Daniel Vreeman

Forum Replies Created

Viewing 15 posts - 16 through 30 (of 54 total)
  • Author
    Posts
  • in reply to: Radiology Order Panels #16703

    Daniel Vreeman
    Moderator

    Take a look at the Radiology tutorial slide deck that is available at:

    http://loinc.org/slideshows/clinical-loinc-tutorial/2013-02-14-clinical-loinc-tutorial-imaging.html

    Then, take a look at the very detailed discussion of Radiology naming conventions in the LOINC User’s Guide (Chapter 4.3):

    http://loinc.org/resolveuid/5b10b078cb8a73852124468027bc9808

    Before you make a submission for new terms, you should scour the examples of radiology terms in LOINC already (there are about 5,000 of them) to be sure the one you’re requesting isn’t already there. You’ll find them in the class of RAD:

    http://search.loinc.org/search.zul?query=class%3Arad

    When you find a gap, then use the naming conventions and patterns above to complete your submission. So, the analyte = the piece of the LOINC Component before the Challenge. The method is the modality.

    in reply to: Web Service to look up code definitions #16702

    Daniel Vreeman
    Moderator

    Sorry, Regenstrief doesn’t produce a publicly available web service or API.

    in reply to: Semantic scope of a code #16666

    Daniel Vreeman
    Moderator

    From LOINC’s perspective, I think it is probably your choice. (Not sure what HL7 SDWG would say). The LOINC code is meant to identify the expected information contents of that item. LOINC’s definition of HPI (which could be improved) is based on the general use of the phrase in medicine…which employs a broad definition of “illness” that would include trauma. In the context of trauma, an HPI might include things like the mechanism of injury, timing of the injury, situational details (entrapment, airbag deployment, etc), other prehospital assessments,etc (per The trauma manual: trauma and acute care surgery). So, I think it would be reasonable to use the HPI template if it fits your needs. The advantage of using an injury-specific code would be that you could find it much more easily amidst all the other HPI’s.

    in reply to: New LOINC terms #16654

    Daniel Vreeman
    Moderator

    I must offer the caution that there are of course lots of important changes in each release (deprecated terms, edits, etc) that are contained in each release, so just looking at new terms added may not be what you need. That being said, the easiest way to find them would be just to run a query between the LOINC table from 2.40 and 2.42. Here’s how you might do it if you are using Microsoft Access:

    Link the tables on the LOINC_NUM field, right click the relationship, and choose Join Properties of “Include ALL records from”…of the LOINC 2.42 table. Then just put both LOINC_NUM fields (and whatever else you want) in your query and indicate the LOINC_NUM field for LOINC 2.40 to be “Is Null”:

    That’ll give you this result, which is the 775 as indicated in the release notes

    in reply to: Harmonization between LOINC and SNOMED CT #16576

    Daniel Vreeman
    Moderator

    You can find updates on the collaboration here: http://loinc.org/collaboration/ihtsdo


    Daniel Vreeman
    Moderator

    We would certainly encourage the construction of local use-case specific hierarchies based on whatever roll-ups you felt necessary. For that you could use the LOINC parts or the strings to do it (e.g. anything with a component of “Admission Evaluation Note”). We also have a recently implemented policy about creating a “generic” term wherever we have document terms that specify a setting or provider. This generic term could be used as a generic parent for the more specific types. E.g. this term:

    67851-6 Admission evaluation note Find Pt {Setting} Doc {Provider}

    Could be thought of as the generic parent for these more specific codes:

    34744-3 Admission evaluation note Find Pt {Setting} Doc Nursing
    34873-0 Admission evaluation note Find Pt {Setting} Doc Surgery
    68552-9 Admission evaluation note Find Pt Emergency department Doc Emergency medicine
    67852-4 Admission evaluation note Find Pt Hospital Doc {Provider}
    68471-2 Admission evaluation note Find Pt Hospital Doc Cardiovascular disease
    64058-1 Admission evaluation note Find Pt Hospital Doc Critical care
    64053-2 Admission evaluation note Find Pt Hospital Doc General medicine
    68483-7 Admission evaluation note Find Pt Hospital Doc Medical student.cardiovascular disease
    64070-6 Admission evaluation note Find Pt Hospital Doc Medical student.critical care
    64054-0 Admission evaluation note Find Pt Hospital Doc Medical student.general medicine
    64066-4 Admission evaluation note Find Pt Hospital Doc Medical student.general surgery
    64078-9 Admission evaluation note Find Pt Hospital Doc Medical student.pulmonary disease
    64074-8 Admission evaluation note Find Pt Hospital Doc Medical student.thoracic surgery
    34862-3 Admission evaluation note Find Pt Hospital Doc Physician attending.general medicine
    64062-3 Admission evaluation note Find Pt Hospital Doc Pulmonary disease
    64060-7 Admission evaluation note Find Pt Hospital Doc Thoracic surgery

    We don’t recommend using the Part codes directly as a “observable” concept code because they don’t have the same change management principles applied (e.g. deprecation, concept permanence, etc) as the LOINC codes do. They have several purposes…some internal.

    We would, of course, welcome new submissions based on your findings.


    Daniel Vreeman
    Moderator

    There is no registration fee and I don’t believe that there is any restriction on who can attend other than that you’ll have to create an account on the Sysmex website (they are hosting it). When I registered on their site, I believe the “country” list is limited to a subset of North American countries (presumably their market area), but Canada is there, so I’d say invite away!

    in reply to: Vitamin D2 and D3 #16564

    Daniel Vreeman
    Moderator

    Take a look at this document from the June 2010 Lab LOINC meeting that has some explanations about LOINC vitamin D terms:

    http://www.loinc.org/meetings/20100607/files/handout-vit-d-explanation.pdf

    in reply to: Excel Export of Panel Order Set/Components #16567

    Daniel Vreeman
    Moderator

    No need to post your question in every forum…just one will do.

    From within RELMA, just search for the panel term you want. Right click the row, then choose “View Panel Children”. That will pop up a window with all the child elements of the panel in a grid. From there you can select all the rows, choose Custom Export to get all the LOINCs in a variety of formats (copy to clipboard, export to excel, etc).

    You may also want to check out the LOINC Panel’s and Forms File under Accessory Files from http://loinc.org/downloads for an export of some of the key panels and forms.


    Daniel Vreeman
    Moderator

    Yes, you can use the CLASS field of the LOINC database for exactly this purpose. Take a look at Appendix B of the LOINC User’s Guide for a detailed listing of the classes. If you are using RELMA, you can restrict your searches to any particular class (or any other hierarchy node) by going over to the “hierarchy and search restrictions” tab and selecting the node at which you want to limit your search.

    The upcoming RELMA release (5.0) will have a way to do this from the search command line, so stay tuned…


    Daniel Vreeman
    Moderator

    You are correct in that not all LOINC codes are present in the multiaxial hierarchy. It includes lab, radiology, clinical note titles, and some other clinical variables, but not the whole database. So it is not a matter of syncing, but rather the scope of content included in the multi-axial hierarchy to date. This is largely a matter of focus and resources; we don’t have a defined plan for adding all LOINCs to this hierarchy. The other hierarchies are stored in the database that the RELMA program uses, but that is available for direct use only under specific permission from Regenstrief (see http://loinc.org/terms-of-use). Those hierarchies are not in the same format as the multi-axial export and could change over time (because they are used by RELMA). We have discussed exporting all the hierarchies in the same format as the multiaxial one, but haven’t nailed that down yet. Is that something you would use?

    in reply to: Short Name vs. all components in HL7 #16543

    Daniel Vreeman
    Moderator

    The short answer is that there is no ‘right’ answer for all circumstances.

    Historically, we (LOINC) have recommended the use of the ShortName (over the colon-separated fully-specified name) in HL7 messages because it fit within the space allocated by most lab reporting systems (30 characters), would work as a column name on a flowsheet, and used common acronyms people would recognize. I think the fully specified name has more instances of ‘reserved characters’ like “^” and “&” which would need to be properly escaped, and we’ve heard a few folks mention difficulty with their systems doing this…though there is a clear cut way to do it.The base HL7 2.x standard doesn’t specify (and actually would be valid if you sent your own label as the display name associated with a LOINC code — something we don’t recommend). More recently we developed (and now have fully-populated) a Long Common Name. These are probably more understandable to human readers and might be preferable, but they can be quite long, so some systems may not accomodate. The HL7 implementation guides for certain use cases (e.g. ELR to public health) may or may not constrain the freedom of the base standard with respect to the preferred display name, thus leaving the choice up to the exchanging partners. This has also come up in the recent EHR certification testing specs, and we have made comments to NIST about it. This display name issue is being discussed as a project in the HL7 vocab workgroup because it is not unique to LOINC.

    So, if there are no other constraints, I’d recommend the Long Common Name, but the Short Name or Fully-specified Name are valid in most contexts too. Since the Long Common Name has only recently reached a state of moderate maturity, we haven’t yet updated the LOINC Users’ Guide, but we probably will revise it to include a summary of the above sometime in the near future. We also definitely recommend the simultaneous messaging of the senders local code and local name (in addition to the LOINC code and Name) to facilitate debugging and detection of mis-mappings.

    in reply to: Orders in LOINC database #16538

    Daniel Vreeman
    Moderator

    Yes, the common order set is small by design…to highlight the most frequently used terms and help prioritize the mapping effort towards those high impact items.

    in reply to: Orders in LOINC database #16541

    Daniel Vreeman
    Moderator

    We populate the ORDERS_OBSERVATION field for all lab terms in the database. The Common Lab Orders value set is purposely a short set of codes that account for the vast majority of lab orders in the US. It was derived by both empirical and consensus-based approaches. You should consider it a minimun “starter set” and definitely doesn’t include all possible LOINC codes or lab orders. You can read more about how it was derived in the Preface, which you can get from:

    http://loinc.org/usage

    in reply to: Orders in LOINC database #16539

    Daniel Vreeman
    Moderator

    This field is our best (informed guess) as to whether this LOINC code can be used as an Order code (e.g. in an OBR segment) and Observation code (e.g. in an OBX segment), or Both. Many terms can be used as both, panel terms will Order only, and things like calculations or impressions will typically be Observations (you can’t order them directly).

Viewing 15 posts - 16 through 30 (of 54 total)