Towards Episode-based Payment in Ambulatory Care

The researchers explored the potential gains and losses to various types of physicians and other providers from applying episode based payment (EBP) in ambulatory care settings. Focusing on a large multi-specialty group practice, the researchers (1) characterized the resources used by type of service within episodes of care, (2) clustered episodes by physician specialty and identified the cost share accounted for by types of services in those clusters, and (3) explored how inter-specialty coordination could yield savings in the context of an episode-based payment system. The goal of this project was to advance policy discussions about episode-based payment in an ambulatory care setting, particularly in the context of developing payment models for accountable care organizations.

Policy Summary:

Fee-for-service (FFS) is often criticized for encouraging the over-provision of services, but even if a physician gets no income for additional tests or referrals, FFS offers no incentive to consider costs attributable to other providers. Episode-based payments are conceptually an attractive alternative to FFS-based payment to encourage the appropriate use of resources.
Episode-based payment shifts financial risk to the recipient of the bundled payment. While manageable for a large organization, risk-bearing is an entirely new activity for individual or small numbers of physicians, e.g., within a department. Clearly, they should not be held responsible for costs due to variations in the patient’s condition. In theory, a good risk adjustment model embedded within an episode grouper will address this. If multiple physicians are involved in an episode of care for a specific patient, however, it is also important to consider what costs should be the responsibility of each.
We sought in this project to examine the resources used by a large (~1000 MD) medical group in the outpatient care of their patients and the potential issues if episode based payment were to be applied. The first step was to categorize episodes by the number of different physician specialties involved. The second step was to modify a “standard” episode to approximate something the initial physician could feel he or she “should” be able to control.
We applied the Symmetry© grouper (Optum) to claims data for 5 years. Analyses of 202 episode treatment groups (ETGs) common in ambulatory care identified distinct patterns of physician involvement and cost responsibility. Illustrative patterns include: (1) one physician specialty responsible for the vast majority of care, e.g., primary care physicians (PCPs) and Urgent Care clinics (UCs) account for 94% of the costs of acute bronchitis episodes; (2) a lead physician who occasionally brings in a specialist, e.g., hypertension is primarily managed by PCPs (68% of total costs), with occasional input from cardiologists (16%); (3) different specialties collaborating, e.g., congestive heart failure is jointly managed by cardiologists (43% of total costs) and PCPs (25%).
Episode groupers were designed for retrospective assessment of claims data and comparison of costs. They bring together all the services associated with a specific problem using proprietary algorithms and windows of time in which a new service can be added to the episode. For chronic conditions, such as diabetes, the window is essentially a year. Acute problems typically have a 30-day window, but continued receipt of services over time can make the episode quite long.
Acute bronchitis (AB) served as an initial case study for an acute problem. Of 78,828 such episodes, 3% extended beyond 30 days, and these were substantially more expensive. At the other extreme, one-day AB episodes (accounting for 78% of the total) would likely be the most acceptable to physicians for episode-based payment because nearly all the costs were paid to, or ordered by, the initial physician. Based on 1-day AB episodes in 2007-2011, there was substantial variability across providers in costs. Some PCPs, moreover, had costs significantly below the average, yet did not have a higher proportion of multi-day episodes or higher cost for their multi-day episodes (indicators of potential under-treatment of their patients) than the other PCPs. With some PCPs repeatedly achieving low 1-day costs without evidence of subsequent problems, the application of episode-based payment may create incentives leading to improved resource use.
These preliminary findings point to ways in which the episode-based payment concept may be applied in ambulatory care. Work is continuing to refine these analyses and to add estimates of prescription drug costs to characterize more completely the episodes.