In healthcare, patient instance data is the focus of what we interact with, move around, analyze, and report on. If you look at the patterns of data we deal with, they tend to fall into a few categories.
Images and video are important parts of a patient’s story. Whether it is an ultrasound, X-ray, photo of a rash, or video of a procedure. This is how we capture things in a relatively objective way which can convey more than words ever could.
I am using narrative to represent any unstructured collection of words, written or spoken. This is an excellent way to capture the essence of an event from the perspective of the “author” of the narrative.
Typically a patient’s allergies, adverse reactions, medications, diagnoses, social history, family history, etc. are represented using some terminology or “code system.” This is a code value and a human readable term used to represent a concept. Often less effective than images and narratives in conveying the true essence of being or thought process, it is more easily leveraged by software that prefers a stable list of codes upon which to reason.
We often want to measure some aspect of what is going on with the patient. Whether it is their weight, current age, a lab result, or a medication dose amount, these quantified values are critical for both the human and the software to understanding what is going on with the patient. Like image data, if done properly, measurements are objective and allow us to see the patient in motion.
This last category is what I am going to focus on in this post.
The Quality of Quantity
Having “cut my teeth” in the reference lab industry and spending my “salad days” in the clinical trials central lab industry, I have a keen awareness of the importance and challenges relating to the quantified information known as laboratory results. The most basic and necessary ingredient in measuring something knowing what you are quantifying and how you are quantifying it.
The “what” you are quantifying is in the domain of terminology. A coded term will describe the property being measured. In some cases, the “what” might be represented by multiple coded terms.
In quantified data, the “how” is the scale I am using to assign the quantity. This my friends is where “units” come in. A unit of measure is necessary in all situations where someone is trying to quantify something. Yes, even where there is no unit (like in the case of arbitrary unit measurements) you still need to understand that the unit is “no units.” Units are important because they allow us the compare quantified information with the same unit and potentially convert and compare quantified information with different units. Units have been scuttling about healthcare information for decades, like lexical trilobites. In lab results, medications orders, patient demographics and other places we have been relying on units to help us make decisions and care for patients long before our current analytics renaissance.
The first questions is why do so many systems represent units as text? It is true, many applications still use unit abbreviations (aka free text) to identify the unit of measure for a quantified value. The answer is, there is no standard code system for units in healthcare. Which leads to a second question.
The second questions is, why is there not a standard code system for units, complete with base unit conversion tables, specifically for healthcare?
If you are thinking to yourself “Yes! I finally know something Charlie doesn’t know!”, you could be right…unless that thought is relative to the “Unified Code for Units of Measure” or “UCUM” from the Regenstrief Institute. Rest assured, I know about UCUM. If you think UCUM in its current form is the de facto code system for units in healthcare, UCUM might not mean what you think it means. That’s because UCUM is not a code system.
No, it’s true. UCUM is a nomenclature. If you go to the UCUM website you would see that UCUM is a set of rules for properly assembling and displaying units. A code system is a ‘system of codes’ that is typically controlled and finite. In the UCUM specification there are no stable, non-semantic identifiers for the different units. Nor is there an ontology that allows you to traverse the implicit relationships between them.
Here is an excerpt from the UCUM website:
The Unified Code for Units of Measure consists of a basic set of terminal symbols for units, called atomic unit symbols or unit atoms, and multiplier prefixes. It also consists of an expression syntax by which these symbols can be combined to yield valid units. The tables of terminal symbols are fixed as of every revision of The Unified Code for Units of Measure, additions, deletions or changes are not allowed. All expression that can be derived from these terminal symbols and the expression syntax are valid codes. Any expression of The Unified Code for Units of Measure has a precisely defined semantics.
The expression syntax of The Unified Code for Units of Measure generates an infinite number of codes with the consequence that it is impossible to compile a table of all valid units.
That the tables of terminal symbols may not be extended does not mean that missing symbols will never be available in The Unified Code for Units of Measure. Suggestions for additions of new symbols are welcome and revisions of The Unified Code for Units of Measure will be released as soon as a change request has been approved.
I am not disparaging UCUM, or the fine folks at the Regenstrief Institute who put it together. It represents a great deal of knowledge and effort. It is important that we understand what we have and what we don’t, if our intent pragmatically solve the issues we face every day. My concern is there are several groups, like HL7 and FHIR to mention a few, that refer to UCUM as a code system and this is misleading. When people try to use UCUM and a code system (read Terminology), they will find that it does not work like one. If you have searched for a download of the UCUM code system you might be disappointed with what you get. If you don’t believe me click here. (My favorite part of the downloaded file is for each sample unit there is a row number and in the header, under where it says ‘row_num’ it also says ‘not a code’ – we love integers and they know it.)
So, if UCUM is not a code system, what should we use?
I have looked around and have not found much. There is an ISO unit standard, but it is not great. The codes are pseudo-semantic and there is not much to it. It seems like it is a general trade standard as opposed to something intended for healthcare (unless you want to give someone a “carload” of diphenhydramine…)
To have a code system for units, it would require a team that reviews and assigns non-semantic identifiers, proper names and recognized abbreviations, assigns unit types and provides an ontology for describing how units relate to each other and how compound units with numerators and denominators break down. Like I said, after some amount of searching I came up empty handed. Nature abhors a vacuum.
Clinical Architecture – Healthcare Units Model
In order to address this gap, Clinical Architecture has created a code system designed for healthcare that we use internally and provide to our Symedical® clients. The CA Healthcare Units Model (HUM) is intended to cover units that are used in healthcare. We follow good vocabulary management practices and assign appropriate aliases, UCUM compliant descriptions and symbols, as well as cross walks any other standard UOM code systems, present or future. We also have base unit conversion and a unit, with unit types and complex scenario indicators.
Our hope is, by providing HUM to our clients, we will help resolve one of the more basic issues they all face. We plan on updating this quarterly (to handle new and exciting units… Or update the mappings and other relevant information). If you are a CA client and have not had a chance to review the HUM with client services, give them a shout. If you are not a CA client and you would like to learn more about HUM, and all the other amazing things CA and Symedical can do to make your life better, please feel free to get in touch here.
If I missed something, or you have any other feedback, feel free to leave a comment or get in touch.