A Short Efficient Update On More Than You'll Ever Want To Know About Accountable Care Organization (ACO) Patient Attribution (Courtesy of Milliman)

Learning obscure concepts and baffling terms is one reason why thousands of readers continue to return to the Disease Management Care Blog. They know they can count on the DMCB to help them bedazzle their competition with enigmatic yet precise insights that amaze and appall. Case in point is the DMCB discovery of this gem of Milliman actuarial summary paper written by Susan Pantely on the attribution of patients to Accountable Care Organizations. The link is here.

Knowing you're busy, the DMCB offers the concise summary below:

"Attribution" is necessary because ACOs will ultimately be held "accountable" for a population's health care cost and quality. Since ACO's will not have a gatekeeper function, patients will be able to seek care from any provider they choose. Data on the performance of any ACO will therefore only be meaningful if patients can be "attributed" to a provider in an ACO.

Here's a quotable definition. ACO atrribution is....

...assigning a provider or providers who will be held accountable for a member based on an analysis of that member's insurance claims data. The attributed provider is deemed to be responsible for the patient's cost and quality of care, regardless of which providers actually deliver the services.

This attribution can be based on costs, visit number or claim type.

You also need to decide on:

"Patient-based" vs. "episode based" attribution: The former assigns all study duration days (typically one year) of data for each patient to an ACO provider. The latter assigns "episodes" of care for each condition from the first to last provider visit. For example, if a patient sees a nurse practitioner for a sinus infection, gets an x-ray and then sees an ENT physician who orders the MRI, all that gets "bundled" in to an episode. Other unrelated health care costs are excluded from the attribution.

"Single" vs. "multiple" attribution: Patients' costs can be attributed completely to a single provider in one ACO or to multiple providers in one or, theoretically, several ACOs.

"Majority" or "plurality" attribution: The former happens if 50% or more can be assigned to an ACO provider; if 50% isn't achieved, there is no assignment. The latter happens to whichever provider gets the majority.

"Prospective" vs. "retrospective" attribution: Based on past utilization patterns, patients can be assigned to a provider today, who then is responsible going forward for the study duration. Retrospective assigns patients based on past utilization. The former has the advantage of being able to yield timely data. The latter has the advantage of being more accurate.

Note that it's possible to simultaneously use different approaches for different specialties within a single ACO.

Other issues to think about are 1) duration (typically one year), 2) continuous enrollment issues, (if a patient leaves a health plan after 10 months, should he or she be attributed?), 3) no claims (these patients are low cost and their attribution could make an ACO look good), 4) family attribution (members of a family tend to cluster) and 5) risk adjustment.

The author points out that the best mix of attribution methods may depend on the organization and its goals. For example, an ACO that values the involvement of multiple providers may favor combined "episodic" and "plurality" approaches in their attribution algorithms.

The DMCB thinks the bad news is that the CMS ACO regulations may be unable to accommodate the multiple methodologies described above. As a result, organizations thinking about becoming ACOs will need to determine if the narrow range of attribution methodologies proposed by CMS in their coming regulations are good organizational fits.

The good news is that there are now readers who can confidently ask business meeting or policy conference show-stopper questions like .... "That was a great talk, Mike. Do you believe prospective, risk adjusted, majority and patient-based approaches based on one year of continuous enrollment is an important option in the emerging science of ACO attribution?"

Enjoy.