Effects of Maryland’s All-Payer Model on Elective Joint Replacement Surgery

ABSTRACT

Objective: To evaluate the Maryland All-Payer Model’s impact on the rate of elective major joint replacement surgery.

Study Design: A retrospective cohort study of patients in Maryland undergoing elective major joint replacement between 2011 and 2018 was performed using a 20% fee-for-service Medicare sample in a difference-in-difference framework with patients undergoing hip fracture repair serving as controls.

Methods: Among Maryland residents, there were 7147 Medicare fee-for-service patients undergoing elective major joint replacement and 1008 Medicare fee-for-service beneficiaries undergoing hip fracture repair. We used patient-level generalized linear models with a negative binomial family function and a log link function to estimate the association of the All-Payer Model with the rate of elective major joint replacement surgery.

Results: Under the All-Payer Model, the rate of elective major joint replacement surgery increased more than that of hip fracture repair (adjusted relative risk, 1.31; 95% CI, 1.15-1.51). Compared with hospitals without affiliates in adjacent states (Maryland-only hospitals), those with affiliates (Maryland hospitals with affiliates) saw rates of elective major joint replacement grow more slowly (adjusted relative risk, 0.87; 95% CI, 0.80-0.95) after the All-Payer Model. Furthermore, major joint replacement rates for Maryland residents at affiliated hospitals in adjacent states increased from 4.26 per 10,000 in the preintervention period to 5.23 per 10,000 in the postintervention period.

Conclusions: Under the All-Payer Model, population-based rates of elective major joint replacement surgery increased more rapidly than did rates of hip fracture repair. Although rates of major joint replacement at Maryland hospitals with affiliates grew more slowly than for Maryland-only hospitals, rates among Maryland residents increased at the affiliates in adjacent states.

Am J Manag Care. 2025;31(5):In Press

Takeaway Points

Maryland introduced the All-Payer Model in 2014 to improve quality and curb cost. The extent to which the policy may impair access to elective surgery or motivate multistate health care systems to shift procedures across states is unknown. We found the following:

• The All-Payer Model did not reduce the use of elective major joint replacements, allaying concerns of impaired access and supporting the policy’s potential expansion beyond Maryland.
• Rates of elective major joint replacements increased less at hospitals with affiliates in adjacent states.
• Although surgical rates significantly increased at out-of-state affiliates, the extent to which this increase represents intentional care shifting is unclear.

Elective hip and knee replacement, collectively referred to as major joint replacement surgery, aim to improve physical function and relieve pain.^1-4 Major joint replacement surgery is common, with more than 400,000 procedures annually in the US.^5,6 As spending exceeds $7 billion annually,⁷ major joint replacement surgery has been targeted by alternative payment models, whereby payment is linked to both quality and spending.^1,8 Such models attempt to disentangle, at least in part, the link between volume and revenue by creating incentives that improve quality and limit waste.

Launched in 2014, the Maryland All-Payer Model was implemented to do exactly that, exempting its hospitals from Medicare’s Inpatient and Outpatient Prospective Payment Systems. The All-Payer Model requires hospitals commit to improve quality (ie, reduce readmissions and hospital-acquired conditions) while establishing annual per capita spending growth caps on payments from all payers for inpatient and outpatient hospital-based services in Maryland.^9,10 Hospitals succeeding in achieving spending and quality benchmarks keep a portion of the savings. Those that fall short repay the difference to CMS and receive an additional financial penalty in the following year.¹¹ Prior work suggests that program spending by Maryland Medicare beneficiaries grew more slowly than for comparable nonresidents, yielding $975 million in savings to Medicare, a 48% reduction in potentially avoidable complications, and a 15.4% decrease in readmissions.^12,13 As a result, CMS has expressed its intent to expand the program more broadly.^14,15 Nonetheless, some worry that Maryland hospitals might try to meet benchmarks by avoiding expensive and/or high-volume elective services,⁷ such as major joint replacement surgery, by reducing access or shifting care to affiliated hospitals in adjacent states, falling out of the All-Payer Model’s purview.

We performed a retrospective cohort study of Maryland beneficiaries with traditional Medicare, assessing the impact of the All-Payer Model on population-based rates of elective major joint replacement surgery. Additionally, we assessed for heterogeneity in the effects of the policy at Maryland hospitals based on their ability to shift care to an immediately adjacent state (ie, hospitals in Maryland only vs hospitals with affiliates in adjacent states as part of a health system).

METHODS

We performed a retrospective cohort study of Maryland beneficiaries with traditional Medicare who underwent elective major joint replacement surgery between January 1, 2011, and December 31, 2018, using a 20% sample of fee-for-service Medicare claims. Although participation in the All-Payer Model was voluntary, all 46 eligible hospitals opted to participate. These participants were identified using the American Hospital Association annual survey.¹⁰ Ten small, rural hospitals previously piloting the model were excluded.^11,16 The Compendium of U.S. Health Systems¹⁷ was used to identify 21 hospitals without affiliates in immediately adjacent states (ie, Maryland-only hospitals) and 15 hospitals with affiliated hospitals immediately adjacent to the state as part of a health care system (ie, Maryland hospitals with affiliates).

For each year, we first used codes (eAppendix Table 1 [eAppendix available at ajmc.com]) from the International Classification of Diseases, Ninth Revision, Procedure Coding System (ICD-9-PCS) and the International Classification of Diseases, Tenth Revision, Procedure Coding System (ICD-10-PCS) to identify patients undergoing major joint replacement surgery. Among these surgical patients, those without ICD-9/ICD-10 hip fracture diagnosis codes listed in eAppendix Table 2 and with an inpatient admission type code of 3 (elective admission)¹⁸ were classified as having received elective major joint replacement. Conversely, patients with any ICD-9/ICD-10 hip fracture diagnosis codes listed in eAppendix Table 2 and without an elective inpatient admission code were classified into the hip fracture repair cohort. We further required that these patients were continuously enrolled in Medicare Parts A and B from 1 year before the procedure through 30 days after the discharge to allow for comorbidity measurement. Because our study is population based, we also identified all eligible beneficiaries residing in the state of Maryland enrolled in Medicare Parts A and B in each year.

Because all eligible Maryland hospitals joined the All-Payer Model on January 1, 2014 (ie, the implementation date of the All-Payer Model), there were 2 separate time periods for each patient: (1) a 3-year preintervention period from January 1, 2011, to December 31, 2013; and (2) a postintervention period extending from January 1, 2014, through December 31, 2018.

Outcomes

The primary outcome was the annual rate of major joint replacement surgery performed in all eligible Maryland beneficiaries. Specifically, the rate of elective major joint replacement was estimated using the annual number of procedures as the numerator and the number of all eligible Maryland residents as the denominator. To account for temporal trends in surgical procedures, we also assessed the annual rate of hip fracture repair using similar methods to serve as controls. Our rationale for using hip fracture repair as controls was based on the fact that rates vary relatively little between regions and largely reflect differences in disease incidence as opposed to surgeon discretion.¹⁹ Selecting the control group within the same state has additional benefits, as it helps minimize confounding caused by variations across health care delivery systems, particularly given that the All-Payer Model is built upon the nation’s only all-payer hospital rate regulation system.¹⁰ It also reduces potential biases resulting from other Medicare alternative payment models (such as the Comprehensive Care for Joint Replacement Model, a major joint replacement–specific Medicare Model), which are implemented in other states.⁷

Statistical Analysis

Comparisons of patient and market characteristics between those undergoing major joint surgery and hip fracture repair were conducted by t tests and χ² tests for continuous and for categorical variables, respectively. Next, the All-Payer Model’s impact on access was assessed by fitting a patient-level, generalized linear model with a negative binomial family function and a log link function to estimate the change in the rate of elective major joint replacement surgery before and after the start of the All-Payer Model on January 1, 2014. We used a difference-in-differences design in which the rate of hip fracture repair served as the control population. The key quantity of interest was the regression coefficient for the interaction term between indicators for the procedure cohort (major joint replacement vs hip fracture repair) and time (post- vs preintervention period). This estimate assesses the average differential change in the rate of major joint replacement surgery relative to the rate of hip fracture repair due to the launch of the All-Payer Model. The validity of the difference-in-differences approach was confirmed by testing the parallel trends assumption, comparing the slope over time in the preintervention period for the major joint replacement and hip fracture repair groups statistically and visually. The outcome did not violate the assumption (eAppendix Table 3 and eAppendix Figure). The models were adjusted for patient age, sex, race/ethnicity (non-Hispanic White, Black, and other, defined with the Research Triangle Institute Race Code), and Medicaid eligibility from the Medicare Beneficiary Summary File Base Segment.²⁰ The model was further adjusted for comorbidity, assessed using Hierarchical Condition Category codes (computed with demographic variables from the Medicare Beneficiary Files Base Segment and ICD-9/ICD-10 diagnostic data from the Medicare Provider Analysis and Review, Outpatient, and Carrier files) enumerated for the 12-month calendar year preceding each index procedure.²¹ Additionally, we adjusted the model for supply-side variables measured at the county level obtained from the Area Health Resources Files, including the numbers of long-term care hospitals, skilled nursing facilities, home health agencies, and acute care hospital beds; per capita income; and the unemployment rate. Secular trends in utilization were captured by including a variable for year.

We explored whether changes in rates of elective major joint replacement surgery implied by the All-Payer Model varied with a hospital’s ability to shift its volume to a financially integrated entity in a state adjacent to Maryland. Using an approach similar to that outlined earlier, we fitted a separate patient-level model. In this model, however, the coefficient of interest was the interaction term between hospital type (Maryland-only hospital vs Maryland hospital with affiliate) and time (post- vs preintervention period). This term provides an estimation of the differential change in the rate of elective major joint replacement surgery performed at hospitals only in Maryland compared with those with affiliates in an adjacent state related to the implementation of the All-Payer Model. As an exploratory analysis, we further assessed rates of major joint replacement surgery among Maryland residents undergoing the procedure in adjacent states (ie, Delaware, District of Columbia, Pennsylvania, Virginia, West Virginia) by Maryland hospitals with affiliates before and after All-Payer Model implementation.

Sensitivity Analysis

We performed several sensitivity analyses to gauge potential sources of bias. First, we examined the robustness of our findings to 2 different preintervention periods (January 1, 2010-December 31, 2013; and January 1, 2010-December 31, 2012). Second, although the All-Payer Model was implemented on January 1, 2014, hospitals spent several months finalizing their contracts with Maryland. Therefore, we performed an analysis in which the year 2014 was considered as a washout period. Third, we conducted falsification tests treating 2013 as a postintervention year.

All analyses were performed using Stata 16 (StataCorp LLC) and SAS 9.4 (SAS Institute Inc) and reported using heteroskedasticity-robust SEs.²² A 2-tailed test with a P value of less than .05 was considered statistically significant. This study was deemed exempt by the University of MichiganInstitutional Review Board.

RESULTS

In this population-based analysis of 315,205 eligible beneficiary-years, we identified 375 unique patients undergoing hip fracture repair and 2173 unique patients undergoing major joint replacement surgery in the preintervention period (January 1, 2011-December 31, 2013), as shown in Table 1. Compared with those undergoing hip fracture repair, elective major joint replacement patients were younger (mean age, 74.7 vs 83.5 years; P < .001) but less likely to be of White race (80.0% vs 84.8%). Notably, those undergoing hip fracture repair had more comorbidity.

After adjusting for patient and market characteristics, rates of elective major joint replacement surgery increased significantly compared with rates of hip fracture repair in the postintervention period relative to the preintervention period (adjusted relative risk, 1.31; 95% CI, 1.15-1.51; P < .001), as shown in Table 2. As illustrated in Table 3, the increase in rates of elective major joint replacement surgery among Maryland hospitals with affiliates in adjacent states was lower than for Maryland-only hospitals (adjusted relative risk, 0.87; 95% CI, 0.80-0.95; P = .002) after implementation of the All-Payer Model. Notably, rates of major joint replacement surgery in Maryland residents at affiliated hospitals in adjacent states increased from 4.26 per 10,000 in the preintervention period to 5.23 per 10,000 in the postintervention period (P = .013).

Sensitivity Analyses

Table 4 shows the findings from the sensitivity analyses, supporting the robustness of the findings of the main identification strategy. Both alternative definitions of the preintervention continued to support the effect of the All-Payer Model on the growth of major joint replacement surgery relative to hip fracture repair. Further, specifying 2014 (ie, the initial year of All-Payer Model implementation) as a washout period had no effect on the magnitude and significance of the effect of the policy. Finally, changing the postintervention period to 2013 as a falsification test decreased the size of effect of the All-Payer Model, which was not statistically significant.

DISCUSSION

In this population-based analysis of Maryland beneficiaries with Traditional Medicare who underwent elective major joint replacement surgery, we found that procedure rates increased significantly relative to hip fracture repair after the implementation of the All-Payer Model. However, rates of major joint replacement surgery increased less rapidly in hospitals with affiliates in adjacent states relative to those without affiliates. Following implementation of the All-Payer Model, rates of major joint replacement surgery among Maryland residents increased significantly at these affiliated hospitals in adjacent states.

That the All-Payer Model did not reduce the growth of elective surgery relative to hip fracture repair allays concerns that such policies might lead to rationing and/or impair access to common surgical services for Medicare beneficiaries. This aligns with national trends that show an increase in the use of elective major joint replacements.^23,24 The policy is primarily geared to affect high-volume/high-cost conditions, such as elective major joint replacement surgery, as they afford the greatest opportunity for savings through clinical redesign efforts.²⁵ Indeed, Maryland hospitals have made substantial investment in this area,¹² bolstering their capacity to provide care for these patients while meeting quality and spending benchmarks. Our findings, combined with prior research demonstrating reductions in hospital costs and avoidable complications in this patient population,²⁶ support the promise of broadening the scope of the policy beyond the state of Maryland.

Notably, the number of elective major joint replacement surgeries grew less rapidly at hospitals with affiliated hospitals in adjacent states. As these hospitals are not subject to the All-Payer Model, Maryland residents managed outside the state have no effect on the budgets of participating hospitals, creating the possibility of intentional care shifting. In this arrangement, Maryland hospitals with affiliates nearby might transfer care of high-cost, elective surgery patients (eg, predisposed to complications) to their out-of-state, affiliated hospitals to circumvent the purview of the policy. To investigate this possibility, we compared the rates of elective major joint replacement surgery in Maryland residents receiving care in hospitals affiliated with Maryland hospitals in adjacent states. The data indicate that rates of surgery among Maryland residents increased significantly in affiliates outside the state after the implementation of the All-Payer Model, supporting the possibility of care shifting, although the intentionality of this finding is unclear.

Limitations

Findings from this study should be interpreted in the context of its limitations. First, the analysis focuses on elective major joint replacement surgery in traditional Medicare beneficiaries, whereas the All-Payer Model targets all patients with various conditions managed in Maryland hospitals. Consequently, the generalizability of our conclusions to Medicare patients in different clinical contexts or non-Medicare patients undergoing elective joint replacements is limited. However, major joint replacements are one of the most common procedures in the Traditional Medicare population, which supports the significance of our findings for policy. Moreover, our findings, derived from evaluating a statewide alternative payment model exclusively implemented in Maryland, may not be generalizable to other states due to significant geographic variations in health care services across states. Second, our estimates may be biased by changes in the traditional Medicare population, the introduction of some Affordable Care Act provisions and alternative payment models (such as the Bundled Payments for Care Improvement Advanced Model that began on October 1, 2018, and targeted elective major joint replacement and hip fracture repair), and other factors influencing the demand for major joint replacement surgery. Nonetheless, we linked Maryland hospital data to accountable care organization and Bundled Payments for Care Improvement Initiative (BPCI) data but found no matches, indicating that no Maryland hospitals participated in the Shared Savings Program or the BPCI program. Additionally, our empirical models adjusted for relevant patient and county-level characteristics, and our findings were robust to multiple sensitivity analyses. Third, this study examined the impact of Maryland’s All-Payer Model on the use of elective major joint replacement using data for only Maryland residents. Future work is needed to compare the rates of elective major joint replacement between Maryland and other states before and after the implementation of Maryland’s All-Payer Model to understand the Model’s impact relative to other states, given the nationwide increase in elective major joint replacement.

CONCLUSIONS

The All-Payer Model did not curtail the use of elective major joint replacement surgery relative to hip fracture repair, allaying concerns of rationing and/or impaired access. The effect was more pronounced in hospitals without affiliates in the adjacent states. As the Maryland All-Payer Model has achieved cost reductions and care improvements¹² and supported hospital financial viability during the COVID-19 pandemic,²⁷ future research should continue to evaluate its effects on quality, cost, and access as CMS considers expanding its scope beyond the state.

Author Affiliations: Department of Foundations of Medicine, New York University Langone Health (MY), Mineola, NY; Department of Urology (AS, VBS, BKH) and Department of Health Management and Policy (RAH), University of Michigan, Ann Arbor, MI; Department of Urology, University of Florida (JMH), Gainesville, FL.

Source of Funding: This study was funded by grants from the National Institute on Aging (R01AG068074) and the Office of the Assistant Secretary of Defense for Health Affairs through the Prostate Cancer Research Program under Award No. HT9425-23-1-0141.

Author Disclosures: Dr Ying reports receiving funding from the Department of Defense. Dr Hollenbeck reports receiving funding from the National Institute on Aging. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (MY, JMH, VBS, RAH, BKH); acquisition of data (MY, JMH, BKH); analysis and interpretation of data (MY, AS, VBS, RAH); drafting of the manuscript (MY, AS, RAH); critical revision of the manuscript for important intellectual content (MY, AS, JMH, VBS, RAH, BKH); statistical analysis (MY, AS); obtaining funding (MY, BKH); administrative, technical, or logistic support (BKH); and supervision (BKH).

Address Correspondence to: Meiling Ying, PhD, Department of Foundations of Medicine, New York University Langone Health, 101 Mineola Blvd, Mineola, NY 11501. Email: meiling.ying@nyulangone.org.

REFERENCES

1. 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;59(2):101-110. doi:10.1097/mlr.0000000000001473

2. Cram P, Lu X, Kaboli PJ, et al. Clinical characteristics and outcomes of Medicare patients undergoing total hip arthroplasty, 1991-2008. JAMA. 2011;305(15):1560-1567. doi:10.1001/jama.2011.478

3. Cram P, Lu X, Kates SL, Singh JA, Li Y, Wolf BR. Total knee arthroplasty volume, utilization, and outcomes among Medicare beneficiaries, 1991-2010. JAMA. 2012;308(12):1227-1236. doi:10.1001/2012.jama.11153

4. Ethgen O, Bruyère O, Richy F, Dardennes C, Reginster JY. Health-related quality of life in total hip and total knee arthroplasty: a qualitative and systematic review of the literature. J Bone Joint Surg Am. 2004;86(5):963-974. doi:10.2106/00004623-200405000-00012

5. Healthcare Cost and Utilization Project (HCUPnet). Agency for Healthcare Research and Quality. Accessed November 11, 2022. https://datatools.ahrq.gov/hcupnet/#setup

6. Thirukumaran CP, Rosenthal MB. The Triple Aim for payment reform in joint replacement surgery: quality, spending, and disparity reduction. JAMA. 2021;326(6):477-478. doi:10.1001/jama.2021.12070

7. Comprehensive Care for Joint Replacement Model. CMS. Accessed November 11, 2022. https://www.cms.gov/priorities/innovation/innovation-models/cjr

8. Dummit LA, Kahvecioglu D, Marrufo G, et al. Association between hospital participation in a Medicare bundled payment initiative and payments and quality outcomes for lower extremity joint replacement episodes. JAMA. 2016;316(12):1267-1278. doi:10.1001/jama.2016.12717

9. Maryland All-Payer Model Agreement. Maryland Health Care Commission. Accessed November 11, 2022. https://mhcc.maryland.gov/mhcc/pages/hcfs/hcfs_hospital/documents/chcf_all_payer_model_agreement.pdf

10. Maryland All-Payer Model. CMS. Accessed February 11, 2022. https://www.cms.gov/priorities/innovation/innovation-models/maryland-all-payer-model

11. Roberts ET, Hatfield LA, McWilliams JM, et al. Changes in hospital utilization three years into Maryland’s global budget program for rural hospitals. Health Aff (Millwood). 2018;37(4):644-653. doi:10.1377/hlthaff.2018.0112

12. RTI International. Evaluation of the Maryland All-Payer Model. CMS. November 2019. Accessed June 10, 2023. https://downloads.cms.gov/files/md-allpayer-finalevalrpt.pdf

13. Monitoring of Maryland’s New All-Payer Model: Biannual Report. Maryland Health Services Cost Review Commission. April 2019. Accessed November 11, 2022. https://hscrc.maryland.gov/Documents/legal-legislative/reports/2019%20Reports/April%202019%20Biannual%20Report%20FINAL.pdf

14. Request for information on concepts for regional multi-payer prospective budgets. CMS. Accessed November 11, 2022. https://www.cms.gov/priorities/innovation/files/x/regprosbudgets-rfi.pdf

15. Regional budget payment concept. CMS. Accessed May 11, 2022. https://www.cms.gov/priorities/innovation/innovation-models/regional-budget-payment

16. Mortensen K, Perman C, Chen J. Innovative payment mechanisms in Maryland hospitals: an empirical analysis of readmissions under total patient revenue. Healthc (Amst). 2014;2(3):177-183. doi:10.1016/j.hjdsi.2014.03.002

17. Compendium of U.S. Health Systems. Agency for Healthcare Research and Quality. Accessed January 28, 2023. https://www.ahrq.gov/chsp/data-resources/compendium.html

18. Inpatient admission type code. Research Data Assistance Center. Accessed September 3, 2024. https://resdac.org/cms-data/variables/inpatient-admission-type-code

19. Birkmeyer JD, Sharp SM, Finlayson SR, Fisher ES, Wennberg JE. Variation profiles of common surgical procedures. Surgery. 1998;124(5):917-923. doi:10.1016/S0039-6060(98)70017-0

20. Ying M, Thirukumaran CP, Temkin-Greener H, Joynt Maddox KE, Holloway RG, Li Y. Association of skilled nursing facility participation in voluntary bundled payments with postacute care outcomes for joint replacement. Med Care. 2023;61(2):109-116. doi:10.1097/mlr.0000000000001799

21. Pope GC, Kautter J, Ellis RP, et al. Risk adjustment of Medicare capitation payments using the CMS-HCC model. Health Care Financ Rev. 2004;25(4):119-141.

22. Wooldridge JM. Econometric Analysis of Cross Section and Panel Data. 2nd ed. MIT Press; 2010.

23. Shichman I, Roof M, Askew N, et al. Projections and epidemiology of primary hip and knee arthroplasty in Medicare patients to 2040-2060. JB JS Open Access. 2023;8(1):e22.00112. doi:10.2106/JBJS.OA.22.00112

24. Thirukumaran CP, Cai X, Glance LG, et al. Geographic variation and disparities in total joint replacement use for Medicare beneficiaries: 2009 to 2017. J Bone Joint Surg Am. 2020;102(24):2120-2128. doi:10.2106/jbjs.20.00246

25. Offodile AC II, Lin YL, Melamed A, Rauh-Hain JA, Kinzer D, Keating NL. Association of Maryland global budget revenue with spending and outcomes related to surgical care for Medicare beneficiaries with cancer. JAMA Surg. 2022;157(6):e220135. doi:10.1001/jamasurg.2022.0135

26. Aliu O, Lee AWP, Efron JE, Higgins RSD, Butler CE, Offodile AC II. Assessment of costs and care quality associated with major surgical procedures after implementation of Maryland’s capitated budget model. JAMA Netw Open. 2021;4(9):e2126619. doi:10.1001/jamanetworkopen.2021.26619

27. Levy JF, Ippolito BN, Jain A. Hospital revenue under Maryland’s total cost of care model during the COVID-19 pandemic, March-July 2020. JAMA. 2021;325(4):398-400. doi:10.1001/jama.2020.22149