Fix Incomplete HCAI Fields

[NickKramer87]

As a data analyst, I need the synthetic database to contain all relevant fields so that I can ensure that the software that I will write to perform the analysis will work on the real-world data.

Proposed Subtasks:

1. Identify all missing fields and determine which are necessary for analysis.
2. Add all additional necessary fields.
3. Create a separate Task for enumerating clearly the fabricated fields, truncated fields, and not-fully-mapped fields for later consideration during analysis.

Acceptance Criteria:

1. A list of all relevant fields with checkmark boxes. Those fields that are currently correctly outputted will have a checkmark.
2. Every item on the checklist is checked off.

Field Notes (running)

Remaining Critical fields with mistakes are:

• Procedure codes - #58
• Facility Identification Number - #84

Remaining fully missing fields are:

• "external causes of morbidity" and POAs for them. Not regularly used, we plan not to do

• Type of Coverage and also Plan Code Number - #68 Appears to be fairly California specific and not really contemplated under the synthea claims data.

• Preferred Language Spoken - #94

• Homelessness Indicator - #70

[TravisHaussler]

This is a bit open ended - sort of a catch-all project. However, there are only a few remaining fully blank fields so lets leave this. Once we add the last few (like external causes of morbidity) we will create a new task to track the fields that need additional work or have important caveats for analysis.

[TravisHaussler]

Remaining fully missing fields are:

• "external causes of morbidity" and POAs for them. I have not found any reference to this information within the synthea outputs or settings or documentation

• Type of Coverage and also Plan Code Number Appears to be fairly California specific and not really contemplated under the synthea claims data. covered in a separate task (#68)

• Preferred Language Spoken (possible) Synthea seems to give us some information for English or not English within the observations.csv so we can at least try to use that at first - @TravisHaussler can implement this at least for now

• Homelessness Indicator covered in a separate task (#70)

[rileeki]

@TravisHaussler and @NickKramer87 - Re: "external causes of morbidity", the analytics team does not regularly use these fields. Also, Medicare's synthetic files leave these fields blank (source pg 29). I think we're good to leave these blank and consider that done.

[TravisHaussler]

Excellent!

On Fri, Nov 3, 2023 at 12:22 PM rileeki @.***> wrote:

@TravisHaussler https://github.com/TravisHaussler and @NickKramer87 https://github.com/NickKramer87 - Re: "external causes of morbidity", the analytics team does not regularly use these fields. Also, Medicare's synthetic files leave these fields blank (source https://data.cms.gov/sites/default/files/2023-05/d51e1218-68c3-4c7c-9598-0b81f22fe903/User%20Guide%20-%20CMS%20Synthetic%20RIF%20Files%20May%202023_AM508_v2.pdf pg 29). I think we're good to leave these blank and consider that done.

— Reply to this email directly, view it on GitHub https://github.com/orchid-initiative/synthetic-database-project/issues/63#issuecomment-1792978938, or unsubscribe https://github.com/notifications/unsubscribe-auth/AZL3CWJOMQYJL4MDBOWI6KTYCU75VAVCNFSM6AAAAAA6H6452OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTOOJSHE3TQOJTHA . You are receiving this because you were mentioned.Message ID: @.*** com>

[beckyphan]

Per thread above, remaining issues are resolved and/or captured in new issues.