- Healthcare providers engaging in value-based contracts with shared savings opportunities earn financial bonuses or receive penalties based on the patients linked to them or their provider system. But which patient attribution model provided the greatest breadth of patients while ensuring stability?
A case study involving a health plan in Minnesota demonstrating a mature shared savings environment showed that retrospective patient attribution models that use primary care visits as the method of identifying patients offered both increased assigned patients and stability.
Providers and payers should also use two years of historical patient population data when attributing patients under value-based purchasing models to provider systems, the authors claimed in the American Journal of Managed Care study.
Healthcare providers should ensure that patient attribution in shared savings value-based contracts balance two conflicting goals. First, the model should maximize the fraction of patients linked to the provider system. Having a large number of patients attributed drives value-based care delivery for a relevant portion of the system’s patients.
Second, the model should also maximize attribution stability. Patients should be consistently assigned to the provider system to establish stable incentives for population health management. If patients are attributed to different provider systems, providers may lose the potential for shared savings for patients they were effectively managing.
However, the two patient attribution goals oftentimes conflict. A large number of assigned patients may increase the chances of patients moving to other provider systems. But strict stability rules could limit the pool of attributed patients.
Providers and payers have negotiated several patient attribution models in an attempt to balance the goals. The models are primarily retrospective, meaning patients are assigned after care delivery. But they contain different methods for attribution.
Value-based contracts assign patients based on the following factors:
• Level of attribution (provider system, clinic location, individual provider)
• Provider type (primary care, specialty, nurse practitioner)
• Encounter type (office visit, evaluation and management visit)
• Unit of measure (visits counts, allowed charges)
• Level of care concentration required (majority, plurality)
For example, the Medicare Shared Savings Program (MSSP) utilizes a primary care provider visit to assign patients to provider systems. An MSSP participant assumes accountability for a patient if the total cost of all primary care provider visits was incurred within that organization.
Using the previous year’s data, CMS retrospectively assigns patients based on the primary care provider visit model and determines if the MSSP participant met their financial goals.
Researchers explored how different patient attribution models impacted the balance between breadth and stability for a provider’s assigned patient population. They tested 16 attribution rules using the provider type, encounter type, unit of measure, and required care concentration factors.
They also explored 16 other attribution rules based on popular assignment methods. First, they included a primary care provider-only method. Then, they compared that to a hierarchical method in which the all-physician strategy is used if the primary care provider-only option is not applicable.
Finally, they examined the lookback method. The method is when a member was not initially attributed, so the model looks back to the member’s experience in the prior year for patient assignment.
The key findings from the analysis of 32 patient attribution rules were:
• The lookback method provided the greatest breadth of patients and stability
• If only a single year of data is available, the primary care provider assignment rule offered more stability, but slightly fewer patients assigned
• Conversely, the hierarchical model of attribution that assigns patients based on the all-physician method attributed more patients, but stability suffered
• All-visit patient attribution rules assigned larger fractions of patients than the primary care provider rule, but it was at the greatest cost to stability under both primary care provider-only and hierarchical models
• For patient attribution models using allowed charges rules, the lookback and hierarchical models captured the most allowed charged with 95 percent, respectively, regardless of other attribution factors used
Based on the data, researchers recommended that providers should use a lookback patient attribution model that employs primary care provider visits for assignment when multiple years of data is available.
If only one year of data is available, providers should decide whether greater stability of all-visit primary care provider rules or greater number of patients attributed under all-visit hierarchal rules is more important.
Although, researchers noted that providers may have more interest in the stability of their assigned patient population as a fraction of their total patient populations instead of a fraction of the payer’s enrolled population. Providers should focus on stability because of provider system size.
“The fraction of patients served by a system that are attributed to that system—and the fraction of revenue generated by those attributed patients—declines markedly with system size,” researchers explained. “Smaller systems see more turnover in their attributed population from year to year. Even the largest vertically integrated systems see significant ‘leakage,’ as only 47 percent to 61 percent of their attributed patients’ revenue is received by the system.”