Care patterns in Medicare and their implications for pay for performance.

Area of Science:

• Health Services Research
• Health Policy
• Medical Economics

Background:

• Pay for performance (P4P) in Medicare assumes patients can be assigned to a primary physician using claims data.
• It also assumes a significant portion of care is delivered to these assigned patients.

Purpose of the Study:

• To evaluate the feasibility of assigning patient responsibility in Medicare fee-for-service.
• To assess the impact of patient care dispersion on P4P initiatives.

Main Methods:

• Analysis of Medicare claims (2000-2002) for 1.79 million beneficiaries.
• Assignment of patients to physicians based on most frequent visits.
• Evaluation of assignment stability and percentage of care from assigned physicians.

Main Results:

• Beneficiaries saw a median of 2 primary care physicians and 5 specialists in 4 practices.
• A median of 35% of annual visits were with the assigned physician; assignments changed for 33% of beneficiaries yearly.
• Assigned patients represented a median of 39% of a primary care physician's Medicare patients and 62% of visits.

Conclusions:

• Patient care is highly dispersed among multiple physicians in Medicare fee-for-service.
• This dispersion significantly limits the effectiveness of P4P programs relying on single retrospective assignment methods.