Should we still be relying on the OMOP_usefule_columns.csv to identify which columns will be available in the UW data? The old one had been labeled deprecated for a while and I see there's a new CSV uploaded but it still does not seem to correlate with OMOP standard columns.
For example, under the person table, it lists time_of_birth as a useful column, however on OMOP and in the synthetic data, time_of_birth doesn't exist at all. Instead birth_datetime is a field, but based on OMOP CDM it's not required.
Should we just presume that required fields within the CDM specifications are what are useful? Or will there be a new useful columns CSV that we can use? (and can we import and use this as part of our training model)
Created by rchen Hi @rchen,
I can create a new useful columns file if that would be helpful. In the meantime, the [new synthetic data](https://www.synapse.org/#!Synapse:syn20685954) should roughly reflect the completeness of the UW Challenge data. If a column in the synthetic data is mostly null, you can assume that the UW data is also mostly null for that column. If you stick with the standard columns like condition_concept_id, observation_concept_id those will all be useful. My main motivation behind the useful_columns file was to keep people from using the souce_value columns since those aren't standard across OMOP repositories and to keep them away from the mostly null columns.
I'll work on creating a new useful_column file and release that in the next week or two.
Hope this helped,
Tim