Abstract
Objective:
To evaluate the impact of the CJR policy on the 90-day trajectory of post-acute care following a total hip arthroplasty (THA).
Design:
Multivariable difference-in-difference models applied to Medicare beneficiaries undergoing a THA prior to (2014–2015) and post-CJR implementation (2017) in areas subjected to or exempt from the policy.
Setting:
Hospitals in Standard Metropolitan Statistical Areas (SMSAs).
Participants:
357,844 elderly Medicare patients nationwide undergoing THA.
Interventions:
None
Main Outcome Measures:
Escalation in care to institutionalization (i.e., admission to an inpatient rehabilitation (IRF) or skilled nursing facility (SNF)) during 90-days post-discharge for those initially discharged to the community and return to the community at the end of the episode of care among those initially discharged to an institutional setting.
Results:
Of the 357,844 elderly Medicare patients nationwide undergoing THA during the study period, 47.6% were discharged directly to the community while 52.4% received post-acute care in an institution. Patients discharged to an institution post-policy in a CJR area were about 10% less likely to return to the community (OR=0.91; 95 %CI=0.84 – 0.98; p=0.02) at the end of the 90-day episode of care than those treated in policy exempt areas. Despite large in magnitude, estimates of escalation in care among patients treated in bundling areas post-CJR implementation were not statistically significant.
Conclusions:
Our findings support further exploration of unanticipated effects of mandatory bundled payment policies on outcomes, as well as further examination of outcomes among policy-relevant sub-groups of patients undergoing hip replacement in the U.S.
Keywords: bundled payment models, total hip arthroplasty, post-acute discharge disposition, inpatient rehabilitation, skilled nursing facility
Introduction
Total hip arthroplasty (THA) provides definitive treatment for refractory, end-stage osteoarthritis.1 Approximately 600,000 THAs are performed in the US each year, with an estimated increase in case volume to nearly 1.5 million by 2040.2,3 Total hip arthroplasty remains one of the most common operations performed in patients with Medicare, accounting for a substantial portion of healthcare expenditures in this patient group.4 In 2014, lower extremity joint replacements were associated with hospital costs of $7 billion.4
In 2016, the Centers for Medicare & Medicaid Services (CMS) implemented the Comprehensive Care for Joint Replacement (CJR) Model,4 the first mandatory, episode-based, prospective bundling payment model for persons undergoing elective joint replacements. Bundles include all relevant Medicare costs during an episode-of-care, defined as the inpatient stay and 90 days post-acute discharge, regardless of post-acute care setting. The CJR model, which was mandatory to virtually all hospitals located in any of the 67 randomly selected SMSAs where the policy was initially implemented, effectively shifted risk from payers to providers, a strategy designed to induce hospitals to optimize post-acute care across settings by choosing the most cost-effective alternative and eliminating the least valued services within the bundle. Although initially slated to occur from April 2016 through December 2020, CMS drastically reduced the number of SMSAs assigned to CJR by 2018. In 2020, a three-year extension was applied to the CMS CJR ruling, which began including joint replacements performed on an outpatient basis in the remaining participating SMSAs.5
Early research evaluating the impact of the CJR model have noted a modest decrease in spending per episode of care, largely driven by a decrease in post-acute care spending.6–8 After implementation of the CJR, fewer patients were discharged to an inpatient rehabilitation facility (IRF) or skilled nursing facility (SNF), and a corresponding increase in the proportion of patients discharged to home was observed.6,7,9 As would be expected, data have shown that discharge to an IRF or SNF results in a significantly higher cost of care compared to those discharged to home.10 This shift in discharge disposition has been reported without a significant increase in rates of postoperative morbidity, mortality, or readmission.7–9 More recently, data have shown that the cost savings observed after the first two years of the CJR bundled payment model have dissipated after the fourth year, likely due to hospitals’ responses to the changes made to the policy in the second year after implementation.11 Policy evaluations of CJR to date have included any type of lower extremity arthroplasty, which presents challenges for clinical considerations in discharge planning for specific procedures, as post-acute care rehabilitation needs can differ significantly between total knee or ankle arthroplasty and THA.
Despite previous research describing the effects of CJR on expenditures, early postoperative discharge disposition, morbidity, mortality, readmission, and disparities therein,12 little is known about the trajectories of patients’ disposition and, more specifically, their setting at the end of the 90-day episode of care. We address both issues in this paper. Taking advantage of the natural experiment afforded by the exogenous regional variation in implementation of the CJR, we test the hypothesis that cost-saving incentives introduced by CJR may, on the margin, lead to inappropriate discharge disposition to home for patients that would have been best served in a post-acute care facility. We do so by contrasting the escalation in care –defined as admission to an IRF or SNF over the 90-day episode of care following initial home discharge– of Medicare beneficiaries undergoing a THA in areas subjected to the policy before and after implementation of CJR to those of a contemporary cohort of beneficiaries treated at hospitals in areas exempt from the policy. We also examine whether, conversely, patients initially discharged to an institutional setting (IRF or SNF) in bundling areas after the policy implementation are less likely than their counterparts in non-bundling areas to return home at the end of the episode of care, even after controlling for prior SNF use and other potential confounders.
Methods
Study Population and Data Sources
Data were drawn from the CMS’s Medicare FFS Research Identifiable Inpatient, Outpatient and Carrier files for the years prior to the CJR policy (2014–2015) and post-CJR policy (2017) for the subset of patients age ≥66 years with DRG 469 and 470 who had undergone a primary THA based on ICD-9 and ICD-10 procedure codes. Additional inclusion criteria consisted of continuous enrollment in Medicare Parts A and B and not C for 12-months prior to admission (for calculation of comorbidities) and for the 90-day period after acute discharge. For the purposes of analysis, two categories of initial discharge and 90-day setting were considered: 1) community (home with or without home health), and 2) institutional (IRF or SNF). The small number of patients with other discharge disposition (i.e., long-term care hospitals, psychiatric hospitals, hospice, or death) accounted for 2.7% of the sample and were excluded from the analyses.
Key Measures
The primary outcomes of interest were either 1) escalation in care level to an IRF or SNF during 90-days post-discharge for those initially discharged to the community, and 2) return to the community at the end of the episode of care among those initially discharged to an institutional setting. Age, gender, race/ethnicity, and dual enrollment in Medicaid or a state buy-program during the month of the index THA surgery was determined from the Master Beneficiary Summary File. Comorbidity was computed by applying the Elixhauser algorithm to inpatient, outpatient, and carrier Medicare claims during the 12-month period preceding the index surgery.13,14 Hip fractures were identified by ICD9 and 10 codes, specifically ICD-9-CM 820.00 – 820.19 and ICD-10 S72.0xxn – S72.099n, where “x” is a wildcard and “n” is “A”, “B”, or ”C.”. Census division of the treatment hospital was determined by the Provider of Service File. Finally, SNF use during the year prior to the index THA was determined from the Medicare Skilled Nursing Facility files.
Statistical Analysis
CMS implemented the mandatory bundled payment model in 67 randomly assigned SMSAs. With the exception of hospitals participating in the voluntary bundled payment program (BPCI), which were allowed to opt-out from the mandatory payment model, all hospitals in SMSAs assigned for implementation of the mandatory bundling payment model were subjected to the CJR Rule.
Leveraging this natural experiment design, we contrast changes over time in outcomes among patients treated at hospitals that were exempt from CJR (controls) to changes over time among patients treated at hospitals where the policy was implemented (treatment cases) to make causal inferences about policy effects. Specifically, we applied difference-in-difference techniques to estimate the policy’s impact on outcomes, after testing for and confirming that the pre-policy parallel trends assumption held for both outcomes.15 By comparing the escalation/deescalation of care level during the 90-day episode of care for patients treated in bundling areas to that of Medicare patients treated in non-bundling MSAs, we identify the effect likely attributable to the CJR policy after factoring out any temporal trend in the overall Medicare joint replacement population during the study timeframe.
A logit specification was used to model both outcomes. These multivariable regressions included an indicator for time period (2017 vs 2014–15), geographic area (bundling vs. non-bundling), and the interaction term between these two factors. All analyses were adjusted for patients’ age, gender, race/ethnicity, poverty status (as measured by dual enrollment in Medicaid or a state buy-in program), number of comorbidities (0, 1, 2, 3, 4+), an indicator for whether the hip replacement was secondary to a hip fracture, prior SNF stay within 6-months from the THA surgery, and Census division of the acute-care treatment hospital. To account for intra-facility correlation due to clustering of patients at the hospital of admission, we applied the cluster Huber-White sandwich estimator. All analyses were conducted using the STATA 15 software package.
Results
There were a total of 357,844 episodes of care associated with THA hospitalizations among Medicare beneficiaries over the 3-year study period. Of those, 170,265 (47.6%) were discharged to the community (home with or without home health) while the remainder were discharged to an institutional setting (Table 1). As expected, relative to those discharged to the community, patients discharged to an institutional setting were markedly older (35.1% vs. 5.5% were 85 or older), had a substantially higher comorbidity burden (37.1% vs. 17.4% with 4 or more comorbidities), had at least one prior SNF stay (13.6% vs. 2.1%, respectively), were more likely to have sustained a hip fracture (27.5% vs. 3.3%), and had a longer length of stay during their acute care surgical hospitalization. They were also more than three times as likely to be poor (10.9% vs. 3%) than patients discharged to the community.
Table 1.
Patient demographics, comorbidity score, and admission characteristics by immediate discharge location.
| Variable | Overall | Initial post-acute setting | p value | |
|---|---|---|---|---|
| Community | Institutional | |||
| N | 357,844 | 170,265 (47.6%) | 187,579 (52.4%) | |
| Age, years; n (%) | <0.0001 | |||
| 66–69 | 62874 (17.6) | 48497 (28.5) | 14377 (7.7) | |
| 70–74 | 82562 (23.1) | 55520 (32.6) | 27042 (14.4) | |
| 75–79 | 74734 (20.9) | 37762 (22.2) | 36972 (19.7) | |
| 80–84 | 62565 (17.5) | 19252 (11.3) | 43313 (23.1) | |
| 85–89 | 46620 (13.0) | 7075 (4.2) | 39545 (21.1) | |
| ≥90 | 28489 (8.0) | 2159 (1.3) | 26330 (14.0) | |
| Sex, n (%) | <0.0001 | |||
| Female | 236914 (66.2) | 99339 (58.3) | 137575 (73.3) | |
| Male | 120930 (33.8) | 70926 (41.7) | 50004 (26.7) | |
| Race, n (%) | <0.0001 | |||
| Non-Hispanic White | 330473 (92.4) | 159192 (93.5) | 171281 (91.3) | |
| Black/African American | 13544 (3.8) | 5646 (3.3) | 7898 (4.2) | |
| Hispanic | 7605 (2.1) | 2961 (1.7) | 4644 (2.5) | |
| Other | 6222 (1.7) | 2466 (1.4) | 3756 (2.0) | |
| Dual Enrollee, n (%) | 25615 (7.2) | 5102 (3.0) | 20513 (10.9) | <0.0001 |
| CJR Policy Area, n (%) | 127573 (35.7) | 59974 (35.2) | 67599 (36.0) | <0.0001 |
| Time Period, n (%) | <0.0001 | |||
| Pre (2014–2015) | 229828 (64.2) | 96485 (56.7) | 133343 (71.1) | |
| Post (2017) | 128016 (35.8) | 73780 (43.3) | 54236 (28.9) | |
| Prior SNF stay, n (%) | 29124 (8.1) | 3564 (2.1) | 25560 (13.6) | <0.0001 |
| Number of comorbidities, n (%) | <0.0001 | |||
| 0 | 55666 (15.6) | 34147 (20.1) | 21519 (11.5) | |
| 1 | 77658 (21.7) | 45979 (27.0) | 31679 (16.9) | |
| 2 | 72458 (20.2) | 37222 (21.9) | 35236 (18.8) | |
| 3 | 52807 (14.8) | 23321 (13.7) | 29486 (15.7) | |
| ≥4 | 99255 (27.7) | 29596 (17.4) | 69659 (37.1) | |
| Hip fracture, n (%) | 57250 (16.0) | 5594 (3.3) | 51656 (27.5) | <0.0001 |
| Length of stay, days; median [IQR] | 3.0 [2.0–4.0] | 2.0 [2.0–3.0] | 4.0 [3.0–5.0] | <0.0001 |
| Census Division of Hospital, n (%) | <0.0001 | |||
| Mid-Atlantic | 46852 (13.1) | 19080 (11.2) | 27772 (14.8) | |
| Northeast | 22286 (6.2) | 8821 (5.2) | 13465 (7.2) | |
| South Atlantic | 80173 (22.4) | 39879 (23.4) | 40294 (21.5) | |
| East North Central | 55792 (15.6) | 25289 (14.9) | 30503 (16.3) | |
| East South Central | 20511 (5.7) | 8760 (5.1) | 11751 (6.3) | |
| West North Central | 27898 (7.8) | 14428 (8.5) | 13470 (7.2) | |
| West South Central | 34532 (9.7) | 15237 (8.9) | 19295 (10.3) | |
| Mountain | 25308 (7.1) | 14626 (8.6) | 10682 (5.7) | |
| Pacific | 44492 (12.4) | 24145 (14.2) | 20347 (10.8) | |
MSA = metropolitan statistical area; SNF = skilled nursing facility; IQR = interquartile range.
Bundled Payment and Escalation to Institutional Care
Multivariable estimates shown in Table 2 indicate that, regardless of treatment at a hospital subject to or exempt from the CJR bundled payment policy, more recent (i.e., 2017 relative to 2014–15) THA patients initially discharged to the community were less likely to experience an escalation in care setting during the 90-day episode of care (OR= 0.85; 95% CI 0.77 – 0.94; p=0.001), suggesting a temporal trend towards more appropriate discharge planning that is unrelated to the bundled payment policy.
Table 2.
Multivariable logistic regression results for escalation to institutional care during 90-days post-discharge among patients initially discharged to the community.
| Variable | Odds Ratio (OR) | 95% CI | p value |
|---|---|---|---|
| Time Period: 2017 (reference) | 0.85 | 0.77 – 0.94 | 0.001 |
| Policy Area: CJR SMSA (reference) | 0.97 | 0.87 – 1.08 | 0.590 |
| Difference-in-difference: Time Period*Area | 1.16 | 0.98 – 1.36 | 0.065 |
| Age | |||
| 66–69 | Reference | ||
| 70–74 | 1.26 | 1.11 – 1.43 | <0.001 |
| 75–79 | 1.86 | 1.64 – 2.11 | <0.001 |
| 80–84 | 2.71 | 2.37 – 3.10 | <0.001 |
| 85–89 | 3.59 | 3.05 – 4.23 | <0.001 |
| ≥90 | 3.37 | 2.64 – 4.30 | <0.001 |
| Male sex | 0.73 | 0.67 – 0.80 | <0.001 |
| Race | |||
| Non-Hispanic White | Reference | ||
| Black/African American | 1.13 | 0.92 – 1.39 | 0.23 |
| Hispanic | 0.95 | 0.71 – 1.27 | 0.74 |
| Other | 0.60 | 0.40 – 0.90 | 0.02 |
| Low income | 1.36 | 1.14 – 1.63 | <0.001 |
| Number of Comorbidities | |||
| 0 | Reference | ||
| 1 | 1.10 | 0.95 – 1.27 | 0.19 |
| 2 | 1.46 | 1.27 – 1.68 | <0.001 |
| 3 | 1.71 | 1.47 – 2.0 | <0.001 |
| ≥4 | 2.55 | 2.22 – 2.92 | <0.001 |
| Hip fracture | 1.50 | 1.27 – 1.76 | <0.001 |
| Prior SNF stay | 1.91 | 1.62 – 2.25 | <0.001 |
| Census Division of Hospital | |||
| Mid-Atlantic | Reference | ||
| Northeast | 0.93 | 0.75 – 1.15 | 0.50 |
| South Atlantic | 0.81 | 0.71 – 0.93 | 0.003 |
| East North Central | 0.97 | 0.84 – 1.13 | 0.70 |
| East South Central | 0.96 | 0.78 – 1.17 | 0.66 |
| West North Central | 0.84 | 0.71 – 1.0 | 0.06 |
| West South Central | 1.02 | 0.87 – 1.21 | 0.77 |
| Mountain | 0.90 | 0.75 – 1.08 | 0.26 |
| Pacific | 0.76 | 0.65 – 0.89 | <0.001 |
CI = confidence interval; CJR = Comprehensive Care for Joint Replacement; SMSA = standard metropolitan statistical area; SNF = skilled nursing facility.
Although not statistically significant at conventional levels, difference-in-difference point estimates suggest a trend towards higher escalation in care among patients treated in bundling areas relative to those in non-bundling areas (OR=1.16, 95% CI 0.98 – 1.36; p=0.065).
Bundled Payment and Return to the Community
As shown in Table 3, regardless of treatment at a hospital subject to or exempt from the CJR policy, there was a temporal trend towards increased return to the community at the end of the episode of care among THA patients initially discharged to an institution (OR=1.41, 95% CI 1.34 – 1.48; p<0.001). Here again, however, the magnitude of the trend differed among the bundling (treatment) and non-bundling (control) groups.
Table 3.
Multivariable logistic regression results for community-living at the end of the 90-day episode of care among patients initially discharged to an institutional setting.
| Variable | Odds Ratio (OR) | 95% CI | p value |
|---|---|---|---|
| Time period: 2017 (reference) | 1.41 | 1.34 – 1.48 | <0.001 |
| Policy Area: CJR MSA (reference) | 1.01 | 0.96 – 1.05 | 0.80 |
| Difference-in-difference: Time Period* Area | 0.91 | 0.84 – 0.98 | 0.02 |
| Age | |||
| 65–69 | Reference | ||
| 70–74 | 0.80 | 0.73 – 0.89 | <0.001 |
| 75–79 | 0.61 | 0.56 – 0.67 | <0.001 |
| 80–84 | 0.47 | 0.43 – 0.52 | <0.001 |
| 85–89 | 0.37 | 0.34 – 0.40 | <0.001 |
| ≥90 | 0.28 | 0.25 – 0.30 | <0.001 |
| Male | 0.80 | 0.77 – 0.83 | <0.001 |
| Race | |||
| Non-Hispanic White | Reference | ||
| Black/African American | 0.89 | 0.83 – 0.96 | 0.003 |
| Hispanic | 1.38 | 1.24 – 1.53 | <0.001 |
| Other | 1.25 | 1.11 – 1.41 | <0.001 |
| Dual Enrollee/Low Income | 0.34 | 0.33 – 0.36 | <0.001 |
| Number of comorbidities | |||
| 0 | Reference | ||
| 1 | 1.09 | 1.02 – 1.17 | 0.02 |
| 2 | 0.99 | 0.93 – 1.06 | 0.83 |
| 3 | 0.90 | 0.84 – 0.97 | 0.004 |
| ≥4 | 0.66 | 0.62 – 0.71 | <0.001 |
| Hip fracture | 0.61 | 0.59 – 0.63 | <0.001 |
| Prior SNF stay | 0.66 | 0.63 – 0.69 | <0.001 |
| Census Division of Treating Hospital | |||
| Mid-Atlantic | Reference | ||
| Northeast | 1.16 | 1.06 – 1.26 | <0.001 |
| South Atlantic | 0.90 | 0.85 – 0.96 | <0.001 |
| East North Central | 0.78 | 0.73 – 0.82 | <0.001 |
| East South Central | 0.79 | 0.73 – 0.85 | <0.001 |
| West North Central | 0.72 | 0.67 – 0.78 | <0.001 |
| West South Central | 0.62 | 0.58 – 0.66 | <0.001 |
| Mountain | 1.18 | 1.07 – 1.30 | <0.001 |
| Pacific | 1.08 | 1.01 – 1.16 | 0.02 |
CI = confidence interval; CJR = Comprehensive Care for Joint Replacement; SMSA = standard metropolitan statistical area; SNF = skilled nursing facility.
The difference-in-difference estimate, which captures the extent to which outcomes among patients treated in bundling versus non-bundling areas over time after the policy implementation, indicate that, among THA patients initially discharged to an IRF or SNF, those undergoing hip replacement surgery in the post-CJR policy period in areas subject to bundled payments had significantly lower relative odds of having returned to the community at 90-days post-discharge (OR=0.91, 95% CI 0.84 – 0.98; p=0.02) than those in non-CJR bundled payment areas.
Discussion
This study of 357,844 elderly Medicare beneficiaries undergoing THA nationwide found that patients in CJR areas who were initially discharged to an IRF or SNF were significantly less likely to be at home at the end of the episode of care compared to those in non-CJR MSAs. This finding raises concerns about CJR’s impact on discharge disposition and successful return to community among patients undergoing THA after CJR implementation.
Previous studies examining the impact of the mandatory bundled payment model for lower extremity arthroplasty have reported a decrease in spending, without a significant increase in morbidity, mortality, or readmission.6–8 Such cost savings have been largely driven by the reduction in discharge to institutional settings.7–9 Although our data, showing a marked reduction in the proportion of patients initially discharged to an IRF or SNF, from 58% in the pre-CJR policy period to 42% post-CJR policy period, support these findings, our study differs from previous literature in important ways.
To date, studies evaluating the impact of the CJR policy have focused on its overall effect, pooling episodes of care at the diagnosis-related group (DRG)- level regardless of procedure type. The heterogeneity within the two DRG groups targeted by the policy (469 and 470), which include patients undergoing arthroplasty at the ankle, knee, and hip level, may have masked important sub-group effects. These procedures differ both in the complexity and intensity of post-acute care rehabilitation requirements. By focusing on the subset of patients undergoing THA, we uncovered policy effects of the CJR bundled payment model on those most likely to require costly, post-acute institutional rehabilitation care. In our analyses, we also control for individuals’ pre-surgery SNF stay, an important confounder of post-acute initial discharge disposition and trajectory of care settings.
Appropriate discharge planning is a critical component in achieving the overall goal of the CJR model and previous research has questioned the wisdom of including hip fractures in the CJR bundled payment model.16–18 Our findings indicating that hip fractures experienced a lower likelihood of return to the community at the end of the 90-day episode of care among those discharged to an institution seem to support that assertion. This group of patients, therefore, may be best excluded from such bundled payment models.
Limitations
Our findings should be interpreted in light of the same inherent limitations of other studies relying on claims data. Our results reflect the experience of FFS Medicare beneficiaries and may not be generalizable to the larger population of THA patients. In addition, changes to the CJR policy altered the proportion of MSAs required to participate in the policy as well as inclusion of lower volume or rural hospitals. In recent years, an increased proportion of THAs are performed at outpatient surgery centers, which are now included in the CJR model.20 Additional research will be needed to assess the impact of the CJR model and initial and 90-day discharge dispositions after THAs performed on an outpatient basis. Most importantly, this was not a traditional randomized trial, and even with state-of-the-art difference-in-differences methods applied to a natural experiment, there remains the possibility of residual confounding. Other changes in the health care delivery system or in the post-acute care landscape during the study period, for example, might have affected our results to the extent that they occurred differentially across bundling and non-bundling areas. Another concern is the potential effect of hospitals with a substantial participation in the voluntary bundling program (BPCI), which were allowed to opt-out of CJR. Following the intention-to-treat principle, we did not exclude such hospitals from our analysis. That said, there is no evidence to suggest that distribution of BPCI hospitals varies systematically across SMSAs randomized to CJR, which limits the concern about bias.19
Despite these limitations, our analysis contributes to the literature assessing the impact of the CMS CJR model on an episode of care by addressing the gap in knowledge regarding patients’ trajectories during the critical 90-day post-discharge episode of care.
Conclusions
The CJR model’s bundled payment policy was associated with reduced return to the community at the end of the 90-day episode of care among those initially discharged to an institution. These findings support further exploration of unanticipated effects of mandatory bundled payment policies on outcomes, as well as further examination of outcomes among vulnerable sub-groups of patients subjected to the policy. Risk-stratification strategies to identify those likely to require escalation of care after THA could be incorporated into the discharge planning process to increase the policy’s overall effectiveness.
Funding
This research was supported in part by a research grant from the National Institute on Aging (NIA) to Drs. Dillingham and Pezzin (5-R01-AG058718).
Abbreviations:
- THA
Total hip arthroplasty
- CJR
Comprehensive Care for Joint Replacement
- SMSA
Standard metropolitan statistical areas
- IRF
Inpatient rehabilitation facility
- SNF
Skilled nursing facility
- CMS
Centers for Medicare and Medicaid Services
- DRG
Diagnosis-related group
- FFS
Fee for service
- BPCI
Bundled Payments for Care Improvement Initiative
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosures
The authors have no conflicts of interest to disclose.
Contributor Information
Kara Kallies, Institute for Health and Equity, Medical College of Wisconsin.
Timothy R. Dillingham, Department of Physical Medicine and Rehabilitation, University of Pennsylvania.
Adam Edelstein, Department of Orthopedic Surgery, Medical College of Wisconsin.
Eric Hume, Department of Orthopedic Surgery, University of Pennsylvania.
Daniel Polsky, Bloomberg School of Public Health, the Johns Hopkins University.
Roy Schwartz, University of Pennsylvania.
Emily L. McGinley, Center for Advancing Population Sciences, Medical College of Wisconsin.
Liliana E. Pezzin, Institute for Health and Equity, Medical College of Wisconsin.
References
- 1.Katz JN, Arant KR, Loeser RF. Diagnosis and Treatment of Hip and Knee Osteoarthritis: A Review. JAMA - Journal of the American Medical Association. 2021;325(6):568–578. doi: 10.1001/jama.2020.22171 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Agency for Healthcare Research and Quality. HCUP Fast Stats - Most Common Operations During Inpatient Stays. Accessed May 20, 2021. https://www.hcupus.ahrq.gov/faststats/NationalProceduresServlet
- 3.Singh JA, Yu S, Chen L, Cleveland JD. Rates of Total Joint Replacement in the United States: Future Projections to 2020–2040 Using the National Inpatient Sample. The Journal of Rheumatology. 2019;46(9). doi: 10.3899/jrheum.170990 [DOI] [PubMed] [Google Scholar]
- 4.Centers for Medicare and Medicaid Services. Comprehensive Care for Joint Replacement Model. Accessed May 31, 2021. https://innovation.cms.gov/innovation-models/cjr
- 5.Centers for Medicare & Medicaid Services (CMS). Medicare Program: Comprehensive Care for Joint Replacement Model Three-Year Extension and Changes to Episode Definition and Pricing; Medicare and Medicaid Programs; Policies and Regulatory Revisions in Response to the COVID–19 Public Health Emergency. Department of Health and Human Services.Centers for Medicare & Medicaid Services (CMS); 2021:23496–23576. [Google Scholar]
- 6.Haas DA, Zhang X, Kaplan RS, Song Z. Evaluation of Economic and Clinical Outcomes Under Centers for Medicare & Medicaid Services Mandatory Bundled Payments for Joint Replacements. JAMA Internal Medicine. 2019;179(7). doi: 10.1001/jamainternmed.2019.0480 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Barnett ML, Wilcock A, McWilliams JM, et al. Two-Year Evaluation of Mandatory Bundled Payments for Joint Replacement. New England Journal of Medicine. 2019;380(3):252–262. doi: 10.1056/nejmsa1809010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Finkelstein A, Ji Y, Mahoney N, Skinner J. Mandatory Medicare Bundled Payment Program for Lower Extremity Joint Replacement and Discharge to Institutional Postacute Care. JAMA. 2018;320(9). doi: 10.1001/jama.2018.12346 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Agarwal R, Liao JM, Gupta A, Navathe AS. The impact of bundled payment on health care spending, utilization, and quality: A systematic review. Health Affairs. 2020;39(1):50–57. doi: 10.1377/hlthaff.2019.00784 [DOI] [PubMed] [Google Scholar]
- 10.Sabeh KG, Rosas S, Buller LT, Roche MW, Hernandez VH. The Impact of Discharge Disposition on Episode-of-Care Reimbursement After Primary Total Hip Arthroplasty. The Journal of Arthroplasty. 2017;32(10). doi: 10.1016/j.arth.2017.04.062 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wilcock AD, Barnett ML, McWilliams JM, Grabowski DC, Mehrotra A. Hospital Responses to Incentives in Episode-Based Payment for Joint Surgery. JAMA Internal Medicine. Published online May 17, 2021. doi: 10.1001/jamainternmed.2021.1897 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li Y, Ying M, Cai X, Thirukumaran CP. Association of Mandatory Bundled Payments for Joint Replacement with Postacute Care Outcomes Among Medicare and Medicaid Dual Eligible Patients. Med Care. 2021. Feb 1;59(2):101–110. doi: 10.1097/MLR.0000000000001473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Quan H, Sundararajan V, Halfon P, et al. Coding Algorithms for Defining Comorbidities in ICD-9-CM and ICD-10 Administrative Data. Medical Care. 2005;43(11). doi: 10.1097/01.mlr.0000182534.19832.83 [DOI] [PubMed] [Google Scholar]
- 14.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity Measures for Use with Administrative Data. Medical Care. 1998;36(1). doi: 10.1097/00005650-199801000-00004 [DOI] [PubMed] [Google Scholar]
- 15.Athey S and Imbens GW, 2006. Identification and inference in nonlinear difference-in-differences models. Econometrica, 74(2), pp. 431–497. [Google Scholar]
- 16.Schroer WC, Diesfeld PJ, LeMarr AR, Morton DJ, Reedy ME. Hip Fracture Does Not Belong in the Elective Arthroplasty Bundle: Presentation, Outcomes, and Service Utilization Differ in Fracture Arthroplasty Care. The Journal of Arthroplasty. 2018;33(7). doi: 10.1016/j.arth.2018.02.091 [DOI] [PubMed] [Google Scholar]
- 17.Cairns MA, Moskal PT, Eskildsen SM, Ostrum RF, Clement RC. Are Medicare’s “Comprehensive Care for Joint Replacement” Bundled Payments Stratifying Risk Adequately? Journal of Arthroplasty. 2018;33(9):2722–2727. doi: 10.1016/j.arth.2018.04.006 [DOI] [PubMed] [Google Scholar]
- 18.Centers for Medicare & Medicaid Services. CY 2020 Medicare Hospital Outpatient Prospective Payment System and Ambulatory Surgical Center Payment System Final Rule (CMS-1717-FC). Accessed May 1, 2021. CY 2020 Medicare Hospital Outpatient Prospective Payment System and Ambulatory Surgical Center Payment System Final Rule (CMS-1717-FC)
- 19.The Lewin Group, Inc. CMS Comprehensive Care for Joint Replacement Model: Performance Year 4 Evaluation Report. September 2021. Available at https://innovation.cms.gov/data-and-reports/2021/cjr-py4-annual-report.