BMI findings: measurement vs interpretation

Clinical Reference Groups Nutrition and Dietetic CRG
,
1

As mentioned briefly at the recent Nutrition & Dietetics CRG meeting, this is a work item already identified for the group. Some recent CRS requests have also highlighted that several different modelling patterns are currently used for BMI-related findings in SNOMED CT, which can lead to overlap between concepts.

As discussed in the meeting, I’m posting here to gather views.

1. BMI measurement combined with interpretation

Examples include:

  • 162863004 |Body mass index 25–29 – overweight (finding)|

  • 162864005 |Body mass index 30+ – obesity (finding)|

  • 408512008 |Body mass index 40+ – severely obese (finding)|

These concepts combine a BMI range with a clinical interpretation which is also reflected in the hierarchy. For example:

162864005 |Body mass index 30+ – obesity (finding)|

is a subtype of:

  • 414915002 |Obese (finding)|

  • 48499001 |Increased body mass index (finding)|

2. BMI findings without interpretation

Example:

  • 427090001 |Body mass index less than 16.5 (finding)|

These represent BMI as a measurement-related finding, without an associated interpretation.

3. Interpretation findings defined using BMI

Examples include:

  • 443371000124107 |Obese class I (finding)|

  • 443381000124105 |Obese class II (finding)|

  • 819948005 |Obese class III (finding)|

These represent interpretations of obesity severity, and their text definitions reference BMI ranges that are widely used in international guidance (for example WHO, CDC and NICE).

However, one issue relates to the presence of BMI range synonyms on some of these concepts. For example:

  • 443371000124107 |Obese class I (finding)| includes the synonym “Body mass index 30.00 to 34.99”

  • 443381000124105 |Obese class II (finding)| includes similar BMI range synonyms

This can make these concepts appear similar to BMI measurement findings, which contributes to overlap with concepts such as:

162864005 |Body mass index 30+ – obesity (finding)|

By contrast:
819948005 |Obese class III (finding)| includes the BMI range only in the text definition and does not include a BMI range synonym.

Potential overlap

Because these patterns coexist, there is some overlap in meaning between certain concepts. For example, CRS requests have suggested the following may represent duplicates:

  • 162864005 |Body mass index 30+ – obesity (finding)|

  • 443371000124107 |Obese class I (finding)|

Possible approach

One possible approach would be to represent BMI findings as the measurement or range only, with interpretations such as overweight or obesity represented separately.

For example:

Measurement

BMI 30+

Interpretation

Obese

rather than a combined concept such as: 162864005 |Body mass index 30+ – obesity (finding)|

This may also provide greater flexibility where interpretation thresholds vary across populations (for example ethnicity or body composition).

Questions

BMI is widely used as a screening measure for adiposity, but it does not directly measure body fat and interpretation thresholds may vary depending on factors such as ethnicity or body composition. Views on the following points would be very helpful:

  1. Representation of BMI findings

From a clinical perspective, would it be preferable for BMI findings in SNOMED CT to represent the BMI measurement or range only, with interpretations such as overweight or obesity represented separately?

  1. Clinical references

Are there specific international references the CRG would recommend using as the basis for reviewing and confirming the BMI ranges currently represented in SNOMED CT?

Of course, any additional comments on this topic would also be very welcome.

Many thanks in advance for any thoughts or suggestions.

2

Feedback from J. Snyder:
The following two concepts are problematic and should be considered for inactivation:

  • 408512008 |Body mass index 40+ - severely obese (finding)|

    • This could be inactivated as outdated – replaced by 819948005 |Obese class III (finding)| that has a synonym of “Body mass index equal to or greater than 40” but unlike the other obese class * concepts does not contain a text definition

  • 162864005 |Body mass index 30+ - obesity (finding)|

    • This could be inactivated as ambiguous – possibly equivalent to

      • 443371000124107 |Obese class I (finding)| with synonym “Body mass index 30.00 to 34.99” and text definition “Body mass index (kg/m²) from 30.00 to 34.99.”

      • 443381000124105 |Obese class II (finding)| with synonym “Body mass index 35.00 to 39.99” and text definition “Body mass index (kg/m²) from 35.00 to 39.99.”

      • 819948005 |Obese class III (finding)| with synonym “Body mass index equal to or greater than 40”

      • 408512008 |Body mass index 40+ - severely obese (finding)|

    • I have had issues with this specific concept in real-world implementation mapping scenarios because it is an open-ended range that technically includes the “40+ concept.

  • 722595002 |Overweight in adulthood with body mass index of 25 or more but less than 30 (finding)| should be inactivated as outdated à replaced by 162863004 |Body mass index 25-29 - overweight (finding)|. If it is deemed necessary to retain the description from concept 722595002, then it should be added as a synonym to concept 162863004.

One of the challenges is decoupling the obesity classification from the BMI range. One reference from 2004 from WHO indicates that these should be retained as the international default range https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(03)15268-3/abstract, but that there is variance in that range based on ethnicity. The problem is, if we leave them and a country makes ethnicity specific content, it will make this part of the hierarchy messy with overlapping ranges.

The only suggestion I can think of this morning is to inactivate the synonyms for the BMI range on the Obese Class I and Obese Class II concepts. Update the text definitions to indicate that these are International default BMI range definitions for the concept, but that there may be variance as each concept is applied to a specific ethnicity.

1 Like
3

Hi Elaine,

our dietetic expert gave the following statement:

“I agree with the view that: BMI is simply a weight-to-height ratio and says nothing about how much visceral fat a person carries. Since visceral fat in particular poses the greatest risk for cardiovascular disease and metabolic disorders, BMI can overlook people at high health risk and incorrectly classify others as being at risk. Therefore, fat distribution—particularly the proportion of visceral fat—is a much more precise indicator than BMI.

I believe it makes sense to separate BMI from its interpretation. I cannot comment on how this can best be implemented in SNOMED CT.”

Additionally, she would recommend waiting for adjustments to international classifications before making changes in SNOMED CT. According to her, the current European guidelines still define obesity (predominantly) with BMI.

@ewooler: I’ve sent you her contact information in case you’d like to invite her to CRG calls as well.

Best regards, Annatina

1 Like