Abstract Body (Do not enter title and authors here): Background: Accountable care organizations (ACOs) often assess provider care quality by using attribution algorithms to determine the provider most responsible for a patient’s care. However, patients with ischemic heart disease (IHD) may receive care from more than one physician, and secondary prevention for IHD is often provided by both primary care physicians (PCPs) and cardiologists.

Aim: To assess the performance of two commonly used claims-based ACO attribution methods for identifying the provider responsible for managing IHD-related outpatient care for patients in the year following incident acute myocardial infarction (AMI).

Methods: Retrospective analysis of Medicare claims data for 491,391 AMI patients who survived for at least 1-year. Assignment of responsibility was made to the physician who saw the patient for the 1) plurality of PCP visits (PCP-method) or 2) plurality of PCP or cardiologist visits (PCP-Cardiologist method). For each method we assessed a) the proportion of patients that could be attributed to a physician, and b) how often the attributed physician and non-attributed physicians provided IHD care. IHD visits were identified using a previously validated list of IHD diagnosis codes.

Results: The PCP-method attributed 352,261 patients (72%) and the PCP-Cardiologist method attributed 396,618 patients (81%) to a physician. The PCP-method attributed physician did not see the patient for an IHD visit in 38% of cases and did not provide the plurality of IHD visits in 67% of cases. The PCP-cardiologist method attributed physician did not see the patient for an IHD visit in 30% of cases and did not provide the plurality of IHD visits in 35% of cases. Overall, the proportion of IHD visits provided by other, non-attributed physicians was 69% for the PCP-method and 39% for the PCP-Cardiologist method.

Conclusions: Patients with IHD often obtain care from multiple physicians. Commonly used ACO attribution methods to identify a single physician responsible for a patient’s IHD care results in a significant number of patients being unattributed. Attribution methods that focus on visit frequency alone often attribute patients to physicians that do not provide any IHD care or less than a plurality of IHD care. Multi-attribution methods which attribute patients to more than one physician and consider both visit frequency and clinical context could result in a more appropriate appraisal and distribution of the responsibility of IHD care.