Student 1 Verification: Crohn's disease (Conditions)

[callahantiff]

Student (GitHub Username): @emswen Verification Number: 1

Verification Assignments: Crohns_Disease_Conditions

• Phenotype Definition
• Spreadsheet

[emswen]

@callahantiff - for this set should I be determining whether the source string and code are a good match or should I be comparing the code to the phenotype (Crohn's disease)? For example, the data set includes the source code "felty syndrome" which is not a good match for Crohn's disease however is a good match when compared to the source string "felty"...

[callahantiff]

@callahantiff - for this set should I be determining whether the source string and code are a good match or should I be comparing the code to the phenotype (Crohn's disease)? For example, the data set includes the source code "felty syndrome" which is not a good match for Crohn's disease however is a good match when compared to the source string "felty"...

@emswen - Great question, this task will actually involve both of the things you describe; thinking about the matches that are returned within the context of the the phenotype definition and the other codes/information we are provided.

For the example you raised, the string "felty" is in reference to criteria that is used to exclude control patients. You can get to these specific details by clicking on the hyperlinks in the phenotype definition. So, you are right that it is not a good match to Crohn's disease, but in this case, it is a good match to Crohn's disease control patient exclusionary criteria.

Does that help?

[emswen]

@callahantiff - for this set should I be determining whether the source string and code are a good match or should I be comparing the code to the phenotype (Crohn's disease)? For example, the data set includes the source code "felty syndrome" which is not a good match for Crohn's disease however is a good match when compared to the source string "felty"...

@emswen - Great question, this task will actually involve both of the things you describe; thinking about the matches that are returned within the context of the the phenotype definition and the other codes/information we are provided.

For the example you raised, the string "felty" is in reference to criteria that is used to exclude control patients. You can get to these specific details by clicking on the hyperlinks in the phenotype definition. So, you are right that it is not a good match to Crohn's disease, but in this case, it is a good match to Crohn's disease control patient exclusionary criteria.

Does that help?

I think so... just to verify - for a code that is a good match to control patient exclusionary criteria (as in the example above) which drop-down phrase should we be selecting?

[callahantiff]

@callahantiff - for this set should I be determining whether the source string and code are a good match or should I be comparing the code to the phenotype (Crohn's disease)? For example, the data set includes the source code "felty syndrome" which is not a good match for Crohn's disease however is a good match when compared to the source string "felty"...

@emswen - Great question, this task will actually involve both of the things you describe; thinking about the matches that are returned within the context of the the phenotype definition and the other codes/information we are provided. For the example you raised, the string "felty" is in reference to criteria that is used to exclude control patients. You can get to these specific details by clicking on the hyperlinks in the phenotype definition. So, you are right that it is not a good match to Crohn's disease, but in this case, it is a good match to Crohn's disease control patient exclusionary criteria. Does that help?

I think so... just to verify - for a code that is a good match to control patient exclusionary criteria (as in the example above) which drop-down phrase should we be selecting?

Good question! Given that the definition specifically included Felty, likely because patient with this syndrome likely look very similar (in some ways) to our Crohns patients (ref), we can assume the clinical relevance is always there, given the string is a good match to the source_label. Does that help? If not, we can think about adding a new label that would allow us to better capture this.

[callahantiff]

• [x] Closing this issue. According to the DropBox paper documentation, this task was completed on 07/13/19. 🎉

Please feel free to re-open if more work is needed on this task.