Protocol of measurements like Height or Weight

Clinical
1

Hi all -

In Wales we have a use case (which I think is universal) to record if a measurement like a weight or a height is Measured, Reported, Estimated and if Estimated, which was the method used to estimate the value

There are archetypes like Body surface area and BMI that have a Method and formula in their Protocols but not in Height/Length or Weight.

The use cases are frail patients, bedridden patients (like ITU patients), or other situations in which it is impossible to measure the value and it is required to calculate a dose of medication, for example.

We would want to record both that it is estimated, and that which formula/method of estimation has been used. Do you agree that this should be part of the archetypes?

This is a link to one paper in which methods of estimation are used: A comparison of three methods for estimating height in the acutely ill elderly population - PubMed (nih.gov)

I would like to hear people’s opinions and if it should be included as a local cluster to the protocol or this requires a change.

Thanks!

2

Hi Marlene,

You will note that there is the following statement in the ‘Use’ of the Body weight archetype:

“Can also be used for recording an approximation of body weight measurement in a clinical scenario where it is not possible to measure accurately body weight - for example, weighing an uncooperative child, or estimating the weight of an unborn fetus (where the ‘subject of data’ is the Fetus and recording occurs within the mother’s health record). This is not modelled explicitly in the archetype as the openEHR Reference model allows approximations for any Quantity data type by setting the attribute Magnitude_status to the value ‘~’. At implementation, for example, an application user interface could allow clinicians to select an appropriately labelled check box adjacent to the Weight data field to indicate that the recorded weight is an approximation, rather than actual.”

There is a similar statement in the Height/Length archetype. The use case driving this statement was an absence of requirement for a formula, maybe even an informal ‘guesstimate’.

There is also a statement in the ‘Misuse’:

“Not to be used to record a calculated body weight, such as an estimation of the body weight of a person with one or more limbs missing. A calculated body weight may be based on, some or all of, the measured body weight, other body measurements and an algorithm. Use other OBSERVATION archetypes for this purpose.”

The other OBSERVATIONs suggested remain unspecified and unavailable, unfortunately.

The reason for the misuse statement is that a calculated weight can be considered a conceptually different entity to an actual measured weight. For example: in individuals with bilateral lower limb amputations we need a calculated weight for medication dosing based on BSA. In that situation we also clearly need to differentiate between the calculated weight and their actual weight which can still be used for tracking weight change etc.

In the situation with height there are similar issues, for example in paediatrics, contractures due to Cerebral Palsy etc will require a calculated height for tracking growth parameters.

To add to the confusion, the term “Estimated” has been used vaguely in many situations - ranging from calculated to vague guesses and we probably need to be explicit in our modelling.

I did some initial modelling on these calculated observations a long time ago, testing how we might specialise these additional concepts for body weight and height/length to support documentation of malnutrition, but it was difficult and it was dropped without a final resolution and no specific user request. However you will note the family of OBSERVATION archetypes for body segments which did arise out of these investigations. For example, Body segment length contains ‘knee height’ and ‘armspan’ which will support our calculations of other body parameters.

This area clearly needs further consideration and investigation as to the best way to define and model these seemingly simple, yet complex, parameters.

Cheers

Heather

1 Like
3

Hi Heather,

Thank you so much for your reply.

This extract that you have copied is something that I find a little bit confusing:

And

So we are accepting estimated values to be recorded, but we are not storing anything stating that the value is estimated - the UI is expected to flag that (where would it retrieve it from? And if it is a different UI reading from that data it wouldn’t have that ‘flag’).

I agree with the misuse 100%. This is for the recording of the weight as-is not as-expected.

In terms of factors affecting a specific weight the intention is to record any ‘special consideration’ to that weight (ie: pre or post dialysis, a person with a cast or external fixation, amputations, etc.) as a coded list with an ‘other’ (the confounding factors). But those are different scenarios to an estimated value due to, for example, a patient being bedridden (ventilated patient) and difficult to weigh and an estimate is calculated from their mid upper arm circumference.

I will look further into the body segments. Is there an incubator with the work of the calculations

I am still unsure if the ‘estimated’ flag and the calculation used to estimate that length or weight should be a part of the protocol or be a separate cluster in the ‘extension’ to have a place to record this.

Thanks,
Marlene

4

Hi Marlene,

I think the heart of the confusion may lie in the use of “estimated”. Maybe we should try to tease this out and avoid using estimated.

In the ‘Use’ the DV_QUANTIFIED data type supports the use of an non-quantified indication of accuracy - “~” : meaning the value is approximately a value, rather than a specific point value. See the Data Types Information model
So this reflects the capacity to record an unspecified level of inaccuracy of a value that is recorded. This is conceptually quite different to a a value calculated using other measured parameters such as body segments.

My suggestion is that a calculated height or weight should be recorded as conceptually separate archetypes - they are not a measured height or weight and we don’t want to accidentally confuse, conflate or graph actual measurements alongside ones calculated based on algorithms and used for a different purpose, especially if there both an actual weight and a calculated weight might be recorded simultaneously but for different purposes eg in an amputee. As you suggest, each of calculated concepts needs to record in the Protocol how the values were reached eg the formulas or algorithms.

In addition, the concept of adjusted and ideal body weight which also needs to be considered within the dietetics domain.

Agree

They are held in the Anthropometry project

1 Like