Many people predicted that the mandatory “bundling” of payments for knee and hip arthroplasty by the Centers for Medicare and Medicaid Services (CMS) that began on April 1, 2016 in several US metropolitan areas would lead to “cherry-picking” and ”lemon-dropping.” In other words, hospitals and surgeons wouldn’t take on more complex and sicker patients for joint replacement for fear that the bundled payment would be insufficient (lemon-dropping), and would instead select the healthier patients (cherry-picking). See related OrthoBuzz post.
In the February 19, 2020 issue of The Journal, Humbyrd et al. compare the characteristics of patients who underwent hip and knee replacement (HKR) from April to December 2015 with those of HKR patients during the same period in 2016, after CMS mandated the bundled-payment program in 67 metropolitan statistical areas (MSAs). The patients were matched so that those treated in bundled and non-bundled settings had similar socioeconomic backgrounds.
The matched groups included 12,388 HKR episodes in 40 bundled MSAs and 20,288 HKRs in 115 nonbundled MSAs. The authors also evaluated pre- and post-policy case-mix changes among 1,549 hip hemiarthroplasties, which are not subject to bundling, in the bundled MSAs.
Among patients who underwent HKR, Humbyrd et al. found no significant differences in patient characteristics—including race, dual Medicare-Medicaid eligibility, tobacco use, obesity, diabetes, and Charlson Comorbidity Index (CCI)—after the bundled-payment policy was implemented. Also, they found that patients in bundled MSAs undergoing hemiarthroplasty had significantly higher CCI values and were more likely to have diabetes than those who underwent HKR. This suggests that some surgeons opt for hemiarthroplasty over total hip replacement in less-healthy patients to avoid treating such patients under a bundled program.
From the MSA perspective, these results suggest that cherry picking and lemon dropping are not occurring in the short term. But we would do well as a profession to ensure that those controversial patient-selection practices are not happening at the individual surgeon level, and that the short-term results demonstrated here by Humbyrd et al. persist over the longer term. Even our sickest joint replacement patients deserve the best surgical care.
Matthew R. Schmitz, MD
JBJS Deputy Editor for Social Media
When Medicare’s Comprehensive Care for Joint Replacement (CJR) program was implemented in 2016, the health care community—especially orthopaedic surgeons— had 2 major concerns. First, would the program actually demonstrate the ability to decrease the costs of total joint replacements while maintaining the same, or improved, outcomes? Second, would CJR promote the unintended consequence of participating hospitals and surgeons ”cherry picking” lower-risk patients and steering clear of higher-risk (and presumably higher cost) patients? Both of these questions were at the heart of the study by Barnett et al. in a recent issue of the New England Journal of Medicine.
The authors evaluated hip and knee replacements at 75 metropolitan centers that were mandated to participate in the CJR program and compared the costs, complication rates, and patient demographics to similar procedures at 121 control centers that did not participate in CJR. The authors found significantly greater decreases in institutional spending per joint-replacement episode in institutions participating in the CJR compared to those that did not. Most of these savings appeared to come from CJR-participating institutions sending fewer patients to post-acute care facilities after surgery. Furthermore, the authors did not find differences between centers participating in the CJR and control centers in terms of composite complication rate or the percentage of procedures that were performed on high-risk patients.
While this 2-year evaluation does not provide the level of detail necessary to make far-reaching conclusions, it does address two of the biggest concerns related to CJR implementation from a health-systems perspective. There may be individual CJR-participating centers that are not saving Medicare money or that are cherry picking lower-risk patients, but overall the program appears to be doing what it set out to do—successfully motivating participating hospitals and healthcare facilities to look critically at what they can do to decrease the costs of a joint-replacement episode while simultaneously maintaining a high level of patient care. The Trump administration shifted CJR to a partly voluntary model in March 2018, and I hope policymakers consider these findings if further changes to the CJR model are planned.
Chad A. Krueger, MD
JBJS Deputy Editor for Social Media
The bundled-payment model has found some early success within the field of orthopaedic surgery, most notably in joint replacement (see related OrthoBuzz post), However, more robust risk-adjustment methods are needed, especially in terms of patient factors. That is the message delivered by Cairns et al. in their retrospective analysis of Medicare data from 2008 to 2012 published in the February 21, 2018 edition of JBJS. The authors make a compelling case for improved risk stratification of hip- and femur-fracture patients to ensure that all patient populations have and retain access to appropriate care.
The authors analyzed reimbursements for the surgical hospitalization and 90 days of post-discharge care among nearly 28,000 patients who met inclusion criteria for the Surgical Hip and Femur Fracture Treatment (SHFFT) model proposed by the Centers for Medicare and Medicaid Services (CMS). Their findings highlighted various inconsistencies that could have unintended consequences if not accounted for in the bundled-payment model. For example, reimbursements were $1000 to $2000 lower for patients in their 80s, who tend to have more comorbidities that require more care, than for younger patients. CMS proposed using Diagnosis Related Groups (DRGs) and geographic location to adjust for risk in its SHFFT bundled-payment model, but Cairns et al. identify several other factors (such as patient age and gender, ASA and Charlson Comorbidity Index scores, and procedure type) that could provide a more realistic stratification of risk.
The article clearly articulates how risk adjustments that don’t include more specific patient factors could lead to a multitude of unintended consequences for patients, providers, and the entire healthcare system. These findings could remain relevant now that CMS has announced an “advanced” voluntary bundled-payment model after the Trump administration cancelled SHFFT in late 2017.
Whatever bundled-payment model takes hold, the totality of the orthopaedic literature strongly suggests that the best outcomes are derived from making specific treatment plans for each patient based on the individual characteristics of his or her case. It seems reasonable that the best bundled-payment plans would do the same.
Chad A. Krueger, MD
JBJS Deputy Editor for Social Media
OrthoBuzz occasionally receives posts from guest bloggers. This guest post comes from Richard Yoon, MD.
In a recent issue of JAMA, Dummit et al. analyzed cost and quality results from the Centers for Medicare & Medicaid Services (CMS) Bundled Payment for Care Improvement (BPCI) initiative. The authors compared joint-replacement results between hospitals that voluntarily participate in the BPCI program versus matched comparison hospitals that do not participate. Nearly 60,000 lower extremity joint replacement procedures from each hospital type were included in the analysis.
Medicare payments declined over time in both groups of hospitals, but the authors noted a greater decline in costs among the BPCI hospitals, primarily due to reduced utilization of post-institutional acute care. There were no statistical differences in quality between the BPCI hospitals and comparison hospitals, as measured by unplanned admissions, emergency department visits, and mortality at both 30 and 90 days. These results echo those reported by other pilots in the United States and suggest that similar programs could reduce cost per episode of care without compromising quality.
However, even proponents of the new programs are cautious. For example, in his JAMA editorial, Elliot Fisher, MD warns readers that because BPCI is a voluntary program, the results may not reflect the true impact of a more widespread bundled-payment model. The incentives, he argues, could end up contributing to volume increases or shifts toward healthier—and “more profitable”—patients. As Fisher concludes, “Bundled payments leave the overarching incentive to increase volume solidly in place.”
In a separate JAMA Viewpoint article, Ibrahim et al. warn that another CMS program, the Comprehensive Care for Joint Replacement (CJR) model, could unintentionally amplify already existing racial disparities in elective joint replacement. CJR is a mandatory initiative in 67 randomly selected US metropolitan areas. The authors say that CJR might improve postoperative quality of care for minority patients after joint replacement, but that the program could also end up favoring healthier, well-insured patients.
Overall, at this early stage, these two CMS models offer promising, comprehensive approaches to joint replacement that may prove cost-saving without comprising quality of care. Results like the ones published by Dummit et al. are hopeful, but longer-term, outcomes-based, and cost-focused studies that include epidemiologic and racial impact must be performed as we move forward carefully.
Richard Yoon, MD is a fellow in orthopaedic traumatology and complex adult reconstruction at Orlando Regional Medical Center.
The answer to that question depends largely on how much the 90-day episode of care actually costs. Virk et al., in the August 17, 2016 edition of JBJS, provide benchmark data that could help policymakers design bundled payments for cervical fusions that are economically viable for providers.
The authors analyzed the Medicare 5% National Sample Administrative Database and found that 4,506 patients in that cohort underwent a one to two-level anterior cervical discectomy and fusion (ACDF) for cervical radiculopathy from 2005 to 2012. The mean cost per patient of the procedure plus the 90-day postoperative period was $15,417. The physician reimbursement represented 20.4% of that total, with the surgeon receiving 18% of the total. Reimbursements for hospitals for inpatient care represented nearly 73% of the total reimbursement. The study did not account for reimbursements from “Medigap” plans or private payers.
The authors also analyzed data from the same database for 90-day episodes of care related to total knee arthroplasty (TKA). The mean per-patient reimbursement for TKA patients was $17,451. The authors noted significant regional variation in reimbursement for ACDF, with the lowest rates in the Northeast and Midwest and the highest rates in the West.
Among the conclusions made by Virk et al. is the following: “Although payments to physicians have been implicated in the rise of health-care costs, the data suggest that the greater opportunity for reducing expenses involves hospital-related reimbursement.”
Click here for more OrthoBuzz coverage of bundled payments in orthopaedics.
“Alternative payment models are here to stay,” according to an AOA Critical Issues article by Greenwald et al. in the June 1, 2016 issue of The Journal of Bone & Joint Surgery. The article identifies successful implementation strategies related to the Bundled Payments for Care Improvement (BPCI) initiative launched by the Centers for Medicare and Medicaid Services (CMS) in 2013.
Alternative payment models represent an opportunity to reduce costs by eliminating waste and unwarranted variation in care by finding efficiencies within the system. One way to achieve this is through gainsharing incentives that align hospitals, physicians, and post-acute care providers in the redesign of care. But participants also assume financial risk.
Orthopaedics plays a big role in the BPCI risk-reward initiative. Sixteen of the 48 clinical “episodes of care” included in BPCI are orthopaedic-related. Moreover, three episodes (major lower-extremity joint replacement, femur/hip/pelvis fractures, and “medical non-infectious orthopaedic”) account for 40% of the 16 orthopaedic episodes being evaluated.
The nuts and bolts—and risks and rewards— of the initiative are well-described in the article, but here are several pearls extracted therefrom:
“Care improvement activities and care redesign…are the necessary prerequisites before entering into bundled payment arrangements.”
“The financial risk is real [because] outliers, those patients whose cost is substantially higher than the mean patient cost, cannot be controlled.”
“It is important that the physician or surgeon responsible for the patient is involved in all stages of the episode of care and interacts with all of the parties involved.”
“Specific to orthopaedics, there are substantial opportunities for cost savings by integrating preoperative and intraoperative processes, reviewing implant purchasing options, and negotiating post-acute care costs.”
“Changes in care delivery often require … managing patient expectations.”
The final rule from the Centers for Medicare & Medicaid Services (CMS) regulating “episode-of-care” Medicare payments to hospitals for hip and knee replacements includes a postponed start date of April 1, 2016. The originally proposed implementation date was January 1, 2016.
Approximately 800 hospitals nationwide are subject to the new payment model, which makes hospitals eligible for bonuses or penalties, depending on their quality and cost performance from the day of patient admission to 90 days post-discharge. Based on comments about the initial rule by 400 key stakeholders, CMS also agreed to eliminate penalty payments during the first year of implementation.
Because the CMS model—dubbed Comprehensive Care for Joint Replacement, or CJR—permits gainsharing, individual orthopaedic surgeons could benefit financially if hospitals they are affiliated with receive bonuses. The AAOS commended CMS for revising the methodology for calculating the composite quality score and said that the delayed implementation “adds some flexibility,” but the group is still calling for CMS to “postpose the mandatory implementation feature of the program until at least 85 percent of providers have attained meaningful use [of EHRs] or another metric of infrastructure readiness.”
Concerns have arisen that the implementation of value-based, alternative payment models pegged to “bundled” episodes of care and/or patient outcomes may make it harder for a subset of patients to access the care they need. Specifically, some surgeons may be apprehensive to treat patients who have substantial medical comorbidities or socioeconomic situations that increase their risk of postsurgical complications and poor outcomes, because these alternative payment models often financially penalize physicians and hospitals for the cost of suboptimal results. The study by Shau et al. in the December 5, 2018 issue of The Journal provides data that sharpens the horns of this dilemma.
The authors used the National Readmissions Database to perform a propensity-score-matched comparison between >5,300 patients with Medicaid payer status who underwent a primary total hip arthroplasty (THA) and an equal number of patients with other types of insurance who also underwent primary THA. Shau et al. found that Medicaid-covered THA patients had significantly increased overall readmission rates (28.8% vs 21%, p <0.001, relative risk=1.37), mean length of stay (4.5 vs 3.3 days, p <0.0001), and mean total cost of care ($71,110 vs $65,309, p <0.0001), relative to the other group. These results strongly suggest that Medicaid payer status is an independent factor associated with increased resource utilization after total hip arthroplasty.
These findings can be viewed from a couple of different perspectives. First, from a preventive standpoint, surgeons and healthcare systems providing THA for Medicaid patients may need to spend more time preoperatively optimizing these patients (both physically and psychosocially) to decrease their postoperative resource burden and increase the likelihood of a good clinical outcome. Second, these results are further proof that any fair and effective alternative payment model needs to take into consideration factors such as Medicaid payer status and patient comorbidities. If they do not, such models will actually throw access barriers in front of patients in this demographic because providers may feel that caring for them increases the likelihood of being penalized financially.
Both perspectives are valid, so Medicaid payer status is a crucial factor to consider as alternative payment programs move forward. Nowadays, controlling costs is an important goal of any healthcare delivery system, but it must not lead to unintended discrimination in patient access to care. As we create further alternative payment models and refine existing ones, we must be careful not to prioritize cost cutting ahead of equitable patient access.
Chad A. Krueger, MD
JBJS Deputy Editor for Social Media
Wide variability in the cost and quality of health care in the US has led some to describe our system as “uniquely inefficient.” Consequently, we continue to study variability intensely, especially in the realm of joint arthroplasty. In the June 3, 2020 issue of The Journal, Schilling et al. elegantly analyze the variations in 90-day episode payments made by Medicare Part A for total knee arthroplasty (TKA) from 2014 to 2016. In so doing, they provide a snapshot of hospital cost performance and, just as importantly, they offer a methodology by which to measure future hospital-level cost performance with this very popular surgery.
The authors reviewed >700,000 TKAs in the Medicare population at a time prior to the full implementation of the Comprehensive Care for Joint Replacement (CJR) model, and they ranked >3,200 hospitals within 9 US regions to determine cost performance. Schilling et al. found that during those 3 years, the mean Medicare episode payment for TKA decreased significantly, due almost entirely to a >$1,500 per-case decrease in post-acute care payments, which included lower costs for skilled nursing facilities and inpatient rehabilitation. Also decreasing during that same period were length of hospital stay and 90-day readmission rates.
These findings highlight the improvements in care and cost efficiency that were occurring even before implementation of the CJR. In a Commentary on this study, Susan Odum, PhD suggests that “the improved value of TKA illustrated by Schilling et al. includes the successful impacts of the BPCI [Bundled Payments for Care Improvement] program,” an alternative payment model that Medicare rolled out beginning in 2013.
On the other hand, the authors also reveal a persistently high degree of variability in episode payments and resource utilization both across and within geographic regions. This strongly suggests the possibility of further improvement. Regardless of which, if any, alternative payment model we participate in, everyone in the orthopaedic community should think about how to become more efficient in our delivery of musculoskeletal care. And this study provides a conceptual framework and benchmarks for identifying where the room for improvement is.
Matthew R. Schmitz, MD
JBJS Deputy Editor for Social Media
During a well-attended symposium on bundled payment initiatives for joint replacement at the 2015 AAOS Annual Meeting, speakers shared enlightening pearls and pitfalls related to Medicare’s Bundled Payments for Care Improvement initiative. But no one mentioned the fact that by 2018, Medicare will shift the 90-day global period for joint replacement—and all other covered surgeries—to a 0-day global period.
This fact is discussed in an eye-opening Perspective by Mulcahey et al. in the April 9 New England Journal of Medicine. Noting that bundled payments in general are designed to improve care and reduce cost, the authors call this decision, which would essentially unbundle postoperative visits, “striking.” The shift to a 0-day global period for surgery is based on an HHS Inspector General audit that found that the number of postoperative encounters between surgeons and patients are actually well below the number paid for in the 90-day bundle. Total knee arthroplasty, for example, includes three inpatient, one hospital-discharge, and three outpatient surgeon visits in its 90-day package.
Mulcahey et al. contend that “removing some or all postoperative visits from global packages will reduce procedure payment rates” for surgeons, but it remains to be seen how surgeons, orthopaedic and otherwise, will respond to the policy change. OrthoBuzz will keep you posted.