OHDSI Home | Forums | Wiki | Github

Survey data discussion (CDM Working Group)

In the 9/5/2017 CDM working group meeting we discussed proposal #90 which is a method devised by Colin Orr and Catherine Kerr, among others, to better handle survey data and responses in the CDM. We will be holding a separate session of the working group for those interested to talk through this proposal. Meeting details are below.

Date: 9/19/2017
Time: 1-2pm est
Link to Webex
1-877-5659999 x881 735 36

We had a great discussion today about the survey data proposal and the recording is located here. We will meet again in two weeks to finalize the proposal for the larger workgroup meeting. Details are below for those who are interested.

Date: 10/3/2017
Time: 2-3pm est
Link to Webex
1-877-5659999 x881 735 36

Are there any minutes or results? Can’t listen to the whole thing.

Hi @Christian_Reich,

We decided to revamp the proposal a bit so that instead of adding two tables to the CDM (one for the survey information and one for for the responses) will we be adding just one optional table called SURVEY_OCCURRENCE to house pertinent information about the survey itself. The answers to the survey will then be housed in the OBSERVATION table since only minor modifications/additions are needed so that all the important response information can be captured. @ColinOrr is going to put the new proposal together and at the next small group meeting we are planning to finalize the details (field names, necessary vocab additions, etc) and then we will bring it back to the larger CDM workgroup.

Hi @clairblacketer

I listened to the recording of the discussion and found it very helpful.

I was wondering how the new proposal relates to the approach of mapping questions to LOINC codes?

e.g. Your poster:

Hi @lee_evans,

That’s a great question. Right now this proposal is focused on adding structure to the CDM to support surveys and responses but a good next step would be to create a standard set of maps for popular open source surveys like NHANES.


Less is more, definitely. And if the OBSERVATION table can hold all the features that Collin has come up with - all the better. Can it? :smile:

That would be great!

There are already some SURVEY panels in LOINC so maybe @ColinOrr would like to review them and also make a recommendation for future ETL about when/if we should map to standard LOINC survey questions versus mapping to the new standard ICHOM survey questions that will be created?

Here is an example LOINC survey panel:


@Christian_Reich it looks like it can. There may be a few fields that need to be added but I don’t anticipate a major change to the table.

Thanks @lee_evans!

My team can also help with mappings to LOINC.

Bringing up an old topic…

Where has this discussion lead? We have pain data, scales, scores, and other attributes in our EHR data. Some of the data fits nicely into the LOINC pain panels, question/answer sets, etc. I brought up the lack of conventions re: pain data in the CDM at Themis sub-group 1 meeting yesterday. It was suggested the Survey table may be the correct place for the data. Is there a roll out date? Or will the data go into the Observation table? Pain scale values (1-10) seem better suited for the Measurement table. But the other pain attributes (quality, location, alleviating factors, etc.) seem better suited for the Observation table.



We will be voting on the Survey_occurrence table at the 2/13 meeting of the CDM working group so it doesn’t have a date yet, though I agree with your assessment that pain data would be better in the Measurement or Observation table. The Survey_occurrence table will be more for information about the survey and the answers to the survey will go to the observation table.


1 Like

Hello Clair,

I have a use case for the suggested SURVEY table as I shall be obtaining a data set produced from surveys taken on patients with hearing disabilities. Any idea on when the SURVEY table and updated OBSERVATION table would be available?


Hi @Vimala_Jacob,

My goal is for the SURVEY table to be added to v6.0 of the CDM, though we still haven’t settled on a date for the release. In the meantime, I will add code and conventions that @ColinOrr2006, @Dymshyts and I have agreed upon to the development branch of the CDM github and I’ll post a link here once it is ready.


Hello Claire,

We are about to start a project which uses survey data gathered from various hearing loss questionnaires. Since you do not have a set release date yet, is it possible for my team to become beta testers of the SURVEY and accompanying tables?

At least this way, we would not need to repeat the ETL process twice. Firstly with the current SOURCE_TO_CONCEPT_MAP & OBSERVATION route. And then secondly when the SURVEY table is made available.


Hi @Vimala_Jacob,

Here is the link to the development branch of the CDM github with the SQL server code for the SURVEY table. Please feel free to use it with the understanding that since it is in dev it is not supported by the tools. I would appreciate if you could log issues for any bugs you find, it will really help when we go to officially release it.