ucum-org
Request for: [arb'U]/l
Many assays in clinical chemistry and laboratory medicine are performed in serum or plasma. Many of these assays are procedure defined and not (yet) international standardized. Often the results can't be measured as mass concentration or substance concentration. So the result of these assay are expressed as arbitraire units, in UCUM that could be [arb'U]/l (or /mL). However, this expression is not excepted.
Sometimes there is confusion around the apostrophe: Make sure that you don't use other characters here, e. g. backquote or single quotes. Otherwise, please contact the authors of the validator you are using.
Gunther Schadow, co-creator of UCUM acknowledged that [arb'U]/L is a valid unit expression for a concentration. The problem, he writes, is concerning the prefixes, like "k" (for kilo), as it is not garanteed that [arb'U] is metric. However when converting [arb'U]/mL to a concentration per liter, the prefix "k" is valid, as liter is a metric unit. So k[arb'U]/L should be a valid UCUM expression.
And what if we 'prove' an [arb'U] is metric unit concentration, though procedure defined? And not a Mcnc or Scnc of course. In the Loinc database there are many concepts with an example unit [arb'U]/mL, which is valid. Suppose we do not want to use [arb'U]/mL (which by the way is archaic), but [arb'U]/L. Nothing has changed on the assay, so clearly -in my opinion- the result will be larger by one decade (x 1000). Isn't that correct?
To the outside world your "arbitrary unit" is just a number. And I do not advocate adding the prefix mechanism to be used in such a case. There is no problem to tackle any problem with orders of magnitude by using the existing mechanism for numerical values in computer languages ( like 1.2345e6 [arb'U]/L or just 1.2345e6 1/L. If your "unit" should be recognised and exchanged between labs, go for a LOINC code that exactly defines what that number means.
I there already a 'verdict' about the use of k[arb'U]/L for metric non-standard units?
We try not to report absurd numbers to doctors as they are easily confused by exponential numbers. Many laboratory systems and electronic patient records have limited positions available for the results. There seems to be a conflict between safe and practical use of test results and their units and the pure theoretical approach of units. How can we solve this issue?
Your suggestion of a result like 12345e6 [arb'U]/L is NOT according to UCUM. Then it should be 12345 10*6 [arb'U]/L. I still urge to consider a solution for metric [arb'U]/L with metric prefixes. The is a request from the Dutch Society for Clinical Chemistry and Laboratory Medicine (NVKC) and Nictiz, National Institute for ICT in Health Care.
Dr. Dirk Bakkeren Clinical chemist Cie AICT, NVKC