In my new role as the “eHealth integration guy” (or whatever), I spend a lot time sitting around thinking about how to make diferent NHS Scotland systems talk to each other. Today, while sitting in a meeting, I had a flashback to my philosophy days: the relationship between patients and the corresponding entries in multiple databases is akin to Frege’s concept of Sense & Reference, and that this could be a very good way to explain to non-techies the concepts and problems associated with multiple identifiers. A quick Google verified that I was remembering the concept of sense & reference correctly (more or less). If you’re unfamilar with it, here’s the basic argument as presented by Wikipedia:

Frege’s distinction rejects a view put forward by John Stuart Mill, according to which a proper name has no meaning above and beyond the object to which it refers (its referent or reference). That is, the word “Aristotle” just means Aristotle, that person, and no more. It does not mean “The writer of Theætetus.” Hence, the sentence Aristotle was Greek says only that that person was Greek. It does not say that the writer of Theætetus was Greek. That is, it permits that Aristotle might not have written Theætetus. More generally, for any given proposition about Aristotle, one can use the name without believing that proposition to be true of Aristotle.

Frege’s central objection to the view that a name’s meaning is no more than its referent is that, if a and b are names of the same object, then the identity statement a = b must mean the same as a = a. Yet clearly the first can convey information in a way that the second cannot; that Samuel Clemens is Samuel Clemens is just trivial, but that Samuel Clemens is Mark Twain is interesting. Why? Or, why is Cicero is Tully more significant than Cicero is Cicero? And, by the same token, Samuel Clemens wrote novels and Mark Twain wrote novels would have to mean the same thing but, again, the two sentences seem to convey different information.

Frege’s distinction is meant to make sense of these phenomena. He postulates that, in addition to a reference (Bedeutung), a proper name possesses what he calls a sense (Sinn), some aspect of the way its reference is thought of that can differ, even between two names that refer to the same object. The important difference between Mark Twain and Samuel Clemens, for example, is a “difference in the mode of presentation of that which is designated”. The sense of an expression is “that wherein the mode of presentation is contained”. Thus, one can know both the names Mark Twain and Samuel Clemens without realizing that they are about the same object, because they present that object in different ways, that is, they have different senses. Another demonstrative example for this is the following: “The Leader of the Labour Party in October, 2007” and “the Prime Minister of the United Kingdom in October, 2007“. These two linguistic expressions differ in sense, but they do have the same referent, that is Gordon Brown.

In the NHS, the reference is the patient. The names are all the identifiers, whether demograhic (first name + second name + date of birth etc) or administrative (CHI, NHS number, the internal IDs used by third-party systems, etc). Using Frege’s argument, it becomes apparent why we have so much difficulty using a single identifier (CHI) across all these different computer systems: the CHI number has a different sense depending on where it’s stored and how it got there, even though on the surface it appears exactly the same. We try to pretend that CHI is an absolute reference in the same way as suggested by John Stuart Mill above, but this is not the case. We need to start recognising that IDs have sense as well as reference. Utilising the same ID (CHI) in multiple systems covers up the fact that they have different sense. This inevitably results in confusion.

I’m not claiming this is any sort of deep revelation, but it’s a badly-needed alternative approach to explaining to non-techies the relationship between administrative information and real-world stuff.