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

The Department of Health and Human Services, Centers for Medicare & Medicaid Services (CMS), formerly known as the Health Care Financing Administration, Acquisition and Grants Group, 7500 Security Blvd. C2-21-15, Central Building, Baltimore, MD, 21244-1850 intends to negotiate a sole source contract with 3M Health Information Systems for the services outlined below: 1) Develop CRG episodes: Using 3M CRG software, develop an episode model. Consider varying anchor record (such as IP admission), including PAC, length of episode (such as 60 days prior to IP, 90 days after), application to chronic and acute, etc. a. Specify the episode trigger event (e.g., hospitalization for coronary bypass surgery for patients with a low severity of illness). b. Specify the episode window or clean period rules (e.g., 30 days prior to hospitalization and 60 days post hospitalization). c. Define the episode service scope (e.g., all health care costs except the cost of acute long term care and skilled nursing facilities). 2) Construct the expected cost profiles for the CRG episode: a. Using claims data, identify the episode cohort (i.e., all patients with the trigger event). b. For each patient in the episode cohort use the patient's diagnosis and procedure profile in the year prior to the initiation of the episode to assign the patient to a CRG episode. c. For each patient in the episode cohort, accumulate all costs within the episode scope that occur during the episode window. d. For each CRG episode, establish, establish a statistical outlier threshold or stop loss based on the distribution of costs for patients in the episode cohort assigned to that CRG episode. e. For each CRG episode establish the expect cost by computing the average cost of patients in the episode cohort excluding outlier patients. 3) Develop Quality Metrics: Using CRGs, MS-DRGs, and APR-DRGs, incorporate simple quality metrics with the episode, such as avoidable admissions. 4) Develop CRG episode distribution: Using CRGs and APR-DRGs develop distributional statistics of the constructed episodes. 5) Analyze data: Using findings from above, analyze data and draw conclusions, and 6) Organize technical panel to review and comment on results: Prepare materials for panel discussion. The Government anticipates a period of performance for eighteen (18) months from date of award. This announcement is for informational purposes only.