Fixing grain of core member months

[leighewarner]

---

Describe the bug - Required
The member_months table is built at the person_id level instead of the member_id level. When a person enrolls in multiple plans (creating multiple member_ids), this causes duplicate person_id entries for the same month, leading to test failures due to unexpected data fanout.

---

Environment - Required

• Tuva project package version (e.g. 0.6.0): 0.14.6
• dbt version (e.g. 1.7): 1.9.4
• dbt type (e.g. dbt cloud or dbt CLI): dbt cli
• Data warehouse (e.g. Snowflake): Snowflake

---

To Reproduce
Steps to reproduce the behavior:

1. create a person_id in the eligibility input layer that has multiple member_ids.
2. Run the tuva project and specifically test the core_member_months table

---

Expected behavior
A clear and concise description of what you expected to happen.

I would expect that the core_member_month table have tests that ensure that each member_id + month is unique, rather than each person_id + month is unique.

---

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

---

Additional context
Add any other context about the problem here.

[chase-jones]

Addressed by #905

[leighewarner]

I don't think this is addressed by #905!

[sarahmcmorgan]

Hi @leighewarner, sorry to hear you're having issues. I see that you're still using version 0.14.6. I would suggest upgrading to the latest version, 0.14.13, to see if you're still getting errors.

In the latest version, there is a new intermediate model called "claims_enrollment__member_months" which creates a member_month_key using person_id, year_month, payer, plan, and data_source in a dense_rank function. This should prevent duplicates as you have described.

In your screenshot, the member ID is different, but they have the same plan. Is that happening in your data? Can a person have multiple member IDs for a given plan? I would expect the member to be given the same member ID for the same plan. That's how I have seen it in the past, but payers continue to surprise me with new data quality issues. 😀