CJR: Navigating the Data Matrix and How to Make Sense of Hospital Data to Maximize Financial Opportunity

Selected hospitals across the country began the mandatory implementation of the Comprehensive Care for Joint Replacement (CJR) model on April 1, 2016. CJR is a 5-year pilot program run by the Centers for Medicare and Medicaid Services (CMS) testing bundled payments for lower extremity joint replacement (LEJR) clinical episodes. By including hospitals from 67 Metropolitan Statistical Areas (MSAs) around the country, CMS hopes to learn more about how to incentivize valued-based care and reduce spending across diverse locations and patient populations.

 

On the spectrum of government healthcare initiatives, bundled payment models are nothing new. CMS implemented several pilot programs during the 1980s, such as the Acute Care Episode (ACE) demo, to assess the effectiveness of bundled payments for various clinical episodes in order to determine if this payment model can reduce cost and overutilization of services while maintaining and increasing quality health outcomes. Further, CMS is currently implementing the Bundled Payment for Care Improvement (BPCI) Initiative that is testing the same idea with a variety of clinical episodes, including both elective and trauma LEJR.

 

In order to profit on a bundled payment model, hospitals and physicians need to continually analyze processes and data as well as coordinate multiple systems. As convener for the largest number of orthopedic physician groups in BPCI, Signature Medical Group (SMG) has gained the experience and knowledge needed to be successful with bundled payments. It can be overwhelming to know where to start when implementing a bundled payment model, but one thing we know for sure: you need to know your data inside and out.

 

In order to succeed in CJR, it is essential for hospital administrators to understand three things about the data:

 

1)  The target price is a combination of hospital-specific data and aggregate regional hospital data, and it will change over time.

2) An experienced team can help interpret data, generate meaningful reports, and drive effective strategy and decision-making.

3) Correct analysis is essential. This data can be interpreted (and misinterpreted) in many ways.

 

1)  The target price is a combination of hospital-specific data and aggregate regional hospital data, and it will change over time.

 

Each hospital receives a set of target prices for LEJR episodes, calculated from both hospital-specific data and regional hospital data for each clinical episode. For the first two years, the target price is calculated by blending 2/3 of the hospital specific-rate and 1/3 of a regional episode rate. In year three, the target price is calculated by blending 1/3 of the hospital-specific rate and 2/3 of a regional episode rate. In years four and five, the target price is entirely based on regional episode payments. According to the CJR final ruling, the gradual transition from hospital-specific target pricing to regional target pricing facilitates competition between hospitals in the same region to provide the best care at the lowest cost.

 

Hospitals who are able to keep their average cost per LEJR episode under the assigned target price will receive cost savings in the form of a reconciliation payment, and those who are over the assigned target price will be writing a check to CMS.

 

Although a hospital’s financial gains are ultimately dependent upon the performance of surrounding hospitals (data which you do not have access to review), the focus needs to be on accurate data compilation and interpretation for the participating hospital’s own data in order to identify overutilization of care.

 

2)  An experienced team can help interpret data, generate meaningful reports, and drive effective strategy decision-making.

 

For CJR, CMS will share retrospective hospital-specific data and performance period data with participating hospitals upon request on a quarterly basis. Performance period data will include raw beneficiary-level claims data and summary beneficiary claims data. Hospitals can analyze their own data, or those without analysis capabilities can review the summaries.

 

I can tell you from experience, the raw beneficiary-level claims data is an overwhelming amount of information provided in an incomprehensible format. The ability to analyze and interpret your own data is invaluable for developing a care redesign strategy. It is essential to have IT professionals on your team who have experience working with CMS data files and data analysts who can interpret episode spending to help guide your decision-making.

 

Creating Reports with Data Sets That Contain Over One Billon Data Points

 

For BPCI, we work with many CMS raw data claims files. We receive multiple flat files containing patient information, episode codes, CPT codes, diagnosis codes, claims information, and costs. Our IT team developed a system to extract the relevant data from each file, incorporate their knowledge of BPCI rules, and import the data into a database that is used to generate reports to monitor spending and guide decision-making.

 

Downloading, compiling, and uploading information to a database seems straightforward, but in this case it is challenging.

 

First, the files from CMS are too big to download in Microsoft Excel. Instead, programming languages, such as SQL, can be used or a third party agency who has previous experience working with CMS claims data can be utilized.

 

Second, multiple files contain information on the same patient and those files are not necessarily linked with the same patient identification information. It takes a fair amount of time to figure out this data puzzle to correctly piece together all of a patient’s information.

 

Third, CMS provided over 30-pages of BPCI rules defining what services are included in the initiative and which are excluded. A similar document has been released for CJR. For example, certain psychiatric services are excluded from the total cost of a patient’s 90-day clinical episode within the BPCI initiative. Although these costs may be included in the claims data, our data analysts incorporate the defined rules to extract the data for a more accurate portrayal of a patient’s total BPCI clinical episode cost by excluding these services and costs in an analysis report. An in-depth knowledge of the rules is necessary to accurately interpret CMS claims data and to find ways to reduce overutilization.

 

Ultimately, a hospital needs to decide whether they have the capacity and time to work with the CMS raw data claim files, understand CJR rules, and develop the reports necessary to interpret performance that will guide their CJR strategy. These reports can evaluate spending by physician, diagnosis codes, service utilization rates, length of stay, readmission rates, and more. In our experience, the time needed to generate meaningful reports has been substantial, but the benefits are worthwhile.

 

3) Correct analysis is essential. This data can be interpreted (and misinterpreted) in many ways.

 

To be successful in CJR, hospitals must find ways to reduce overutilization of services while maintaining quality care. We have already established that accurate data analysis is critical for success. However, interpreting raw claims data is like taking a Rorschach inkblot test: you are not entirely sure what you are looking at and everyone sees something different. Below is an example of a graph generated from claims data that demonstrates the average cost distribution for services utilized within a 90 day LEJR clinical episode. The question is: what information can we glean from this graph?

 

Chart Data

In this scenario, the target price is $21, 274. According to the graph, the average cost per case is $22,774. A significant amount is spent in the hospital acute care setting and the skilled nursing facility (SNF). There are several interpretations a hospital may make about this graph to guide strategic spending decisions.

 

Interpretation 1: Hospital acute care services cost the most and that is where the most money can be saved. As a result, a hospital may focus on reducing patient length of stay in the acute care setting. However, hospital acute care costs are mostly fixed. Thus, efforts to save money in this category of care would not be very effective.

 

Interpretation 2: SNF costs are the second highest expense and need to be reduced. While it may be true that SNF costs need to be reduced, it may not be the first place that costs should be lowered. It can be misleading to assess cost independently of other critical factors.. Although inpatient rehabilitation facility (IRF) costs are often lower overall because less patients are admitted to IRF, IRF costs can be three timesx the daily cost of SNF. The average cost per case can be dramatically reduced if IRF utilization is reduced by even 2%. The key here is to include utilization rates as well as average facility costs in the calculations and presentation of data.

 

Interpretation 3: The average cost per case is $1,500 higher than the target price. Thus, we need to reduce each clinical episode by $1,500. Hospitals may focus on cost reduction per case rather than identifying the treatment plans or patient populations creating overutilization. Hospitals need to standardize care and protocols to provide appropriate and quality care while reducing variance and inconsistencies in care. The clinical pathway and quality care cannot be ignored since they reduce costly preventable readmissions.

 

Data Takeaways

 

Ultimately, interpreting the data requires a team that is knowledgeable about CMS claims data, bundled payment data rules, provider services, and has experience with best practices in managing bundled payments. Optimal data interpretation incorporates utilization rates, provider costs, and knowledge of clinical pathways and evidenced-based treatment protocols to effectively achieve savings.

 

As participating hospitals are approaching the half way mark for the first year of CJR, they must identify areas of overutilization and implement strategies to reduce spending. Simultaneously, the target price will go through changes as the regional rate is phased in and hospital performance is increasingly compared to the performance of other hospitals. Fortunately, hospitals will have all the resources they need if they can assemble the right team of people.

 

For more information or assistance in working with bundled payments and value-based care, visit SMGbundles.com.

Ann Conrath