OHDSI Home | Forums | Wiki | Github

Federal Information Processing System (FIPS) codes?

Vocabulary Users
#1

Hi there,

Is there or will there be support for Federal Information Processing System (FIPS) codes?
https://transition.fcc.gov/oet/info/maps/census/fips/fips.txt

I don’t see it listed searching through Athena.
I’m specifically interested in deriving the FIPS county codes using ZipCode and state and storing them in LOCATION.county as a means of de-identifying patient address data, and for eventually creating analyses and visualizations by US county.

Thanks!
May Terry

#2

@mlterryMitre

Counties are there. They are four Middlesex Counties in the FIPS, which correspond to Middlesex County in MA (where I live), Middlesex County in NJ, Middlesex County in CT and Middlesex County in VA, all in the OSM vocabulary. I don’t know how FIPS works, but I am sure there is a link to the states as well. So we could create a mapping.

Not sure where the Zip code comes into play. Do you want it as a separate geographical region, or just as a way to unambiguously map? In case of the former: This is very tricky. Zip codes are not geographical regions (polygons), but a number of streets the postal service trucks drive up and down. They can cross county and even State borders. One could create some kind of awkward Zip inside the county maps with exceptions, though. We were thinking about that before,

If the latter - don’t think this is a big deal. There are 1,980 of them. Should take somebody a day to map them just using State information. Want to have somebody volunteer? :slight_smile:

#3

@mlterryMitre

Hello!

Because of the oddities in geospatial data, (some of which @Christian_Reich has pointed out), the GIS work-group is currently developing an add-on to OMOP that would consist of a separate database with GIS capabilities that works side-by-side with the OMOP database. This add-on would facilitate associating geospatial attributes to patients and roll up patient level attributes to geospatial attributes.

De-identifying patient address data, calculating FIPS county codes, and creating analyses / visualizations by geographic area are exactly some of the core use cases that this group has considered. We would love to hear more about your use case and assist in whatever ways we can to support it. The next work-group call is this Wednesday (March 31) at 2:00PM ET (meeting link). I invite you to join us there, reach out to me directly, (or continue the conversation here) to talk more about this!

#4

Hi Christian,

Yeh to your point, these are not FIPS codes. OSM codes (https://osmcode.org/) are a different standard and probably too granular for our needs in which we need to translate zip codes to some region level.
Per FIPS (https://transition.fcc.gov/oet/info/maps/census/fips/fips.txt), for Middlesex County, NJ, the FIPS code should be 34023 which is a pre-coordinated code that contains the state code “34” and the county code for that state “023”.
Where would I find this code for Middlesex County, NJ within Athena if available?

I guess in your other suggestion, the request to bring in code mappings from zip to FIPS county/state can potentially be done and there is a cross-walk mapping (potentially outdated since the last release was in 2018). If I get instructions on how to to incorporate that mapping into Athena, then I suppose I can try. Just let me know.

Thanks!
May

#5

Correct. They are what we use as a standard for geographic information. So, your FIPS codes would need a mapping to the equivalent OSM concepts.

They have all granularities: town, county, state, region, country.

The FIPS codes are not in Athena, yet. The OSM concepts are (Middlesex County, NJ).

Yes, please. :slight_smile: Except, we would need the zip codes mapped to the standard OSM concepts. But same idea.

Athena is just a browser. The vocabularies are maintained in Pallas. Don’t worry about it. A spreadsheet with four columns (zip code, FIPS code, FIPS name, OSM concept ID) is totally sufficient, and we bring it in.

#6

Thanks for the clarification re: Athena and Pallas. As you can tell, I’m still pretty new to the OHDSI’s tools and process. :slightly_smiling_face: Let me take a pass at the concept mappings with the 4 columns noted and send them to you. Probably early next week.

t