Last week I wrote up my newcomer’s understanding of ACO quality reporting options and teased a new little simulation project. Well, here goes: an open source simulator for understanding how reporting choices flow through to shared savings: https://joshuamandel.com/app-plus-quality-lab/

I’ve attached a ~20 minute walkthrough if you’re interested in a guided tour:

The gist is: take an ACO participating in CMS’s “Enhanced Track” with 24,000 assigned lives and a $330M benchmark. Beating the benchmark by 3% savse CMS saves ~$10M. The ACO is eligible to receive shared savings of 75% (a $7.4M check), provided they pass the quality standard.

Whether they an “eCQM” or “CQM” reporting paths has a significant impact on the actual scores (with a whole lot of nuance we explore in simulation — see video). The ACO can pass by reporting all five clinical measures on your full (all-payer) population, in which case they’re “deemed” to meet the standard at full rate, almost regardless of their actual scores. Or the ACO can report on assigned Medicare patients only (which precludes deeming), and then their composite score (8 measures, 0-10 points each) has to clear ~74%. Falling short scales the payout down.

Some things I learned building this:

• Medicare-population reporting uses flat benchmarks (if the CY 2027 proposal is finalized), so every 10 points of blood pressure control earns a point. eCQMs grade on a historical curve, so 66% control rate might earn 3 points under the all-patient ladder instead of 6 under the Medicare-patient ladder.
• eCQMs are only permitted see what’s structured in the EHR and slotted into predefined fields. Good care documented in a note (or an obscure data element) doesn’t count.
• Quality matters differently in bad financial years. Overspend by 3% and a strong score caps your shared-loss rate at 40%; a weak score pushes it toward 75%. Deeming doesn’t help here, only points do.
• The “Complex Organization Adjustment” adds a point per full-population eCQM measure — a special incentive CMS only applies to full-population eCQM reporting

Under the hood

Under the hood, eCQMs are based on an XML standard called QRDA, though CMS is pushing toward a “dQM” future leveraging FHIR Bulk Export of US Core data + CQL over FHIR “QI Core” profiles.

But CMS isn’t really providing a pathway to incentivize (or explore / debug) early adoption (which is too bad because that is probably the best way to discover problems with this approach and fix them before they hit everyone).

If this all seems a bit complicated — well yes, for sure! That’s why I wanted a systematic, visual way to think about it.

Your feedback and (especially) corrections are most welcome.