# Is there a loinc code I can use for an unknown observation? **URL:** \ **Category:** General\ **Tags:** fhir\ **Created:** [July 11, 2024, 1:16pm UTC](https://forum.loinc.org/t/is-there-a-loinc-code-i-can-use-for-an-unknown-observation/2037 "2024-07-11T13:16:05Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![edmondumolu](https://forum.loinc.org/letter_avatar_proxy/v4/letter/e/7c8e57/32.png) [@edmondumolu](https://forum.loinc.org/u/edmondumolu)\ **Post date:** [July 11, 2024, 1:16pm UTC](https://forum.loinc.org/t/is-there-a-loinc-code-i-can-use-for-an-unknown-observation/2037/1 "2024-07-11T13:16:05Z")
We’re working on a system that predicts loinc codes for observation resources. I wanted to know if there was an `unknown` / `unspecified` observation code that I could use for cases where our system cannot predict the code with a high enough accuracy. Thanks in advance! ---
**Author:** ![pdbanning](https://forum.loinc.org/letter_avatar_proxy/v4/letter/p/9f8e36/32.png) [@pdbanning](https://forum.loinc.org/u/pdbanning)\ **Post date:** [July 11, 2024, 6:11pm UTC](https://forum.loinc.org/t/is-there-a-loinc-code-i-can-use-for-an-unknown-observation/2037/2 "2024-07-11T18:11:51Z")
Hello and Welcome Edmond! The purpose of LOINC is to act as an identifier of the assay based on the attributes specific to that instance. As such, there is no “unknown/unspecified” observation code that could be used when the prediction doesn’t achieve high enough accuracy. I would hope the system can identify it has completed its evaluation differently than a LOINC code. Possibly a text phrase “Unable to accurately identify”, “Prediction not made”, etc.? Just a thought. Best regards, Pam ---
**Author:** ![edmondumolu](https://forum.loinc.org/letter_avatar_proxy/v4/letter/e/7c8e57/32.png) [@edmondumolu](https://forum.loinc.org/u/edmondumolu)\ **Post date:** [July 12, 2024, 6:21pm UTC](https://forum.loinc.org/t/is-there-a-loinc-code-i-can-use-for-an-unknown-observation/2037/3 "2024-07-12T18:21:57Z")
Ahh that makes sense, thanks Pam!