4 Validation

We can validate using the following methods:

1. 🚧 Re-Pricing: Re-price claims using the new grouper. Prices should not be significantly different from old groupers, especially directionally between low vs high severity groups.
2. ✅ CMS: Use CMS Severity as a proxy source of truth.

Re-Pricing:

🚧 We can't do this yet

CMS:

ssp_setting	pct_MCC_in_tier1	pct_No_CC_in_max_tier
IP	14	2.9
OP	44.8	15.1

• Inpatient:
  • 13% of MCC diagnosis codes are in tier 1 (lowest severity tier)
  • 2.9% of "No CC" diagnosis codes are in the max severity tier
• Outpatient:
  • 44.8% of MCC diagnosis codes are in tier 1 (lowest severity tier)
    • Note that MCC diagnosis codes are rare in Outpatient (2% of OP claims)
  • 15.1% of "No CC" diagnosis codes are in the max severity tier

To-Do:

• Investigate the MCC diagnosis codes that are in tier 1. Are they truly low severity? Are there certain SSPs where this is more common? Are these Diagnosis Codes common for those SSPs? Which of the 14 factors are driving down the severity for these codes? Do the claims look correct?
• Investigate Outpatient in general:
  • Why are there so many MCC codes in tier 1?
    • Look at some of these outpatient claims with MCC diagnoses. I wonder if they are due to things like cardiac rehab.