Claims conditions mapping differs between ICD10 and ICD9 inputs

[davidshimamoto]

---

Describe the bug - Required
The handling of ICD10 and ICD9 codes from medical claim is inconsistent. This applies to both Diagnoses and Procedure codes. In both cases, in the "staging core" model, if the code is ICD10, it will be joined over to the reference table and populate the normalized fields. In the "core" output, this is treated as a "manual" mapping method. Meanwhile, if the code is ICD9, the normalized field is left blank (in staging), but then in the "core" output, this is treated as "automatic" mapping.

---

Environment - Required

• Tuva project package version (e.g. 0.6.0): 0.15.0
• dbt version (e.g. 1.7):
• dbt type (e.g. dbt cloud or dbt CLI): CLI
• Data warehouse (e.g. Snowflake): DuckDB

---

To Reproduce
Steps to reproduce the behavior:

1. Have a medical claim set that has ICD9 and ICD10 codes.
2. Run the project to populate core.condition
3. Check the values in mapping_method

---

Expected behavior
These two code systems are treated the same when they exist with valid values. In both cases, I believe mapping_method should be automatic.

---

Screenshots
If applicable, add screenshots to help explain your problem.

---

Additional context
Add any other context about the problem here.