UK Pathology Units List – the What, Why and How…

UK Pathology Units List – the What, Why and How…

The UK Pathology Units List (UKPUL) is a table of standard UCUM unit codes and supporting information designed to facilitate the exchange and storage of measured quantities as elements of pathology test results, in computable form.

TL;DR

It comes with a documentation over 70 pages long and an Excel spreadsheet with 8 tabs. 🤓 This article and others in the series are the TL;DR version for the busy folks who want to know what it is about, why they should look at it and where to get started!

Background

Like many other countries, England has been working on a set of national pathology standards for sharing lab results across health and care. The broad remit of these standards is to cover:

  • The names of tests – standardised using the Unified Test List – based on SNOMED CT and covered in past articles
  • The units accompanying the test results – the UK Pathology Units List – the focus of this article
  • The structure of the tests – the UK Pathology FHIR specification – based on HL7 FHIR

Note each time I say `tests`, the word covers the request and result sides too. But let’s not digress and focus on the UK Pathology Units List (UKPUL)… Arrgh can we come up a better acroynm please?

UK Pathology Units List – the What

What is the UK PUL you ask? This is the description provided by NHS England:

The UKPUL is a table of standard UCUM unit codes and supporting information designed to facilitate the exchange and storage of measured quantities as elements of pathology test results, in computable form.

I’d rephrase that as:

A collection of standardised units of measure and their representations, which are recommended for use in lab results across health and care in England.

I deliberately say – `recommended` since the documents do not hint at any mandate to use them. In fact the UKPUL includes a list of `preferred`, `accepted` and a category of `discourage` units and representations.

In the current version, only 65% of units are recommended for use. There are a total of 228 `active` units of which 78 are `discouraged` from being used.

If that sounds a little confusing, then you aren’t alone – but there is a reason for this. There are a lot of variations of the units that can be used to report a given test. For example, let us look at the various units that can be used to represent haemoglobin concentration in a blood.

Scientifically, a concentration is a ratio of a given substance (e.g. mass) within a substrate (represented as the denominator) – usually volume for fluids like blood. So Haemoglobin concentration is typically reported as: `grams per litre of blood`. Here are some possible examples in the Mayo clinic site:

  • For men, 13.2 to 16.6 grams per deciliter
  • For women, 11.6 to 15 grams per deciliter

Let us take the first example 13.2 and ignore the range for now… So mathematically the following are equivalent

Possible units for reporting Haemoglogin concentration

A machine could do the conversions in less than a second, but most humans become familiar with patterns over time when it comes to interpreting data. So it makes sense to always share Haemoglobin concentration using a single `preferred representation` even if there are many mathematical equivalent representations.

Interestingly in this example, Mayoclinic uses units of measure (g/dL) which the UKPUL flags as `discouraged` and instead recommends using grams per litre (g/L)!

So part of the reasons why I said  `recommended` previously is the motivation to eventually move all lab results to a single `preferred` representation when dealing with units of measure.

Units representations and conversions – the Why

In the above example, we already saw how it is possible to have multiple mathematically/computable equivalent representations of a given result. Remember, the focus of a lab result is not the units of measure, but rather the finding aka result and the interpretation that a clinician can safely make from it…

So if the following results are equivalent, then how do we convert them all into a `preferred human representation` – 132 g/L?

Haemoglobin Concentration result showing units separately

The answer is – UCUM, the Unified Code for Units of Measure going back a few decades. I won’t go into what UCUM itself is, but think of it as a syntax/convention of how different units of measure can be safely represented and converted by a machine. Yes, that’s right – UCUM is really for machines and not humans. But given a lab result and its associated units represented in UCUM, a machine could trivially (and safely) convert it into a preferred representation.

Given a result with the corresponding UCUM representation for g/dL (which also happens to be g/dL), as below:

Haemoglobin Concentration result with units and associated UCUM

A computer (software) could trivially convert it into any of the equivalent representations, including the one that we prefer. This is in fact, one of the primary reasons that HL7 FHIR, the messaging standard used for sharing lab results, recommends using UCUM to represent units of measure.

UK Pathology Units List – the How

Okay, now that you were with us till here you might want to know how exactly you can make use of the UKPUL in your software or in day to day use. Welcome to the somewhat interesting part of the experience… 😅 We support a reference list of units in our product and wanted to understand `the how`. Here is an animated preview of what is in the 8 tabs of the Excel spreadsheet that is what the UKPUL is published as:

Preview of the UK Pathology Units List – showing various tabs and content

There is a lot of information there to unpack and figuring out what tabs and columns could be ignored and how to consolidate relevant information was not straightforward. Trying to outline the content of each tab and how we approached this would make this article really long. 🤓 Since this is the TL;DR version, for most cases you only need to look at the following tabs:

  1. Unit_List
  2. Unit_List_Detail

I am not saying for a minute that you can ignore the rest of the tabs. Just note there is a lot to unpack and understand there. Clearly a lot of thought went into the creation of this artefact and perhaps this is still evolving. Here are some thoughts about the future use cases for an artefact of this nature:

  • While going through the associated document of this artefact, my mind was drawn to the parallels between the schema of this UKPUL artefact was trying to solve the representation issue and how SNOMED CT already solves the same issue.
  • Will there ever be a more standard distribution format for this artefact – since this would make life easier for suppliers.
  • There is a missing association between a given test (in the UTL) and its corresponding `preferred` or accepted units of measure. Perhaps future work is planned in this space.
  • Having the ability to immediately associate the replacement of a discouraged representation/unit with the preferred version perhaps that needs more work.
  • I should write this in small print – but a future national interoperable standard which is clearly a technical artefact being published as an Excel spreadsheet with multiple tabs, colours, embedded pivot table fields does not feel `interoperable` or `standard`… 🙈

From the perspective of our users who want a human readable representation and a quick lookup for units, we created something like this:

Simpler interface for looking up units of measure in Pathnexus

How we went about doing this and what way we use the UKPUL and the associated representations will be the focus of the next article in this series, if that is of interest.

What are your thoughts about the UKPUL? Are you using it or do you already use something like this in your own setting? We would love to hear your thoughts and experiences… Lets share our experience and move lab results to a safer place!