Page tree

2486 View 31 Comment In discussion Comments enabled In the category: Undefined

This discussion thread was created to share some initial findings and start a discussion about how we might make concepts related to assessment scales and score more fit-for-use in clinical care (and research to accelerate knowledge discovery to improve clinical care). 

Assessment Scores

There are currently many, many concepts in the observable entity (OE) hierarchy for assessment scores.

  • Some concepts are child concepts of | Assessment score (observable entity) |
  • Other concepts are child concepts of more specific observable entity concepts
    • e.g., 105 assessment score concepts that are direct child concepts of | Mental state, behavior / psychosocial function observable (observable entity) |

If there is any Editorial Guidance around how assessment score observables should organized in the hierarchy, can we add it to this thread?

I extracted all assessment score concepts in the | Mental state, behavior / psychosocial function observable (observable entity) | hierarchy and started a very preliminary categorization of assessments based on the construct they measure. I would like to work with others who may be interested to more formally classify the instruments so we can represent them more precisely in the terminology. The document is attached, and can also be accessed here

Assessment scales 

See attached spreadsheet of assessment scale concepts extracted from SNOMED. This document can also be accessed here

Most of the concepts we have in this hierarchy are not specific enough to allow us to identify the specific assessment instrument. See table 3 in this document for an example of the many different versions of the Hamilton Rating Scale for Depression.  

Recommendation

Create a formal project plan for evaluating assessment instrument content in SNOMED. 


Contributors (6)

31 Comments

  1. Hello Piper, the UK would support the development of a formal project plan to develop editorial guidance and ideally modelling guidance for this area of SNOMED CT. 

    It is a question / problem for the UK on how to model and parent this type of concept within the context of the larger observable entity model. 

    We would be willing to actively support such a project with some time from our authors. 

  2. Terrific.

    Thank you Stuart! 

    I think we need to set up a document sharing space, rather the cluttering Confluence with working documents. 

    I will work with Elaine Wooler and Monica Harry to get some guidance around how to proceed.

    Piper

  3. We have two people interested, Fatima Haidaree and Emma-jane Macdonald .

    Please let them konw once we have a Working Group.

  4. Thank you Stuart Abbott.

    I'll send an initial email to find a time for a kick off meeting. 

    Best,

    Piper

  5. Stuart Abbott , Fatima Haidaree , Emma-jane Macdonald ,

    Are you able to access Microsoft Box for shared documents, or is Google Drive better for you?

    Thanks,

    Piper

  6. Hi Piper Allyn Ranallo in theory we can access both, im not familiar with box though, but im sure it wouldnt be too difficult - may be a test document to start?

  7. Hi Emma-jane Macdonald 

    I have created a Box folder here and a folder in the MABH-CRG google drive here.

    I believe I set up the Box folder so anyone with a link can access it, but I may need to add email addresses. 

    Let me know if you have any trouble access either location.

    My direct email address is sven0018@umn.edu.


    Best,

    Piper

    1. sorry Piper Allyn Ranallo  i didnt get a notification for this. 

      Box - I have signed up  - but get an error message saying i dont have access - as you say I guess you would need to add email addresses emma-jane.macdonald1@nhs.net

      Drive - I can see a folder and a doc - but cannot add anything - again i presume you woudl need to authorise me? - ejmacdonaldnhs@gmail.com

      email - Ihave sent a test email so we are in contact

      Thanks

      1. Thanks Emma. 

        I've added your email address to both Box and Google Drive. 

        Let's give those both a shot and see whether you have full access.

        Piper

        1. super - ive uploaded a blank test doc to test the functionality

  8. Hi all.

    Wow - the categorisation work already done in the two 'SNOMED assessment...' excel files is impressive!

    I wonder if the instruments listed on the NHS's NCCR library site will complement those from OMOP and RDoC in identifying priority areas to test the approach? If needed I'm sure we can coerce the page into another spreadsheet.

    As noted in the proposal there tends to be a lack of subscales (sometimes related to licensing concerns). The following 'observable'-based refset includes a lot of UK extension sub-score codes which might help highlight those instruments where it is felt this extra precision is needed for meaningful recording:

    991471000000104 |Mental health assessment observables simple reference set (foundation metadata concept)|

    There are other related refsets but this one seems to be the most useful starting point.

    Ed

  9. Ed Cheetham -

    Thank you for sharing the refset and the link to the NCCR library. I was not aware that either existed.

    This is exceedingly helpful!

    May I add your name to the list of folks interested in working on this content? 

    Piper

    1. Hi Piper.

      Yes, please do add me. I've spent today playing with chatGPT to see if it can help categorise these long lists of assessment instruments and I'd love to feed back what I've done!

      Ed

  10. That's brilliant Ed! 

    I can't wait to see what you find. Thanks for being one of us (smile)

    Piper

  11. Hi Piper Allyn Ranallo  unfortunately the link re Hamilton Rating Scale for Depression is no longer workign for me. Please could you put the relevant document on Box or Drive? Thanks

  12. Hi Emma-jane Macdonald -

    I copied a folder named 'Library' into the Box Drive.

    It needs some organization, but this has contains a) publications about assessment instruments and b) copies of assessment instruments. 

    We'll need to figure out a good way to organize links, documents, etc.

    Piper

    1. sorry Piper Allyn Ranallo  i dont know why - but im not getting these notifications .

      I have uploaded document " Project proposal_NHS England" to Box. Hopefully covers the angle Fatima and I are coming from.

      1. Hi Emma-jane Macdonald -

        Not sure why you aren't receiving notification or confirmation either.

        I do see this document, so can confirm you uploaded it successfully.

        Piper

  13. I've posted some historic and background stuff on Box (in 'Background' folder):

    • 2005: An unattributed 'Proposal for representing SNOMED CT Assessment Scales in HL7 V3'
    • 2007: A text archive of the original 'Using SNOMED CT in HL7 Version 3' implementation guide (very old, none of the links work, not a lot relevant to this work but section 3.3 deals specifically with '3.3 Assessment Scale Results').
    • 2009/10: A UK document titled 'Use of SNOMED CT UK Edition for Scored Assessments Implementation Guidance' - a long (and I'm not sure how enlightening now) document. Quite a lot on explicit version identification.
    • 2020: UKSNOMEDCT_Editorial_Principles_v26 - largely restates the international position in section 6.2.

    It's also worth considering these two external resources:

    • HL7 FHIR's Questionnaire and Questionnaire Response resources; I note the former has a version attribute.
    • openEHR's CKM provides glimpses of how they handle structured instruments (just, for example, type 'scale' in the search box on the left) but I cannot see a general pattern for assesssment instruments.

    Downstream analysis standards such as OMOP CDM may have target patterns that are of importance here too.

    The reason I mention these is to keep in mind the opportunity for complementary approaches to representing the various facets of assessment instruments, but history tends to tell us that most things end up (by design and because of heterogeneous user groups) duplicated across terminologies and information models so we're probably on our own however unsatisfactory that may seem.

    I've also posted (in the 'EdData' folder) an attempt to associate SNOMED CT 'assessment' codes from the various chapters (trying to line up, programmatically, each scale, procedure and observable 'set'). The 'TripleMatch' tab is where I was able to associate all three, 'OneMatch' includes incomplete match rows and 'NoMatch' is, well, codes with no match. There are, of course, a lot of false positives and negatives, but it might help bootstrap any work that needs to do such an association, and certainly gives some indication of the scale of the problem.

    It'll be easier to explain what I've done on a call, but the main build code is in 'instruments.py' and 'instruments2.py' in this repo. I've run it against the latest UK data because I know we have added a lot of extension content in the past to try and complete each scale-procedure-observable 'triple' (but this does mean an international group will be looking at a lot of UK codes!). It's worth noting there are smatterings of relevant 'findings' content (such as scored GCS values) in SNOMED CT but considering them is for another day.

    I note the GCS content was discussed a few years ago in the Anaesthesia group (you need to be logged in to see the discussion) which is a reminder of how many groups have requirements in this area and how little agreement there still is.

    Ed

    1. Ed Cheetham -

      This is great information.

      Thank you for sharing this! 

      Piper

  14. FYI - came in an email from our rep. sorry for posting in full - link not working

    Dear Member Forum,

    I hope this message finds you well. 

    Following the productive face-to-face meeting in October, during which we discussed the requirement of integrating assessment scales into SNOMED CT, we are writing to provide an update and request your input.

    As agreed upon during our meeting, Jane Millar and Jim Case will collate your requirements and explore further how to manage the addition of in scope content into SNOMED CT. We are pleased to announce that we are now in a position to move forward with this work.

    As next steps, we kindly request that you share the following details at your earliest convenience:

    • The specific assessment scale(s) you require.
    • The underlying reason or use case for the incorporation of this assessment scale in SNOMED CT.
    • The level of detail of the specific assessment you envision being available in SNOMED CT.

    Additionally, should you have any specific issue in relation to the associated Intellectual Property of an assessment scale, it would be helpful to understand this aspect too.

    We appreciate the proactive engagement of members who have already submitted their requests, however, to guarantee that we do not overlook any previous input, we kindly ask that those who have submitted requests and information resend them so that we ensure we have a full picture of Member requirements.

    Please forward your request to Jane Millar at  jmi@snomed.org

    Your input is appreciated so that we can get a picture of requirements at the international level and options for going forward. We anticipate providing an update on this issue during the upcoming face-to-face meeting in April. 

    Thank you for your contribution to this topic which we know created a lot of interest at the October meeting.

    Kind regards,

    Anne Højen

    1. Emma-jane Macdonald,

      Thank you for posting this.

      I tried to link to the following page: Follow-Up on Integration of Assessment Scales into SNOMED CT but was unable to access it.

      Jane Millar, Jim Case - is there a way to get folks in this group access to this page so we can begin to follow SI work about assessment scales?

      Also... it sounds like there is already a project in place to address assessment scales in SNOMED, and therefore this group does not need to propose a formal project, bur rather work to provide input to the terminologists assigned to the project. How can this group be most helpful to SI as you work to address the content?

      Best,

      Piper



      1. Piper Allyn Ranallo 

        To my knowledge there is not yet a formal project to address assessment scales.  The information we are requesting is to help provide the substance needed for a formal project proposal.  We have not gotten a comprehensive list of requirements and use cases that we could use to guide the approach to managing assessment instruments and scores beyond what we already provide in SNOMED, which is why we need the project proposal.

  15. Jim Case Thank you, this helps.

    Just so we're clear on how we can be most helpful in getting this project prioritized and moving the work forward - Do you need us to respond to the 3 bullet pointed items in the letter, and send this information to you and Jane Millar? Or do you need someone to prepare and submit a formal project proposal using the SI project template?

    Thanks,

    Piper

    1. Please send the information related to the three bullet points and we can work with you to develop the project proposal.

      cc: Jane Millar 

        1. Piper Allyn Ranallo  there is an agenda item re assessment scales on the upcoming SI member forum business meeting. 

          Did you send information re the bullet points to Jim?

  16. Wonderful discussion!

    I'd like to share an innovative approach we've been taking that uses self-reported data based on a scale of the degree of affective pleasantness or unpleasantness related to meaningful life situations experienced within a biopsychosocial context. These emotional valence data are analyzed to produce measures of psychological well-being (the subjective state of overall life satisfaction from good physical, mental, and emotional health) that can be tracked over time. Etiological factors can be determined via an in-depth mindset assessment we're developing that defines a person's adaptive and maladaptive thought patterns and mindset-shifting recommendations.  

    I realize this is a significant departure from conventional psychological assessment approaches. Still, it can provide valuable diagnostic information that offers practical support for mental health therapeutic practitioners' clinical insights and decisions. 

    I welcome questions and comments.

    Piper Allyn Ranallo Jim Case Jane Millar Emma-jane Macdonald Ed Cheetham Stuart Abbott 

    1. Hi Stephen Beller

      Great to hear from you again. This sounds like interesting work!

      Just to clarify, the work this group is doing around assessment scales is related to how they will be represented in SNOMED, and ensuring that procedures and scores for assessment instruments used in clinical practice can be encoded using SNOMED concepts. The group is not developing or evaluating the quality of assessments - just working to ensure we have adequate coverage and representation in the terminology.

      That said, once your assessment instruments is published, it can be submitted for inclusion in SNOMED so those who use it in clinical practice can encode associated procedures and scores.

      Warmest regards,

      Piper

  17. All -

    Since we first met the following have occurred:

    1) SI made some decisions and announced some guidance around assessment scales (no concept model will be created, guidance provided around how to represent assessments and the associated observable and procedure concepts)

    2) Ed Cheetham has developed some methods and tooling for automated extraction of candidate constructs measured by specific assessment instruments

    This is critically important b/c we still have a lot of work to do to clean up the hierarchical representation of both the scores (observable entities) and the assessment scales 

    3) I created and have been testing a very preliminary concept model for the scales (to propose to SI in the event they are willing to consider one day implementing a concept model for assessments);  We will need to demonstrate the value for patients, clinicians, and researchers to justify the resources required to implement some final, production version of this model SNOMED 

    I would like to share the model onour next call -  I will send an email with a doodle poll to get folks' availability

    Finally, wondering if anyone knows how (and when) existing LOINC Survey content and SNOMED assessments scales and scores will be integrated in the LOINC-SNOMED ontology?  Jim Case not sure if any decision has been made yet? 

    Best,

    Piper


    1. Piper Allyn Ranallo 

      The representation on survey instruments in the LOINC extension was discussed at our last face to face meeting in March.  The consensus was that the LOINC extension will subscribe to the proposed policy for assessments for the International release, for which the LOINC extension itself would only represent the name of the assessment and the score.  If LOINC has obtained permission for the individual questions from an instrument to be added, then those would also be added to the extension as would the value sets associated with each question.  However, what would not be added to the extension would be anything related to the structure of the instrument (i.e. order of questions) or any related logical structure (e.g. skip logic).  Thus it would not be possible, using the extension alone, to recreate the native instrument.  This was determined to be an implementation issue outside of the purview of a terminology.

      Aside from what already exists in the International release of SNOMED CT, all additional survey content will reside in the LOINC extension.  It has been recognized that we may need a "new" concept model to represent assessments and surveys and that is a future effort in the LOINC extension project as we are currently working on getting a more comprehensive transform of laboratory LOINC terms for the first release.  So, both "how" and "when" are not yet determined.