For most of the nine study conditions examined, the condition-related episodes involved a median of one PCP (meaning half of the episodes for any given condition involved only one PCP). Episodes related to AMI involved a median of two PCPs, and episodes related to breast cancer and low back pain involved a median of zero PCPs.34 For episodes that did not involve any PCPs, this may pose challenges for determining who to hold accountable and who would be responsible for coordinating care. Among those episodes involving a PCP, the PCPs could be located in ambulatory, hospital outpatient, or inpatient settings; as such, an AMI-related episode which involved two PCPs could indicate care from two hospitalists during a single inpatient stay. Because most episode types typically involved a single PCP, these PCPs could potentially provide a foundation for coordinating the care for a beneficiary, if the PCP is also managing care for other episode types the beneficiary may experience. This study did not use a cross-condition approach to examine whether there were multiple PCPs involved in managing a beneficiary's care across different episode types. Future work should explore whether there are multiple different PCPs involved in managing care across the entire set of episodes for any given Medicare beneficiary to ascertain whether a single PCP exists to coordinate care.
To the extent that co-occurring and related conditions (e.g., hypertension and hyperlipidemia) are grouped into a single broader episode construct, there may be a greater number of physicians involved in management of the patient. Our analysis examined only the number of providers involved within a single episode. To better understand the opportunities for and challenges associated with coordinating care, assigning responsibility for management, and aligning financial incentives, future analyses could look across all episodes for a beneficiary to estimate how many different providers are caring for a beneficiary.
The median number of specialists involved per episode was generally higher. The lowest median number of specialists was for diabetes-related episodes (zero), and low back pain and congestive heart failure involved a median of one specialist. The episode types related to conditions typically involving inpatient stays-AMI and hip fracture-involved the largest median number of specialists (six and five, respectively). These medians reflect specialists in both inpatient and outpatient settings, and so include anesthesiologists, radiologists, pathologists, and other hospital-based specialty care, including consultations. For episode-based performance measurement and payment approaches, the number of specialists raises a question about how many of these specialists should be held accountable for episode performance. Are all five specialists involved in the median hip fracture-related episode responsible for the performance measures available for hip fracture? Are different performance measures available for the care provided by different specialties? How would financial incentives, such as pay-for-performance be directed-to one, some or all physicians involved in the episode?
Although we observed a fairly high degree of dispersion of care during most episode types among multiple physicians, the dispersion was not as great as that observed by Pham et al. (2007) in per-capita analyses of FFS Medicare beneficiaries. The Pham study, which found multiple physicians involved in a FFS beneficiary's care within a given year (frequently there was more than one PCP caring for the beneficiary as well as multiple specialists) concluded that the dispersion of care across so many practitioners would prove challenging to assigning responsibility for all care to any single physician or group of physicians in a pay for performance context. Using narrower constructions of episodes, in contrast to examining all care received by a Medicare beneficiary within a year, could mitigate these concerns to some degree. The dispersion we observed will be an important design consideration, particularly in the attributing episodes to physicians for measurement or payment purposes. With multiple providers involved in the care delivery, several questions arise that warrant further investigation: 1) Who is accountable for the care delivered (one, some or all providers) with an episode of care and how might that vary under more narrow versus broader episode constructions? 2) What operational challenges exist related to being able to measure and assign responsibility to one or more physicians (i.e., unique physician IDs would need to exist on all Medicare claims and include the provider who rendered the service)?35