A LOINC code is a highly pre-coordinated, 6-axis data model, which gives me fits. My current hair-pulling exercise is providing LOINCs for a number of microarray PCR tests. Our lab operates on a more post-coordinated model, where they send specimen in a different field (like in the SPM segment), and store it in the system in a specimen table. In that regard, they will create test codes within a panel for a specific method and specimen type. Great, LOINCing is easy. Later, they might add that same test to a different panel for a different specimen. Sometimes the PCR method is different as well (probe vs non-probe). They also will “internally validate” a vendor cartridge for a different specimen, such as using the blood culture cartridge for the synovial fluid panel. There are many instances of a previously-LOINCed test becoming unLOINCable through this process.
The combinations are endless and I’m going insane trying to provide LOINC codes for them. I’m so far down the rabbit hole of varying combinations, there’s no way I’m going to figure it all out.
For example, the vendor provides this LOINC for one of the particular components of their 40+ component microarray:
Our lab has shared some of the components for CSF and Stool and Joint Fluid panels, and probably others. The methods even cross non-probe and probe PCR.
I said all that to say this. I need more generic LOINCs for PCR that are agnostic to specimen and probe vs non-probe methods.
I see the towel has already been thrown in for Micro susceptibilities (some of which are included in these microarrays). Has consideration been given to do the same thing for organism identification?
Is this a reasonable request, or do our tests need to remain without LOINCs until such time as I can convince the lab to consider LOINC more carefully during their builds? (This means currently-built tests will remain unLOINCable.)
I’m sorry to hear this is giving you fits. Past experience had LIS assays built for different specimens, to account for different reference ranges. While this might be true for chemistries and hematology, I can see it might not be followed for microbiology nucleic acid testing. Does your database administrator have an explanation for why a result field crosses CSF, Stool and fluids?
I checked for micro LOINC with specimen of Isolate and method of Molgen, to see what may intersect. The precedent there currently is for antimicrobial resistance predictions based on genes acquired by the microorganism.
I would recommend a sit down with the LIS database crew to show where their workflow processes are not enabling an international standard application and come up with improvements. Bring in the Instructions for Use to their attention and the FDA requirements on IVD vendors to designate the specifications to run an essay. These requirements shouldn’t be lessened from the FDA to the local LIS. And there shouldn’t be code re-use if the substitute reagent/kit is not exactly the same. If the LIS refuses to tighten up result field usage, you could create a submission of the generic ones you want Regenstrief to consider. This isn’t a best practice that I would encourage however. It’s also possible that the LIS module they’ve built these microarrays in is not conducive to the standard either. I’ve seen microbiology LIS modules that are stringent for stain, culture and abx…nothing for nucleic acid testing: this would be an issue to bring up to the LIS vendor.
I’m sorry to hear this is giving you fits. Past experience had LIS assays built for different specimens, to account for different reference ranges. While this might be true for chemistries and hematology, I can see it might not be followed for microbiology nucleic acid testing. Does your database administrator have an explanation for why a result field crosses CSF, Stool and fluids?
I checked for micro LOINC with specimen of Isolate and method of Molgen, to see what may intersect. The precedent there currently is for antimicrobial resistance predictions based on genes acquired by the microorganism.
I would recommend a sit down with the LIS database crew to show where their workflow processes are not enabling an international standard application and come up with improvements. Bring in the Instructions for Use to their attention and the FDA requirements on IVD vendors to designate the specifications to run an essay. These requirements shouldn’t be lessened from the FDA to the local LIS. And there shouldn’t be code re-use if the substitute reagent/kit is not exactly the same. If the LIS refuses to tighten up result field usage, you could create a submission of the generic ones you want Regenstrief to consider. This isn’t a best practice that I would encourage however. It’s also possible that the LIS module they’ve built these microarrays in is not conducive to the standard either. I’ve seen microbiology LIS modules that are stringent for stain, culture and abx…nothing for nucleic acid testing: this would be an issue to bring up to the LIS vendor.