I am trying to map certain vitals that I receive from FHIR data using Loinic Codes that I have downloaded. I am currently mapping the Loinic codes I get from FHIR data to the Lonic codes to the display name of these vitals. However some these display names are shown as NAN.
Is this something on the LONIC side that needs to be resolved or a better implementation of how mapping should be done? I have provided an example of what I am talking about.
Hello Sheen,
The code snippet you are sharing, is this within a FHIR resource?
Hi Tim,
Thanks for the quick response. No this is not within a FHIR resource. This snippet is from the LOINIC code PDF from zip file I downloaded. I wanted to know if you have recently updated the PDF and when because I see on your site the zip file was last updated which was in August which is when I last downloaded it. Also do you happen to have an API we can call that can easily map LOINIC codes to vitals?
I would not use the PDF from our Loinc_2.81.zip distribution (or whichever LOINC version you are using). Use the LoincTable/Loinc.csv file for all the concepts. The database structure is defined in our Knowledge Base.
We have a Terminology Service (API) using HL7 FHIR which you can also use to lookup the concepts. See our documentation on this resource. You may also be interested in our Search API.
Mapping of vitals to LOINC codes is outside my expertise so I will allow any other members of our community offer their suggestions. I do know that many of the common vital observations appear within our Top 20,000 codes as recorded in the COMMON_TEST_RANK column.
Sure will look into the API. Also in regards to the Loinic codes. I was mistaken and we are using the Loinic Table/CSV for mapping purposes and that’s where we see the nan in the display of Loinic codes.
That is odd. Can you share which column names from the Loinc.csv you using for “description” and “display”?
If you’re using the import scripts defined here: Knowledge Base – LOINC then the DisplayName column will be NULL when that column is blank, hence the NaN (Not a Number) in the json. There are almost 42000 blank DisplayName rows. For these rows, the options to use for display is then to use LONG_COMMON_NAME or SHORTNAME. The only one of these to be always populated is LONG_COMMON_NAME (This is what Ontoserver uses as the Preferred display). See Shrimp/🔥
