Patient attribution: why the method matters

Abstract

Objectives: To assess the impact of 5 commonly used patient attribution methods on measured healthcare cost, quality, and utilization metrics within an integrated healthcare delivery system.

Study design: Cross-sectional analysis of administrative data of all patients attributed (by any of 5 methods) and/or paneled to a primary care provider (PCP) at Mayo Clinic Rochester (MCR) in 2011.

Methods: We retrospectively applied 5 attribution methods to MCR administrative data from January 1, 2010, to December 31, 2011. MCR is an integrated healthcare delivery system serving primary care and referral populations. The referral practice is geographically colocated but otherwise distinct from 6 primary care practice sites that include pediatric, internal medicine, and family medicine groups. Patients attributed by each method were compared on their concordance with PCP empanelment, quality measures, healthcare utilization, and total costs of care.

Results: The 5 methods attributed between 61,813 (42%) and 106,152 (72%) of paneled patients to a PCP at MCR, although not necessarily to the paneled PCP. There was marked variation in care utilization and total costs of care, but not quality measures, among patients attributed by the different methods and between those paneled versus not paneled. Patients with more primary care visits were more likely to be attributed by all methods.

Conclusions: Reliable identification of the physician-patient relationship is necessary for accurate evaluation of healthcare processes, efficiencies, and outcomes. Optimization and standardization of attribution methods are therefore essential as health systems, payers, and policy makers seek to evaluate and improve the value of delivered care.

FIGURE 1.. Number of Patients Attributed and Proportion of Attribution Match With Panel Assignment, by Attribution Method^a

ACO indicates accountable care organization; HP, HealthPartners; MNCM, Minnesota Community Measurement; PCP, primary care provider. ^aUnattributed percentages for each method are calculated from the number of paneled patients in 2011 (N = 146,469). ^bDartmouth attribution method used by CMS ACO attribution and Medicare Shared Savings Program. ^cAttribution method used for performance-based reimbursement and incentive contracts by a health plan providing Medicaid coverage in a Midwestern state. ^dAttribution method used for provider profiling by a Midwestern health plan. ^eAttribution method provided by HP with National Quality Forum total cost of care submission.

Proportion of Attribution Capture by Method for Categories of Overall Total Charges in 2011.

ACO indicates accountable care organization; HP, HealthPartners; MNCM, Minnesota Community Measurement. ^aDartmouth attribution method used by CMS ACO attribution and Medicare Shared Savings Program. ^bAttribution method used for performance-based reimbursement and incentive contracts by a health plan providing Medicaid coverage in a Midwestern state. ^cAttribution method used for provider profiling by a Midwestern health plan. ^dAttribution method provided by HP with National Quality Forum total cost of care submission.