I cannot find the Leukemia data at present - I'll follow that up and get back to you.
For creating a cohort in general, using Lip Neoplasm as an example :
Use UKB showcase to identify the fields that will be useful. For example, if you are looking at cancers you might read category 100092, field 40011 and data-coding 38, or possibly Field 40006 and data-coding 19.
Suppose you want lip cancer, that would be code C00 in coding 19 for field 40006.
In your project on the RAP, double-click the name of the dataset record, to get to the cohort browser.
Select Data Preview tab.
Select Add Filter,
navigate through Health related outcomes > cancer register > Type of cancer ICD10 | instance 0
select Add Cohort Filter
enter C00, select Malignant neoplasm of lip, ignore the "excludes" option,
select Apply Filter.
Repeat for Instance 1.
Click on the AND between the two filters to change it to OR.
Repeat for Instances 2 to 21.
Click the Save icon, enter a name for your cohort, save to your project.
For those fields such as cancer occurrences that have a lot of Instances, creating a cohort using the cohort browser can be a bit long-winded, and you might prefer to use a JupyterLab instance to run a python script to select your cohort instead.
the leukemia data is a bit odd, in that the codes don't show up in the List View within the Apply Filter. However, the data is present, and can be used for filtering. Instead of entering C00 (per instructions above), try entering acute , and selecting each of the acute myeloid leukemia options.
I will follow up on why the codes don't show up, but it may take a while to sort out.
I tried the C92 (myeloid leukaemia) code in a filter of p40006_i0 again today, and it seemed to be fine. I don't know what was wrong last week (user or system) but it seems to be fixed now.
Comments
4 comments
I cannot find the Leukemia data at present - I'll follow that up and get back to you.
For creating a cohort in general, using Lip Neoplasm as an example :
Use UKB showcase to identify the fields that will be useful. For example, if you are looking at cancers you might read category 100092, field 40011 and data-coding 38, or possibly Field 40006 and data-coding 19.
Suppose you want lip cancer, that would be code C00 in coding 19 for field 40006.
In your project on the RAP, double-click the name of the dataset record, to get to the cohort browser.
Select Data Preview tab.
Select Add Filter,
navigate through Health related outcomes > cancer register > Type of cancer ICD10 | instance 0
select Add Cohort Filter
enter C00, select Malignant neoplasm of lip, ignore the "excludes" option,
select Apply Filter.
Repeat for Instance 1.
Click on the AND between the two filters to change it to OR.
Repeat for Instances 2 to 21.
Click the Save icon, enter a name for your cohort, save to your project.
For those fields such as cancer occurrences that have a lot of Instances, creating a cohort using the cohort browser can be a bit long-winded, and you might prefer to use a JupyterLab instance to run a python script to select your cohort instead.
Hello again,
the leukemia data is a bit odd, in that the codes don't show up in the List View within the Apply Filter. However, the data is present, and can be used for filtering. Instead of entering C00 (per instructions above), try entering acute , and selecting each of the acute myeloid leukemia options.
I will follow up on why the codes don't show up, but it may take a while to sort out.
I tried the C92 (myeloid leukaemia) code in a filter of p40006_i0 again today, and it seemed to be fine. I don't know what was wrong last week (user or system) but it seems to be fixed now.
Got it. Thanks a lot!
Please sign in to leave a comment.