Pulling It All Together

(Post Seven in a series)

By Joanne Conroy, MD

Here we are, one week before CMMI/CMS releases the Hospital Referral Cluster data to  applicants so we can assess the “opportunities” for savings. We have had our final meeting on the definition of the episodes, and a number of our members have weighed in with detailed comments about how to define the episode of care.

We are trying to construct an intelligent definition of the episodes. We know we may get pushback from CMS to have a broader definition, but we will start by defining the episode of care in an clinically driven fashion, including what should and could reasonably happen during care. The seven episodes we have chosen are Congestive Heart Failure, Pneumonia/Chronic Obstructive Pulmonary Disease, Cardiac Valve Surgery, Coronary Artery Bypass Grafts, Spine Surgery, Total Joint Replacement, and Stroke/ Intracranial Hemorrhage.

The group moved from creating very narrow definitions (in an attempt to minimize risk) to broader inclusive episodes (in order to maximize saving opportunities).

Although many valiantly went through line by line of the readmission DRGs and the 15,000 ICD-9 diagnosis codes for post-discharge services, they realized that such a task is beyond any one person’s ability because of the sheer number of codes and that the process in detail does appear relatively arbitrary. But collectively, this process builds a strong episode.

  1. We took the “broad strokes” approach, asking everyone to give us guidelines around how they would construct the episode, with readmission inclusions, exclusions, and post-discharge services.
  2. We asked them to include services that would be responsive to management of care processes and would be a part of a high-quality discharge process to home, where support and effective aftercare would be provided.
  3. We asked them to include services that were part of care pathways that would favorably impact readmissions.
  4. We asked them to consider services that would minimize overall risk for events that put both patients and institutions at risk — for example, anoxic brain injury that led to both devastating patient impact and an incredibly expensive hospital course that could have been prevented with standardized care processes.

We then took everyone’s input — including identifying what their “hard stops” were — and put all of the episode definitions on a color-coded master spreadsheet.

This is where I now understand how the facilitator/conveners can meaningfully contribute to the application process, in addition to hiring a great consulting firm to write the application and finding data experts to analyze and interpret the firehose of data that we will receive next week. With the master spreadsheet, we were able to compare the episode definitions across all of the readmission DRGs and post-discharge services and harmonize them. There were services and conditions that 90 percent of the participants thought were important for inclusion; it was relatively easy to create episodes that included service such as infections, medication management, and treatment complications across all patient groups. The spreadsheet easily highlighted services that were specific to procedures, versus chronic diseases and services specific to cardiac, neurological, or musculoskeletal care — across all 15,000 ICD-9s , seven disease episodes, from 22 institutions, in 6 hours. The spreadsheet was a thing of beauty, representing the cumulative energy of many thoughtful people who have made the definition of the episode a very logical process.

Our colleagues at Brandeis will finish writing the code this weekend for the episode grouper and run it against the CMS data. Institutions will be eager to see the historical cost of care for their institutions and the total cost of care for their patients within the anchor DRG for the defined episode of care. Then they can begin to discuss where the quality and cost opportunities lie.

We do love the data, but the real work is still ahead. We can define the perfect episode and analyze the data down to the provider, but if we don’t identify and impel change in how we transition care, ensure care coordination post-discharge, and create and nurture partnerships through some degree of gainsharing, this will be only an academic exercise.

—Joanne Conroy, MD, is Chief Health Care Officer at the Association of American Medical Colleges. She can be reached at jconroy@aamc.org. Follow her on Twitter @joanneconroymd.