Mapping

This section last updated: 2020-05-19 (4 years ago)

Where do I begin when I want to start mapping?

We suggest that you obtain a sample file of your HL7 messages for the area you are interested in mapping. You may want to adjust your message collection window, but we have found that a sample as short as one month captures a list of the most common terms. RELMA can create a "synthetic" master file of all tests that have been ordered and/or resulted within your time frame. You can import your local terms into RELMA from an ASCII delimited file or import local terms from files which contain HL7 messages. Complete directions for this are in the RELMA Manual (See "Import Local Terms into RELMA from Delimited File, Import Local Terms into RELMA from HL7 File). Suggested pre-mapping activities are also in the RELMA Users' guide. After completion of these activities, you should be ready to map your local terms to LOINC. The RELMA program is specifically designed to search for LOINC terms.

How do I build a test directory or master file?

You can choose the tests you want to include using RELMA to build your own test directory. There are two primary ways to build a list of local tests in RELMA:

1. Importing from a text file you create
2. Reading from HL7 messages

Your local terms can be imported into RELMA from an ASCII delimited file created from your system or from an HL7 sample message file. When creating your ASCII delimited file, your test code and test description are required fields. Optional fields are listed in the RELMA Users' Manual. The file should be delimited using the tab, vertical bar, comma or other character that is not contained in your code or description fields. Terms may also be imported into RELMA from one or more files which contain HL7 messages. At a minimum, your HL7 messages will need to provide OBX codes and descriptions. Please consult the RELMA Manual for detailed information on importing local terms into RELMA from a delimited file or a HL7 file.

When do I use a methodless code?

As a general rule, the methodless LOINC code is meant to include the LOINC codes with methods. In the hierarchy tree, the methodless terms pertain to methods that are unknown or multiple methods using the same reportable detail. Use a method specific LOINC code when the method has a meaningful clinical difference in the results. Refer to Type of Method (6th Part) of the LOINC Users' Guide for more information.

I run the same analyte on two different vendor's analzers, which LOINC code do I use?

The LOINC code may be the same for multiple analyzers due to similar methodologies or it may be different. If the analyte is reported out using the same units with both instruments and the reference ranges are the same, they may have the same LOINC code, unless the method makes a difference to the report. If the method makes a difference, you should use that method specific LOINC code. If one does not exist, you should request a new term (see proposing new terms below).

What if I have test that is only an order, only a result, or may be used as both an order and a result?

LOINC includes codes for all these situations. LOINC has codes that are only an order, such as a CBC differential count. Codes that are only a result, such as %basos. LOINC codes exists for those tests that are used both as an order and as a result, such as serum glucose.

When do I choose Ord versus Nom codes?

Nom (Nominal) really means the choice is from a selection of non-orderable entities. These coded items are not ordered items. Ord (Ordinal) codes should be used for terms that have answers that can be placed in a order, such as negative, intermediate, or reactive.

When should I choose the "XXX" LOINC code system (sample type)?

"xxx" should be used when the test you are mapping contains specimen information and is located elsewhere in the report or HL7 message. "xxx" should not be used to represent tests in the system that have a specimen in the name nor those that have an assumed specific specimen. For example, do not use "xxx" for a sodium result by itself as this is normally a serum sodium level. "xxx" should be used if that sodium level is on some other miscellaneous body fluid, not serum or plasma.

The LOINC terms in Microbiology appear to be based on what is identified, which can't be known a priori. How should I map my culture terms?

The LOINC model for routine cultures is “Bacteria identified:Prid:Pt:<specimen>:Nom:Culture”, where <specimen> = urine, stool, blood, etc.

So, for example, a routine urine culture, would be mapped to this code:

630-4:Bacteria identified:Prid:Pt:Urine:Nom:Culture

Long Common Name: Bacteria identified in Urine by Culture

For most routine cultures the users are looking for bacteria. Such cultures may also grow out some fungii, and if so they too would also be reported under this test name. But, the cultures would not be optimized for growing out fungi hence our bacteria-focused name.