CDISC have announced a public review of LOINC TO LB Mapping. Niels Both, our Partner and longtime CDISC Consultant, considers the practical implications for your work.

LOINC coding provides a complex and sophisticated system to precisely identify lab parameters. For several years, LOINC codes have been a part of SDTM LB domains as a permissible variable in LB. This is the case, at least since SDTM IG version 3.1.0, as far as I am aware, and means it has been in use for more than a decade.

Personal Experience

It is my experience that it was rarely used in SDTM, and that the typical knowledge about LBLOINC amongst SDTM users stretched no further than: “LOINC is some kind of coding for lab parameters – and we are not using it!” This notion is changing! In 2017, FDA showed public interest in LOINC codes for SDTM lab data (https://www.fda.gov/media/109376/download) and is now requiring their use for all studies starting after 15th March 2020.

Visit LOINC

I would recommend that all programmers and data managers in the pharmaceutical industry visit loinc.org to gain a basic insight into the LOINC codes. For people that are more curious about the details, there is some fantastic supporting material available there. And if you are used to working with lab data and mapping lab data, try the LOINC search engine (https://search.loinc.org/searchLOINC/).  I personally learned a great deal about the system, by just browsing the dictionary via that search engine.

Technical Conformance Guide

In the most recent technical conformance guide FDA provides additional guidance:
When submitting LOINC codes you should:

  1. Continue submitting laboratory data in the CDISC SDTM format using CDISC laboratory terminology alongside the LOINC code for a given laboratory test.
  2. Enter LOINC codes in the LBLOINC field of the SDTM LB domain and populate LBMETHOD when available. When LOINC codes are unavailable, leave the field blank.
  3. Submit LOINC codes only when they are available from the clinical laboratories as a pass-through only, i.e. reporting the codes as received from the laboratories with no modifications. FDA understands that there may be inconsistencies in the specification and interpretation of LOINCcodes submitted across tests, studies, and subjects.
  4. Provide in-vitro diagnostic (IVD) device information in the SDTM Device Identifiers (DI) domain, when available. This information will help inform further FDA guidance on the consistency of LOINC codes associated with laboratory devices

From my perspective, numbers 3 and 4 are particularly interesting. I am pleased to see in number 3 that FDA recommends to only set LOINC codes prospectively / upstream so that it is the experts in the laboratories doing it, rather than a lab experienced Data Manager or statistical programmer doing it retrospectively. However, number 4 might become a challenge. Such technical metadata about the lab equipment/devices are not typically delivered, as part of the raw data from a laboratory vendor to a company in the pharmaceutical industry. It’s an interesting challenge and something the industry will have to address in discussion with FDA.

So what should you do if you are working with LOINC & SDTM?

I have three recommendations for a pharmaceutical company to consider regarding LOINC at this point in time:

  1. Make sure Data Managers and Statistical Programmers are trained so they have a basic knowledge about LOINC codes for LB.
  2. When agreeing lab data transfer specifications with your lab vendors, make sure to include LOINC codes and the IVD device information.
  3. Remember – LOINC is a powerful system to identify lab assessments. It might prove to be very useful to have on your datasets, once processes has been established.

If you have any questions or require any support, do get in touch and me or one of our other experienced consultants and we will be pleased to help and take a look at our CDISC page to find out more about our work.