OHDSI Home | Forums | Wiki | Github

ICD10CM T39.1 Extensions Mapping Question

Hi all,

Hope everything is going well.

Thanks again for the help about KCD T39.1 mapping question, and now it is mapped to SNOMED - Poisoning caused by acetaminophen, same as ICD10 T39.1. Previous topic link

However, for ICD10CM, there is T39.1 code and 25 sub concepts under it, but the mappings for those 26 concepts are to Drug-related disorder or Accidental poisoning by analgesic or similar concepts that are an upper level match.

I was wondering if ICD10CM T39.1 and its extensions need to be consistent with ICD10 T39.1 as well, which maps to Acetaminophen intoxication level, instead of analgesic drug level.

Please let me know what you think.

@Polina_Talapova @Dymshyts @Christian_Reich

Thanks,
Jing

Hi Jing Li, thank you for shedding light on the issue.
The reasons why you see this confusing mapping difference between ICD10CM and other ICD-related vocabularies are the following:

  1. Discrepancies in the formulation of ICD concept names.


    The ICD10CM concept of “T39.1 Poisoning by, adverse effect of and underdosing of 4-Aminophenol derivatives” is applicable not just to poisoning but also to an adverse effect as well as underdosing, while T39.1 codes in ICD10, ICD10CN, and KCD7 indicate poisoning only. According to the OMOP mapping rules, we are not allowed to map a concept indicating “condition A (poisoning) OR condition B (adverse effect) OR condition C (underdosing)” just to “condition A (poisoning)”. In this case, such a concept has to be mapped to the closest common semantic ancestor (more generic concept) in SNOMED.

  2. Different mapping approaches:

  • Mapping of T39.1X%-codes in ICD10CM is a comparatively new one and based on the information obtained from ICD list , under which 4-Aminophenol derivatives include not just acetaminophen, but rarely used phenacetin and acetanilide. Due to this, mapping just to Acetaminophen poisoning is not recommended according to the mapping rules in order to prevent data loss. Moreover, the T39.1-related subset of ICD10CM concepts cannot be represented through the “one-to-one” mapping due to complex semantics, in which an intent of poisoning also matters. That is why a “one-to-many” mapping is applied.


    The purpose of the approach is to provide a background for the creation of OMOP Extension Standard concepts for all ICD codes with non-equivalent mapping and for embedding them into the existing SNOMED Hierarchy. However, the vocabulary team needs to get the green light before starting this challenging task.

  • Mapping of T39.1 code in ICD10, ICD10CN, and KCD7 vocabularies is older than the respective mapping of ICD10CM. The logic was established on the relevant assumption that APAP is the most widely used 4-Aminophenol derivate and we can neglect others. But this still tends to be an open issue.

In conclusion, we can deal with the issue in two steps:
Step 1: A) to include APAP only, replace " 437158 241748001 Poisoning by analgesic drug" with “4322306 70273001 Poisoning caused by acetaminophen” OR B) to include APAP and other para-aminophenols, replace “4322306 70273001 Poisoning caused by acetaminophen” with “Poisoning by analgesic drug”.
Step 2: create OMOP Extension Standard concepts for ICD codes with non-equivalent mapping and include them into the SNOMED Hierarchy (at least, from this chapter for a start)

@Christian_Reich, @aostropolets, @Dymshyts, @Alexdavv, please feel free to share your thoughts.

Thanks Polina!

Appreciate your thoughts on this question. @Christian_Reich @aostropolets @Dymshyts @Alexdavv

Thanks,
Jing

@ljkaka, the good news is that the first step is done. It was agreed that “437158 241748001 Poisoning by analgesic drug” has to be replaced with “4322306 70273001 Poisoning caused by acetaminophen” where possible. You can take a look at this here, in Athena.

t