The Michigan Value Collaborative

Helping Michigan hospitals achieve their best possible patient outcomes at the lowest reasonable cost

Tag: CMS (page 1 of 4)

Moneyball in Medicare

Edward Norton

Edward Norton, Ph.D., is a health economist working with MVC.

The Center for Medicare and Medicaid Services (CMS) is increasingly using financial incentives in pay-for-performance programs to encourage higher quality care at lower cost. Michigan hospitals might want to know: “How much are we penalized if one additional Medicare patient dies?”  A National Bureau of Economic Research (NBER) working paper by several members of the MVC team addresses that question for the Hospital Value-Based Purchasing Program (HVBP), with surprising results.  This study finds that about one-third of Michigan hospitals face no financial penalty if one additional patient with AMI, heart failure, or pneumonia dies.  For most other hospitals, the penalties for an additional death are modest, typically less than $10,000, but a few hospitals face penalties of up to around $40,000.

CMS created HVBP to reward or penalize hospitals based on their quality and episode-based costs of care. Within HVBP, each patient affects hospital performance on a variety of spending and quality measures (including mortality), and that performance translates directly to changes in program points and ultimately dollars.  For example, when a patient with AMI dies, the hospital’s AMI mortality rate increases, which reduces their points for the mortality measure, which reduces their total performance score, which lowers their percent bonus paid in two years, which lowers their future Medicare revenue.  But until now, the magnitude of this penalty was unknown. 

A recent NBER study — authored by myself, doctoral student Jun Li, medical student Anup Das, and MVC Associate Director Lena M. Chen — estimates how much money each Michigan hospital would lose if mortality increases by one, for each of three conditions, AMI, heart failure and pneumonia. The MVC data were essential to conduct the simulations.

One reason that the magnitude of financial incentives are hard to calculate is that there is no simple formula. Hospitals are rewarded more points if they have a low mortality rate relative to other hospitals or if they improve their mortality rate relative to their own performance in a prior year. 

The magnitude of the HVBP penalty for one additional death ranges widely across Michigan hospitals, from $0 to more than a $40,000 penalty (see Figure 1). For roughly one-third of all hospitals, there is no penalty.   The specific numbers are 17 hospitals (out of 50) for AMI, 33 out of 73 for heart failure, and 25 out of 75 for pneumonia (some hospitals do not have enough patients to meet the minimum threshold).  On the other hand, for the two-thirds of hospitals that do face a penalty, it can be as large as -$44,683 for AMI, -$41,303 for heart failure, and -$29,345 for pneumonia. 

Michigan hospitals with larger penalties tend to be larger and to be safety-net hospitals. These hospitals also tend to have mortality rates in the middle of the distribution, because those that have the best or worst rates are not penalized much by a single additional death. 

We do not know yet if these financial incentives affect behavior, although this is the premise of the HVBP Program. However, if provider behavior responds to financial incentives, these findings suggest that CMS may need to adjust how it calculates points and creates incentives for hospitals in the HVBP program.  The research by the MVC team on the national CMS pay-for-performance programs will help us design better pay-for-performance programs in Michigan, and achieve the goal of delivering high-value care to Michigan residents.

Figure 1.

This work was supported by the National Institute on Aging (P01-AG019783). Support for MVC is provided by Blue Cross Blue Shield of Michigan as part of the BCBSM Value Partnerships program; however, the opinions, beliefs and viewpoints expressed by the author do not necessarily reflect those of BCBSM or any of its employees. 


Questions or comments?  We’d love to hear from you!  You can comment on this post, or use the form at the bottom of the page to contact us.

Optimizing Post-Acute Care Spending: Care Setting or Intensity?

Scott Regenbogen

Dr. Regenbogen is a Co-Director of MVC and an Assistant Professor of Surgery at the University of Michigan.

MVC hospitals confronting the CMS Episode Payment Models, BCBSM Hospital P4P metrics, and other episode-based measures ask themselves (and us): “Where can we find opportunities to reduce episode spending?” A publication in last month’s Health Affairs authored by several members of the MVC Leadership, may help. This study suggests that hospitals can most effectively address post-discharge ancillary care costs by examining how frequently they choose to send their patients to inpatient rehabilitation care facilities after discharge.

In the past, we have found the greatest point of leverage in overall episode spending involves post-acute care services. An often cited paper from our previous MVC Director and current Associate for Strategy, David Miller, found that post-acute care was responsible for the greatest share of variation in episode payments after inpatient surgery, ranging from 31% for coronary bypass (CABG) to 85% for hip replacement (THR). However, it is less obvious how a hospital would set about addressing their patients’ post-acute care spending if they found themselves to be high utilizers.

This new study — authored by MVC Associate Director Lena Chen, Economist Edward Norton, and Co-Director Scott Regenbogen — asks a simple question: do high post-acute care expenditures result from (i) the decision to use any post-acute care; (ii) the choice of what kind of post-acute care to use; or (iii) the quantity of services used once enrolled in post-acute care.

To answer these questions, the study identified hospitals nationwide in the highest and lowest quintiles of post-acute care spending for fee-for-service Medicare beneficiaries undergoing colectomy, CABG, and THR. It then performed both price standardization and risk-adjustment, as we do in MVC data, to allow uniform comparisons of the spending that would be expected if all hospitals treated clinically similar patients under similar regional price and wage constraints. Next, it evaluated the degree to which remaining differences in utilization and spending could be explained by the choice of setting (home without services, home with visiting nursing or in-home rehabilitation, inpatient skilled nursing, inpatient rehabilitation etc.) versus the intensity of care (the quantity of services billed within the chosen modality of care). This analysis used the Oaxaca-Blinder decompensation method to identify the separate contributions of factors within and between the care settings.

The choice to use any inpatient rehabilitation explained the largest share of all variation in adjusted post-acute care spending, ranging from 26% for colectomy to 52% for CABG. The intensity of use of inpatient rehabilitation, in contrast, explained 3% or less of differences. Choice of skilled nursing was responsible for 16-25% of variation, and choice of home health accounted for another 8-13%. The only site and condition for which the intensity of care made a meaningful contribution was in skilled nursing facilities after colectomy, where the quantity of billed services contributed to 20% of variation.

These findings suggest that hospitals seeking to reduce their post-acute care costs after major inpatient surgery should first examine the proportion of their patients that enter the most expensive types of care after discharge. In general, they need worry less about the quantity of services that get used once the patients arrive there.

The research was supported by the National Institute of Aging (Grant No. P01AG019783-0751). Dr. Chen was also supported by the Agency for Healthcare Research and Quality (Grant No. K08HS020671).


Questions or comments? We’d love to hear from you!  You can comment on this article, or use the form at the bottom of the page

Older posts