This video shows how to use Snow Owl to search for terms within the UK Edition of the SNOMED CT terminology. I also provides a bit of background information on the content of SNOMED CT.
Once you’ve seen the video, you can find additional documentation at:
- Getting started with browsing and searching
- The SNOMED CT view: Top-level concepts and dedicated icons
- Examining concepts with the concept editor
- Quick search
- Filter search
- Advanced search
You might also be interested in our videos about browsing the SNOMED terminology and examining its concepts:
- Browsing SNOMED CT with the concepts view
- Introduction to the SNOMED CT top-level concepts
- Searching SNOMED CT: Quick search, filter search, and search profile
- Examining SNOMED CT concepts
Hello! This is Sonja from Bi Healthcare. Today I would like to show how to use Snow Owl to search for terms within the UK Edition of the SNOMED CT terminology. As we go through the examples, I will also provide a bit of background information on the content of SNOMED CT.
When you open Snow Owl you will see a screen similar to this one. This is the text field where you can enter a search term. Let’s look for Myocardial infarction. I enter myo as a search term, hit TAB to autocomplete and the matching term is already displayed in the list. All I have to do now is click the item to open the concept editor where we can find detailed information about this concept.
The concept editor consists of three different sections, which can be opened and closed using the triangles. Let’s look at the first one. Descriptions are terms or clinical phrases that describe a concept. Each concept has a fully specified name, (it’s this one here), this is the name that is unique for this concept. A fully specified name also has a term in parentheses, which is called the semantic tag. The semantic tag refers to the top-level hierarchy that the concept belongs to. In the hierarchical view we can see the SNOMED CT root concept on the top and the 19 top-level concepts, each of them representing a different hierarchy. For example, Myocardial infarction belongs to a subcategory Clinical finding.
There are also several synonyms. These are names or terms with the same or similar meaning for example Cardiac infarction, Heart attack, or MI are synonyms for Myocardial infarction. One term is marked with a rosette icon. It’s this one here called Myocardial infarction. The rosette indicates the preferred term. The preferred term is a description that is most commonly used by clinicians, it can vary from country to country. Since we’re looking at the UK Edition, it indicates the preferred term for the UK. Another country might have a different preferred term. The preferred term is also what is displayed in the search list up here, because it’s the term that is most likely searched for and known. By the way the way the terms are sorted here is based on their global usage frequency, this means that terms that are used more often in practice have a higher ranking and show up higher on the list.
Let’s look at the next section, the Properties. It displays the concepts’ relationships, this gives you information on how the concept is connected to other concepts. There is always at least one IS A relationship, which indicates the parent concept. In this example we have three parent concepts. These are concepts that are one level higher up in the hierarchy.
Let me show what this looks like. I’m just going to connect the editor with the hierarchical view and now we can see where Myocardial infarction is located within the SNOMED CT hierarchy. If we go one level higher up, we can see Myocardial disease, which is one of the parent concepts – it’s this one here.
To see other parents I can take a look at the Parents view where Myocardial infarction is on top and its parents on the bottom, so it’s exactly the other way around than in the hierarchical view. Parent concepts are more general terms. We can see that there are a lot of myocardial diseases and Myocardial infarction is one of them.
However, there are different types of heart attack, which are the children. This one has 62 children. If you hover over a concept you can see the number of children displayed. Children are more specific terms such as Acute myocardial infarction or Postoperative myocardial infarction. If I open this one and browse further down in the hierarchy, I can see different subtypes of Postoperative myocardial infraction.
Myocardial infarction has also some other relationships, which are down here. One is the Finding site, this tells us where the disease occurs, which is in the heart muscle, so in the Myocardium structure. We can also learn something about the associated morphology. This is the form in the structure of the disease, which is an infarct, this refers to tissue death.
Let’s take a look at the last section, the SNOMED CT Properties. Here we can find metadata about the concept, for example, the concept ID, which is a unique numerical identifier of the concept. We can see when the concept was published (this is the effective time), and which module it belongs to (this one here is the SNOMED CT core module).
If it was a concept that was created for a particular country or region, for example, the UK you could see this here as well. Let’s look at an example for a concept that is specific for the UK, the Pain self-efficacy questionnaire. I don’t have enter in the entire term I can just use an acronym (PSEQ) and here’s already the concept that I was looking for. We can see that this is the United Kingdom clinical extension module and if we look at the hierarchical view, we can see that the icon has a little flag of the United Kingdom. This way we can very easily distinguish which concepts are specific for the UK Edition of SNOMED.
OK, let’s close the hierarchy and just look at the top-level concepts. As I already mentioned the fully-specified name is unique, so two concepts cannot have the same fully-specified name. However, the same preferred term can have different meanings. Let me give you an example. If I enter the term Dressing, it shows up three times in my list: dressing, dressing, dressing. So I don’t know which is the dressing I’m looking for.
These icons help to identify what kind of concept we are looking at. This one – the magnifying glass – refers to the top level category Observable entity. This would be dressing as a personal care activity – so getting dressed. This one is a Qualifier value that refers to a unit of drug administration. And this one is the Physical object that is applied on a wound. It has the icon for the glasses. The icons help to instantly pick the concept that you’re looking for, it’s not necessary to open each of them up to find out what category it belongs to.
It’s also possible to restrict a search, for example, if you were looking for all procedures associated with dressing, you can use this button that opens up a search dialogue and enter a search term here (Dressing) and then restrict your search to a certain top-level concept. There are a lot of other possibilities but I just wanted to show you this one. Click search to display the results in this window here (it can be maximized), we have 93 results. To take a closer look, I just double-click an item and this opens up the concept editor about the procedure Application of dressing.
Even within the same category you can have synonymous terms. A good example for this case is Fundus, which refers to the bottom section of an organ and dependent on what specialist you are talking to the term might refer to different organs. But since the preferred term (and not the synonym) is displayed here in the list you can actually instantly see what kind of fundus this concept refers to. For example, fundus of eye, fundus of gallbladder.
It’s also possible to create a customized search profile this means that your concepts are sorted based on your specialties. For example if you are an eye doctor, you might modify your search results so that this fundus would show up higher in the list or you could exclude certain terms, for instance the non-human terms, from the non-human reference set. Then this fundus here would not be displayed at all because it’s a veterinary term.
Finally, I would like to show searches for reference sets. There’s a reference set called Occupational therapy assessment scales. Here are the SNOMED CT concepts (I just entered occupational), this retrieved all terms with occupational. In this category down here you can see our reference set. These results are from ICD-10 codes, because the search extended to other terminologies as well.
But let’s look at the reference set (I can maximize the view). This reference set here has 117 members. We can do a filter search that only searches within this reference set. Let’s look for the Allen Cognitive level screening tool, that’s already here. If I click it, we can see detailed information displayed about this concept.
If I did the same kind of search up here, in the quick search, it would search the entire ontology. Let’s see what happens: Here are also the two results that I had in the reference set, but I also have some results in grey. The grey font indicates that these are not exact matches but approximate matches. I have an Observable entity, which is the score. I also have a Procedure, the assessment procedure of actually administering the Allen cognitive level screening test.
That’s all I wanted to show today. It was just a short introduction. If you want more detailed information, please go to our website at http://b2i.sg. There are lots of videos there and documentation about different kinds of searches you can do with Snow Owl and about the other features of Snow Owl. I hope you found this video useful and interesting and thank you very much for your attention.