OHDSI Home | Forums | Wiki | Github

Vocabulary for Respiratory Infection Screening

Vocabulary Users
#1

Hello,

I have a screening questionnaire in my source data that covers respiratory infections. I checked the vocabulary table and came across a LOINC concept , “Respiratory infection [Minimum data set]”. And there were 3 values for possible responses. However how do we represent the individual questions such as respiratory symptoms and travel to high prevalence of acute respiratory infections?

Regards,
Vimala

#2

@Vimala_Jacob:

Those LOINC surveys are standardized in the US. They work in a question-answer scheme, where you put the question into the observation_concept_id and the answer into value_as_concept_id. If you can use them - good. If not, you need to either tell us to add them, or cobble along with the existing Concepts.

Alternatively (and probably easier) would be to find the appropriate Concepts and just record them as conditions. They should be in SNOMED.

Not sure what that means.

#3

Hello Christian,

Thanks for your explanation about LOINC and suggestion to use SNOMED-CT. I did think of using it. However I would have to post-coordinate 2 concepts to reflect the meaning of a particular question. So I am not sure how that would be done within the OMOP model. Have multiple entries in the “Observation” table for a single question.

I missed out a key word that would have made my “travel” comment easier to understand. I should have said “Have you travelled to areas of high prevalence for Acute respiratory infections?”.

Regards,
Vimala

#4

@Vimala_Jacob:

Why do you want to pre-coordinate? Do you really need the question? Why can’t you just encode the answers as conditions?

:slight_smile: Funny idiom. I am sure you don’t mean to ask whether I have actually travelled somewhere, e.g. to a day care center, which surely often has a high prevalence for acute respiratory infections (aka all the little monsters are sporting a running nose). But you probably mean whether anybody has had to code concepts of that particular discpline. Let’s see if somebody picks it up.

#5

Hello Christian,

I am thinking of post-coordinating, not pre-coordinating. I believe that we would need to encode the question, as the answer is a yes or no response. And we would need to somehow link those responses to the respective question.

Yes you were right. I did not mean to ask you whether you had travelled somewhere with a high prevalence of respiratory infections. But rather that is one of the questions in the respiratory infection screening questionnaire.

Regards,
Vimala

#6

@Vimala_Jacob:

So, yes. As I said. It can all nicely go into the OBSERVATION table. The question (or the subject, doesn’t have to be posed as a question) goes in observation_concept_id, the answer in value_as_concept_id. Makes sense?

Aaaah. I got it. That is an example for a question you want to encode.

Do you have a prespecified questionnaire? We could add it, like we did the standard ones in the US in LOINC (usually the name of the survey is in parenthesis and the concept_class_id=‘Survey’).

t