Evaluation of Clinical Risk Groups (CRGs) Episodes as an Approach to Measuring Physician Resource Use

The successful implementation of the Medicare Diagnosis Related Group (DRG)-based Inpatient Prospective Payment System (IPPS) in 1983 clearly demonstrated that bundling of all inpatient services into a single, per case payment amount, creates an effective incentive for hospitals to utilize resources efficiently. Despite the success of the bundling of services inherent in IPPS, there have been limited efforts at creating larger payment bundles that go beyond a single encounter (i.e. beyond an admission or a visit). Payment bundles for an episode of care can provide the opportunity to expand the range of services included in a payment bundle. MedPAC notes, "[B]undling Medicare payment to cover all services associated with an episode of care has the potential to improve incentives for providers to deliver the right mix of services at the right time." (MedPAC, June 2008 Report to Congress). The Medicare Improvements for Patients and Providers Act of 2008 requires CMS to establish a physician feedback program in which physicians would receive confidential information on their resource use based on episodes of care. The initial motivation for developing the CRG classification system was for risk adjustment of capitated payments. However, the development focused on a management tool for Managed Care Organizations (MCOs), since the success of a capitated payment system is dependent on MCOs being able to respond to the incentives in the system and deliver care efficiently and effectively. The classification system that resulted is not only a management tool but can also be used as a basis for risk adjusting capitated payments. Because the high utilizing population is characterized by multiple co-morbid conditions, it is extremely difficult to accurately attribute the use of individual services to a specific health care event. For example, for a patient who has congestive heart failure, diabetes, renal failure, and is hospitalized for acute heart failure, there is considerable ambiguity in identifying precisely which services are related to the heart failure hospitalization episode as opposed to the diabetes or renal failure (e.g., whether a post discharge emergency room visit for syncope related to heart failure, diabetes, or renal failure). As a result, the definition of an episode needs to be patient-centered rather than health care event centered. The development of an episode profile requires the following steps for construction: 1) create the episode, and 2) construct the expected cost for the episode. Simple to determine quality metrics, such as avoidable admissions, should also be included with the episode if possible. Using established attribution rules, we hope to determine the distribution of CRG episodes for a sample of physicians (using Tax IDs or NPI). The tasks outlined in this contract are exploratory in nature.