13:22
I agree with Phlogiston,
Medical doctors rarely know what is going on behind the screens, or what is the real mechanism behind a decease. Because they don't have to know.
On the other hand, I would definitely not want a scientist/researcher at my bed... Yes sir, this medicine could possibly work... or not. but anyway,
the outcome is the most important! See the work of Edward Jenner
vmelkon - 1-2-2016 at 13:25
Thanks,
I also found
http://www.acb.org.uk/Nat%20Lab%20Med%20Hbk/ALT.pdf
which happens to be from acb.org.uk as well.
Dr.Bob - 28-2-2016 at 18:14
There are specific courses and programs in clinical chemistry and lab tests, many doctors don't know much about the assays details. Now many are done
with antibodies, bead based assays, fluorescent assays, and many other fancy methods. People have been trying to find simpler tests, based on DNA
chips, microfluids, and other new technologies, but even those are mostly tested and calibrated by older chemical methods, colorimetric assays, and
more.
I think I even have a few books on some of those methods. I'll have to look for them.