Darcie HurteauParticipants in the Medicare Comprehensive Care for Joint Replacement (CJR) program have their work cut out for them with the challenge of responding to changes throughout the program in the makeup of target pricing, derived from a three-year moving baseline.

As year three of the CJR program closes out, participants should prepare for another critical change in the target price. In program year four, which began October 1, 2018, the target price changed from a mix of hospital and regional-specific data to 100 percent regional data.

As outlined in the exhibit directly below, the target price during program years one and two was largely based on participants improving against their own historical performance. During those years, hospitals achieved savings by reducing episode spend below their own historical levels. 

CJR Target Prices Derived from 3-Year Moving Baseline
CJR target prices derived from 3-year moving baseline

In program year three, that flipped, and the target price was largely based on participants’ performance compared with the performance of other hospitals within their region. That has meant that each participant’s original delta—the difference between its historical baseline and its regional baseline—has started to dictate by how much and how quickly it needs to improve.The exhibit directly below illustrates six possible change scenarios based on whether a hospital starts off being more or less efficient than the average efficiency of organizations in its region, and how fast its rate of change is compared with the regional rate of change. If a hospital has a target price based solely on its historical baseline data, its aim is to reduce spend by at least enough to cover the target discount and maintain that level of performance. Therefore, it pays to be inefficient in the baseline. 

Rate of Change Comparison, Hospital Versus Region
Rate of change comparison, hospital versus region

Moving to a regionally based target, hospitals must either begin with costs less than the regional baseline and maintain that relative positioning, or they must reduce costs below the regional baseline and at the same rate of change as the regional rate. This situation explains why having regional targets doesn’t work for a voluntary program: Only hospitals that already are efficient would apply.

The exhibit directly below compares four hospitals’ performance with the respective baselines over three years (represented by $0). Actual performance is shown for program year two (PY2), where the target price is largely based on hospital-specific data, and PY3, where it is based largely on regional data, and with projected performance shown for PY4, where the target is based on 100 percent regional data. The data indicate the following:

  • Hospital 22 has made significant improvements against its historical performance, but its costs are still significantly higher than the regional performance level.
  • Hospital 30 has made significant improvements against its historical performance but not enough to get below the regional performance target.
  • Hospital 35 was not able to improve performance based on its historical experience and its costs were significantly higher than the regional average.
  • Hospital 40 made improvements against its own historical performance but was already better than the regional performance.  

DRG 470 Non-Fracture Performance Variance PY2 to PY4*
DRG 470 non-fracture performance variance PY2 to PY4

With regard to the comparison laid out in the exhibit directly below, baseline 1 consists of 2012­–14 national DRG 460/470 episode experience data, whereas baseline 2 uses 2014–16 data. The Bundled Payments for Care Improvement (BPCI) initiative began in earnest in 2014, so these reductions reflect the effect of those major joint replacement of the lower extremity (MJR) participants.

Regional Average Episode Change Baseline 1 to Baseline 2 DRG 470 Non-Fracture
Regional average episode change baseline 1 to baseline 2 DRG 470 non-fracture

Looking at the reductions in regional episode spend, the data show that if the target price was 100 percent regionally based from program year one to program year two, participants in some areas would have had to reduce Medicare payments by almost 5 percent to break even. 

Baseline 3 reflects projected data for 2016–18, during which time BPCI and CJR programs were in effect, so regional averages can be expected to decrease even further in PY5.

Participants will need to determine by how much their target is changing. If a hospital had set a goal to reduce its target by a certain amount, that reduction may not be enough to meet the new target price in performance year four.

It therefore is critical for participants to reassess their program strategy in light of this change to a 100 percent regionally derived target price. Even for those who have been successful under CJR, a hospital’s performance thus far may be very different from what it will experience moving forward. 


Darcie Hurteau is senior director at DataGen, Rensselaer, N.Y.

Publication Date: Tuesday, October 02, 2018