ALLERGY UNITS

[Gunther Schadow]

Requested for FDA/CBER in order to support biologic structured product labeling:

ALLERGY UNITS (typically abbreviated "AU") - a biological potency unit assigned to several standardized allergenic extracts, following in-vitro comparison of the test extract to an FDA CBER reference standard. The FDA CBER reference standard may be assigned a specific AU unitage based on quantitative skin testing in allergic individuals, or based on specific allergen content.

See also #17 and #19

[Gunther Schadow]

The definition is not sufficient to establish this notion as a unit.

This is probably the one worst possible case of an arbitrary unit which we have to deal with somehow. I don't think that all AUs are using FDA reference standards. I read they may just create idiosyncratic scales based on skin-testing separately for each product / manufacturer.

[Gunther Schadow]

Use of Skin Testing for Evaluation of Potency

[Gunther Schadow]

We can't argue against this. There is too much precedence already. Instead we will have to expand on the notion of procedure defined units.

Christof has an issue with the arbitrary unit flag blocking any comparability. We can use a simple id for the procedure that defined the unit. This has to extend the base unit vector into a sparse variable length structure.

This is still a problem but no worse than with the arbitrary units which we already have.

It is not a problem when it comes to proper measurement data structures with kind of quantity (fully specified), the values can be dimensionless and it still works. But when (as in this case) the units are used in a simple "amount of substance equivalent" for ingredient quantity, we have a problem, because we can't assume that dimensionless quantities actually compare.

With the choice of [AU] and [BAU] we can't even be sure for the same ingredient.

This too has a precedence, such as in Gray vs. Sievert, and here these units are even SI units.

[Gunther Schadow]

So, adding [AU]