Fig. 1 Line graph showing average expenditures per episode by period and condition.Data are from the authors' analysis of Medicare claims data. The analysis is from April 1, 2011, to June 30, 2012 (baseline period), and from October 1, 2013, to December 31, 2014 (risk-bearing period). The prebaseline period (July 2009 through March 2011) is included to show the general trend before the start of the risk period. Averages are adjusted to reflect changes in the proportion of patients with complications. Expenditures are adjusted to 2014 dollars based on a gross domestic product deflator. Reproduced from Jubelt LE, Goldfeld KS, Chung WY, et al: Early lessons on bundled payment at an academic medical center. 2017;25(9):654-663.


Published 11/1/2017
Terry Stanton

Bundled Payments: Early Lessons from NYU

Study in JAAOS shows savings in TJA, higher costs in spine—with caveats

The payment model for healthcare services is shifting, spurred in no small part by the federal government. In 2015, the U.S. Department of Health and Human Services set a goal of providing 30 percent of Medicare payments through an alternative payment model by the end of 2016. That goal was met by March 2016.

At the spearhead of the push toward alternative models is bundled payment, which sets a fixed price for all services during a single episode of care.

An early adopter, New York University Langone Medical Center (NYU) began participating in the Bundled Payments for Care Improvement (BPCI) model, which was initiated by the Center for Medicare & Medicaid Innovation (CMMI) in 2013. In a study published in the September 2017 issue of the Journal of the AAOS (JAAOS), an author team of NYU faculty details the institution's experience with bundled payments and shares lessons learned.

Cost savings were variable
The primary study finding was that savings, where they were achieved, varied considerably. The greatest cost reduction was seen for lower extremity joint arthroplasty episodes. Here, the adjusted average episode cost decreased by $3,017. Cardiac procedures displayed a similar average dollar reduction, of $2,999, but the decrease was not statistically significant. For spinal fusion bundles, on the other hand, episode costs went up markedly, by $8,291.

Where savings were achieved, they were largely attributable to a shift of postdischarge care from inpatient rehabilitation facilities to home. The spike in spine fusion costs was related to changes in surgical technique—increased use of interbody fusion, for example—that were not accommodated by the BPCI.

NYU engaged in bundled payments under the BPCI Model 2, in which it took on financial risk for three procedures— lower extremity joint arthroplasty, spinal fusion except cervical, and cardiac valve procedures—with episode lengths of 90 days. NYU selected the procedures based on three criteria: sufficiently high volume, opportunity to reduce postacute spending, and opportunity to reduce readmission rates.

To evaluate the effect of the interventions at NYU, the authors used a quasi-experimental, pre/post study design with a control group (difference-in-difference design) to determine the changes in the average 90-day cost per episode between the baseline and risk periods. The unit of observation was a clinical episode. The study covered Medicare fee-for-service patients hospitalized from April 2011 to June 2012 and October 2013 to December 2014 for lower extremity joint arthroplasty, cardiac valve procedures, or spine surgery (exposure group). Patients hospitalized during the same time periods for congestive heart failure, major bowel procedures, medical peripheral vascular disorders, medical noninfectious orthopaedic care, or stroke served as the control group. The investigators examined total episode costs and costs by service category.

In the pre-BPCI baseline period for the study, the exposure group had 1,352 episodes and the control group had 1,558 episodes. During the intervention period, the exposure group had 1,588 episodes and the control group had 805. Thus, overall there were 2,940 intervention episodes and 1,474 control episodes. As noted, lower extremity joint arthroplasty episodes achieved the greatest savings; the average decrease of $3,017 with 95 percent confidence interval [CI] reflected a range of net change of −$6,066 to $31. For cardiac procedures, the adjusted average episode cost decrease of $2,999 with 95 percent CI had a range of −$8,103 to $2,105; for spinal fusion, the cost increase of $8,291 with 95 percent CI had a range of $2,879 to $13,703 (Table 1). Average expenditure per episode is shown in Fig. 1.

J Am Acad Orthop Surg

The authors estimated the primary outcomes—the difference in differences for the three at-risk episodes—by assessing the interaction of time and treatment group. Their model controlled for age, sex, and the presence of a major complication or comorbidity (as determined by a "major complication or comorbidity" principal Diagnosis Related Group [DRG]).

Across all service categories, postacute facility-based spending was the single category in which savings were achieved from a Medicare perspective. Savings in this category were realized for cardiac valve replacement and major lower extremity joint arthroplasty.

Interpreting the findings, the authors write: "This study shows that the change in postacute spending was driven predominantly by reducing discharges to inpatient rehabilitation facilities, and not by reducing discharges to skilled nursing facilities or by shortening post-acute facility length of stay. Among patients discharged to an inpatient rehabilitation facility, the proportion of those with a major complication or comorbidity increased between the baseline and intervention periods. This finding suggests that the substantial decrease in the number of patients discharged to inpatient rehabilitation was largely among less severely ill patients, increasing the average severity of illness among the patients who were still discharged to inpatient rehabilitation in the risk-bearing period. At the same time, home health expenditures increased across conditions, suggesting that NYU shifted some postacute care to home health services."

The authors write that the inability of NYU to reduce Medicare costs for spinal fusion "highlights the fact that the CMMI bundled payment initiative does not account for changes and innovations in medical care." They explain that "Medical innovations may increase the costs of care in absolute terms, but if they may improve patient outcomes, they may in fact be cost effective. As currently structured, the BPCI initiative negatively incentivizes providers from implementing such innovations because, if the DRG payment remains constant, the provider would bear the difference in cost of the new innovation compared with the cost of the standard care. To remain effective, the BPCI initiative will need to monitor for and adapt to innovations in medical care."

The authors conclude: "If a newly developed technology is costly and requires several years for the improvement in outcomes to become evident, a bundled care provider may be discouraged from adopting the innovation. Despite this potential unintended consequence, our findings suggest that bundled payment may be an effective way to lower healthcare costs across an episode of care without affecting quality, particularly for lower extremity joint arthroplasty."

To learn more about the study, AAOS Now conducted a Q&A with senior author, Leora I. Horwitz, MD, MHS, and three of her coauthors—John A. Bendo, MD; Joseph A. Bosco, MD; and Richard Iorio, MD.

AAOS Now: What prompted you to undertake this study?

Dr. Horwitz: Bundled payments are a major component of the new movement toward alternative payment models to prioritize value over volume. Our institution was an early adopter of bundled payments for several conditions and made major efforts to succeed in the new environment. We were curious to see, in a rigorous fashion, whether these efforts had borne fruit, and, particularly, to explore whether results were consistent across conditions and where the cost savings were achieved.

AAOS Now: Are there novel aspects to the study design?

Dr. Horwitz: We take a very rigorous difference-in-difference approach to the question, rather than relying on simple pre/post analyses, as other early studies have done. This is crucial because many other changes were happening at the same time in the greater healthcare environment, potentially creating secular trends toward lower utilization and lower readmission rates. It is critical to analyze results in the context of ongoing changes so that we can have more confidence that the changes we are observing in the bundled payment group were associated with the program. Our difference-in-difference design permits us to control for secular trends. Moreover, we examine a longer intervention period and include more conditions than prior studies.

AAOS Now: What were the most noteworthy findings? Were there any surprises?

Dr. Horwitz: We went into the program expecting to achieve savings in all three conditions and were surprised to learn along the way that the very different clinical contexts of joint, spine, and cardiac surgeries made our efforts variably effective. In both the joint and cardiac group, where we had a high proportion of patients being discharged to inpatient rehabilitation, we had a lot of opportunity for improvement and were in fact able to reduce expenditures on postacute care by even more in the cardiac group than the joint group. However, because the index admission for cardiac surgery is so expensive—and because we obviously could not change the cost of the cardiac surgery DRG payment—the reduction in postacute care use for the cardiac group was a smaller fraction of overall costs than in the joint group and was not enough to achieve overall savings. In the spine group, we had less preexisting opportunity in postdischarge acute care and experienced an unanticipated change in surgical approach that increased costs.

Dr. Bosco: Success with the lumbar spine fusion bundle was impossible because the standard of care (more fusion levels) had changed since the baseline period and the bundle covered surgeries, which were too heterogenous.

Dr. Iorio: Bundled payments have increased quality, decreased cost, and improved the patient experience for total joint replacement. The benefits of improved quality and efficiency extend far beyond Medicare beneficiaries and have resulted in a much better experience at NYU Langone Medical Center for all patients, as well as better financial and quality outcomes.

AAOS Now: Can you expand on your finding that the BPCI yields negative incentives for innovation?

Dr. Horwitz: This is an important point for any payment structure. If incentives are geared toward short-term outcomes or healthcare-specific outcomes, then there will be less focus on or even worsening of longer-term outcomes or non–healthcare-related outcomes if they are not aligned with incentivized outcomes.

We often assume that interventions that improve short-term outcomes also improve long-term outcomes, but that is not always the case. In this case, our institution shifted toward providing a new, more complicated, more expensive, and potentially longer-recovery approach for spine patients that has been shown to result in better long-term outcomes. Although the long-term benefits may outweigh the short-term costs from the patient perspective, a payment structure system focused only on short-term costs will devalue such an intervention.

The same is true, for example, for preventive care (such as vaccinations, nutrition counseling, diabetes prevention, behavioral health, supportive housing, etc.) for which short-term costs are high and long-term health gains are only realized years or decades in the future, or which produce non–healthcare-related benefits such as higher school attendance or reduced incarceration rates that are not captured as outcomes.

AAOS Now: What are the lessons that hospitals and care providers can take from your study? What changes in the BPCI are warranted?

Dr. Horwitz: The main lessons here are that all conditions are not equally viable when it comes to bundled payment, but that success can be achieved through careful attention to main drivers of cost and variability. The BPCI program should be flexible enough to adapt to innovations in practice; incentivize evidence-based practice and high-quality care, not just cost reduction; and consider longer-term outcomes when appropriate.

Dr. Bendo: Although the bundled payment model for lumbar spine fusion did not prove to be financially advantageous, we were able to streamline and standardize our treatment protocol, thus serving to ultimately improve clinical outcomes and patient satisfaction. The inherent challenges in the spine bundle were the broad DRG (including 1-7 level posterior lumbar fusions), diversity of surgical techniques and instrumentation, and maintaining a similar case mix over time relative to the negotiated baseline.

Dr. Bosco: This topic is especially timely since the U.S. Centers for Medicare & Medicaid Services (CMS) has just proposed major changes to the Comprehensive Care for Joint Replacement Model (CJR) and the other mandatory bundles. They have made participation in the Surgical Hip and Femur Fracture Treatment (SHFFT) model and the cardiac bundles completely voluntary. Additionally, they have eliminated 33 of the 67 Metropolitan Statistical Areas (MSAs) for CJR, and in the remaining 34 mandatory MSAs have exempted rural and low-volume hospitals from mandatory participation. However, this is not a defeat or a retreat from bundled payments. The fact that CMS kept the 34 largest urban MSAs in the mandatory program is telling.

Additional authors of "Early Lessons on Bundled Payment at an Academic Medical Center" (J Am Acad Orthop Surg 2017;25(9):654-663) are Lindsay E. Jubelt, MD, MS; Keith S. Goldfeld, DrPH, MS, MPA; Wei-Yi Chung, MS, ASN; Thomas J. Errico, MD; Anthony K. Frempong-Boadu, MD; and James. D. Slover, MD, MS.

Terry Stanton is the senior science writer for AAOS Now. He can be reached at