New Shared Savings and Losses and Assignment Methodology: The Impact to SNF Attribution

Attribution is one of the central features of ACOs: it is used to identify which beneficiaries the ACO is responsible for. Lists of attributed beneficiaries are used to create quality and cost reports to help manage the ACOs population, and subsequently are used to calculate whether an ACO achieved shared savings.

In the February 2019 release of the Shared Savings and Losses and Assignment Specifications, CMS ‘removed use of POS 31 modifier and in its place indicate[d] more generally that [they] will exclude services billed under CPT codes 99304 through 99318 when such services are furnished in a SNF.’1

This is not the first change that CMS has made in an effort to more fairly provide attribution methodology for providers who see patients in a nursing facility. For example, in the 2017 performance year and beyond, CMS excluded nursing home evaluation and management (E&M) visits (HCPCS 99304-99318) coded with place of service (POS) 31 as a qualifying claim type for beneficiary assignment. However, nursing home E&M claims coded with POS 32 still qualified for attribution to the ACO. To clarify, POS 31 is a Skilled Nursing Facility–primarily provides inpatient skilled nursing care and related services to patients who require medical, nursing, or rehabilitative services but does not provide the level of care or treatment available in a hospital. POS 32 is a Nursing Facility– primarily provides to residents skilled nursing care and related services for the rehabilitation of injured, disabled, or sick persons, or, on a regular basis, health-related care services above the level of custodial care to other than individuals with intellectual disabilities.

However, according to a study published by Milliman in April 2018, POS 31 and 32 codes are often inaccurately coded. This can occur, for example, if the SNF provides services under POS 31 and 32, and the provider’s biller doesn’t have adequate documentation to determine whether the patient was receiving services for post-acute short-term SNF stays or whether the patient was a long-term institutional nursing home resident. The study of Medicare claims found that, “32% of POS 31 E&M claims do not have a corresponding Part A SNF claim and 49% of POS 32 claims do have a corresponding Part A SNF claim—this translates to an overall miscode rate of 36%.” 2

Since most nursing home patients typically have high expenditures, a few beneficiaries can significantly impact the ACO’s PMPY expenditures. The new rule should help stabilize the fluctuations in PMPY due to inaccurate POS coding, since CMS will exclude all codes 99304 through 99318 if they are furnished in a SNF. Avoiding drastic performance year historical benchmarks and performance expenditures that are not particularly grounded in actual ACO performance, can help ACOs more accurately manage their population and achieve shared savings.
As attribution methodology continues to evolve, this is certainly a welcome change that will hopefully help ACOs that provide the right care at the right time to the right patients aren’t negatively impacted by administrative constraints.


1 https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/sharedsavingsprogram/Downloads/Shared-Savings-Losses-Assignment-Spec-V7.pdf

2 http://www.milliman.com/uploadedFiles/insight/2018/exclusion-nursing-facility-visits-unintended-consequences.pdf

Maria Nikol

About the Author

Maria Nikol
Maria Nikol

Maria Nikol is a senior business consultant with 10 years of experience in healthcare operations and healthcare strategy. Her specialization includes process improvement and the application of emerging models of healthcare delivery such as Patient-Centered Medical Homes and Accountable Care Organizations. She holds a Master of Jurisprudence in Health Law from Widener University and a Bachelor of Science in Pharmaceutical Marketing and Management from the University of the Sciences in Philadelphia.

Leave a Reply

Your email address will not be published. Required fields are marked *