This section last updated: 2020-12-07 (1 year ago)
The LOINC committee approved inclusion of standardized assessment measures (e.g., survey instruments) with version 1.0p. Representing the observations in these assessments within LOINC required a modest extension of the System axis to include aggregate units of analysis, such as ^family, and storage of additional attributes within the LOINC database. Bakken1 provides a detailed description of the methodology for inclusion and evaluation into LOINC and the extensions to the LOINC axes.
The initial corpus of material represented in LOINC came from standardized nursing assessment instruments. We have since expanded the content to cover standardized assessment instruments in many other domains. As we have added content, we have iteratively refined our modeling. Vreeman et al have published a summary of our approach and its evolution.2
The overall organization of the survey instruments and other collections are represented in LOINC using a nested panel structure consistent with the existing model for laboratory panels. LOINC terms are created for the individual questions/items within an instrument, as well as for the panels/groups of terms representing the hierarchical nature of the instrument. We recognize that standardized assessments have psychometric properties that are essential to their interpretation, and so the data model includes elements such as the actual question text and allow answer options as attributes of the LOINC observation term.
LOINC creates terms for the collection as described in the section on Order Panels (batteries). Typically, the Property and Scale attributes are a hyphen (-), because the child elements (questions) vary in these axes. We create different panel terms for different versions of the same instrument where there are meaningful changes (e.g. different questions asked, different answer choices, etc). Some assessment instruments of the same version have different “forms” for various purposes that contain unique collections of items. For example, the OASIS-C has five unique forms that represent different subsets of questions that are used at various times: Start of Care, Resumption of Care, Follow-up, Transfer to a Facility, and Discharge from Agency. LOINC creates different panel terms for each of these forms.
The LOINC table contains fields that store additional attributes relevant to many terms for individual questions or variables. Examples include the fields that store the exact question text, external copyright notice, example units, and the HL7 field where the content should be delivered (if Null, presume OBX).
LOINC does not usually create terms for information that has a designated field in an HL7 message. But, in order to create complete sets of items for some instruments we have created LOINC terms for items such as patient first name. To identify this content, we added an HL7 Field Sub ID to the LOINC table and indicate the corresponding HL7 place (e.g. PID-5.2).
The questions/items in standardized assessment instruments often have highly specialized, fixed answer lists. As described in Section 9.2, we have created a structured representation of the answer lists for the questions in assessment instruments represented in LOINC.
Some important attributes of a LOINC term for a question or variable may vary when that concept is used in different assessments or on different forms of the same assessment. For example, “measured body weight” is a variable of many different assessments. In the context of each instrument, that same concept could have different local codes, help text, validation rules, or associated branching logic. We therefore store these attributes at the level of the instance of the item in a particular panel.
As discussed in the previous sections, we have three fields that can capture the exact display of the question/item on the form in question. (For some instruments, it can be difficult to determine what exactly is the question text.) The Component of a LOINC term represents the thing or attribute being measured, and is the default for capturing the item text. However, there are several reasons why the Component may not be the exact item text. The most important ones include: our LOINC naming conventions do not permit certain characters (e.g. "/" or "?") because of our internal "Part" parsing rules, that some aspects of the question are modeled in other parts of the formal LOINC name, or that there is some important aspect to the "thing being measured" (e.g. a lookback period of the last 7 days) that is not represented explicitly on the form for that particular item.
In general, for purposes of displaying the item text as it appears on the instrument, one can follow this rule:
In addition to these fields, some LOINC codes used in survey instruments may have other LOINC name fields such as a Long Common Name and Consumer Name. These additional names may be useful in some contexts for these items, but we will still use the above rule to capture the item's representation in the instrument. Some of the original survey instruments modeled in LOINC may not follow this rule exactly, in part due to the fact that we did not have the full survey representation model as we do presently. Ongoing work includes reviewing where modifications may be needed.
For variables linked to a defined answer list that come from an external source, it has become our policy to enter the initial source as the Method. For example, variables first entered for the Minimum Data Set (MDS) will have a Method of MDS. This does not mean that the term cannot be used in any other contexts, such as other panels or surveys. As long as the defining context of use is understood, the variable can be used in other contexts where the term is exactly the same, including the answer lists, as the original term.
Beginning with the June 2015 LOINC release, the Time Aspect for survey questions that include a lookback period specify that lookback period as the Time rather than using Pt. This change was approved by the LOINC Clinical Committee at the February 2015 Clinical LOINC meeting and will apply to all new terms moving forward.
At the August 2015 Clinical LOINC meeting, further naming conventions were approved. These conventions include:
As of LOINC version 2.54, nearly all of the existing survey instrument terms that include a lookback period have an updated Time Aspect part, and we will continue to make these changes moving forward for any terms that were not updated for the 2.54 release.
Assessment instruments often contain additional text either preceding or following an item to guide users in completing the form or provide additional context about the item. LOINC uses two fields, form CodingInstructions and Context, to represent this information.
When the text gives details about completing the form, for example, "Enter 88 if patient does not answer or is unable to respond", that text is represented as form CodingInstructions.
When the text provides additional explanation or description, it is typically represented as Context. For example, along with an item recoding the number of stage 2 pressure ulcers a patient has, the form may display a clinical explanation of what a stage 2 ulcer is: "Stage 2: Partial thickness loss of dermis presenting as a shallow open ulcer with a red or pink wound bed, without slough. May also present as an intact or open/ruptured blister." Such text is represented as Context.
These form-specific attributes are displayed on the LOINC details pages and are present in the LOINC Panels and Forms File.
Beginning with LOINC version 2.26 (January 2009), an export of panels and forms content has been available as a separate download in the LOINC release called the LOINC Panels and Forms File. This package contains the full assessment content, including:
This content is available for download from the LOINC website. Please see the PanelsAndFormsReadMe.txt file in the LOINC Panels and Forms File for more details. All of the assessment content is also included within the RELMA program.
Through a collaboration with Centers for Medicare and Medicaid Services (CMS), we are updating the current LOINC representation of CMS Long-term Post Acute Care (LTPAC) assessments and creating new LOINC representations for CMS PAC assessments not currently represented in LOINC. The work includes the Nursing Home Minimum Data Set (MDS), Long-term Care Hospital CARE Data Set (LCDS), In-Patient Rehabilitation Facility Patient Assessment Instrument (IRF-PAI), and the Home Health Outcome and Assessment Information Set (OASIS).
Our work on this content began in August 2016, and starting with LOINC version 2.63 (December 2017), new and updated content represents current versions of all four CMS LTPAC assessments. As we continue to collaborate with CMS, we plan to release the LOINC representation for each new version of these assessments prior to its CMS final publication date so that the LOINC content will be available for implementation at the time of final publication.
In undertaking this work, the LOINC Committee agreed to adopt a unified Method of CMS Assessment across all instruments, rather than having a Method specific to the assessment instrument that was first modeled in LOINC. This change encourages and facilitates re-use of assessment variables across instruments where appropriate.
We have also enhanced the search functionality in RELMA and on search.loinc.org with links to the CMS assessment item codes that are searchable as "Related Codes". For example, searching "RelatedCodes: BB0700" will return the LOINC term that represents the CMS assessment item "Expression of ideas and wants".
Vreeman DJ, McDonald CJ, Huff SM. LOINC® - A Universal Catalog of Individual Clinical Observations and Uniform Representation of Enumerated Collections. Int J Funct Inform Personal Med. 2010;3(4):273-291. Epub 2011 May 23. PubMed: 22899966.↩