- Email: [email protected]
Do Medicare Accountable Care Organizations Reduce Disparities After Spinal Fracture?
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Do Medicare Accountable Care Organizations Reduce Disparities After Spinal Fracture? Shaina A. Lipa, MD,a Daniel J. Sturgeon, MS,b Justin A. Blucher, MS,a Mitchel B. Harris, MD,c and Andrew J. Schoenfeld, MD, MSca,b,* a
Department of Orthopaedic Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts b Department of Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts c Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts
article info
abstract
Article history:
Background: National changes in health care disparities within the setting of trauma care
Received 5 April 2019
have not been examined within Accountable Care Organizations (ACOs) or non-ACOs. We
Received in revised form
sought to examine the impact of ACOs on post-treatment outcomes (in-hospital mortality,
19 August 2019
90-day complications, and readmissions), as well as surgical intervention among whites
Accepted 5 September 2019
and nonwhites treated for spinal fractures.
Available online xxx
Materials and methods: We identified all beneficiaries treated for spinal fractures between 2009 and 2014 using national Medicare fee for service claims data. Claims were used to
Keywords:
identify sociodemographic and clinical criteria, receipt of surgery and in-hospital mortal-
Accountable care organziations
ity, 90-day complications, and readmissions. Multivariable logistic regression analysis ac-
Surgical disparities
counting for all confounders was used to determine the effect of race/ethnicity on
Trauma care
outcomes. Nonwhites were compared with whites treated in non-ACOs between 2009 and
Spinal fractures
2011 as the referent.
African Americans
Results: We identified 245,704 patients who were treated for spinal fractures. Two percent of the cohort received care in an ACO, whereas 7% were nonwhite. We found that disparities in the use of surgical fixation for spinal fractures were present in non-ACOs over the period 2009-2014 but did not exist in the context of care provided through ACOs (odds ratio [OR] 0.75; 95% confidence interval [CI] 0.44, 1.28). A disparity in the development of complications existed for nonwhites in non-ACOs (OR 1.09; 95% CI 1.01, 1.17) that was not encountered among nonwhites receiving care in ACOs (OR 1.32; 95% CI 0.90, 1.95). An existing disparity in readmission rates for nonwhites in ACOs over 2009-2011 (OR 1.34; 95% CI 1.01, 1.80) was eliminated in the period 2012-2014 (OR 0.85; 95% CI 0.65, 1.09). Conclusions: Our work reinforces the idea that ACOs could improve health care disparities among nonwhites. There is also the potential that as ACOs become more familiar with care integration and streamlined delivery of services, further improvements in disparities could be realized. ª 2019 Elsevier Inc. All rights reserved.
* Corresponding author. Department of Orthopaedic Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115. Tel.: þ330 329-2594; fax: þ617 278-6919. E-mail address: [email protected] (A.J. Schoenfeld). 0022-4804/$ e see front matter ª 2019 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jss.2019.09.003
124
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
Introduction Health care disparities after orthopedic and spine surgical interventions have been widely reported for racial and ethnic minorities in the past.1-8 Documented disparities in the orthopedic literature include reduced access to surgical services, health care segregation, higher rates of perioperative complications, and inferior outcomes in several respects.2,3,5,9e12 Recent health care reform efforts, such as Accountable Care Organizations (ACOs), have been proposed as initiatives that would not only reduce costs but also lead to reductions in health care disparities.13e19 Through incentives and accountability, it was believed that ACOs would lead the U.S. health care system to realize the “the triple aim”: better population health, better patient experience, and lower costs. Via actuation of these mechanisms, one of the highly touted benefits of ACO implementation included amelioration of health care disparities.15,16 In the past, however, other health reform efforts have been found to unintentionally aggravate existing disparities.20e22 Thus far, studies have shown mixed results regarding the impact of ACOs on surgical health care,17,18,23 including disparities in care among disadvantaged patient populations.13,19,24 For example, Schoenfeld et al. documented persistent differences in the rates of elective surgery, including hip replacement and lumbar spine procedures, among minorities treated in ACOs.13 In the past, surgical interventions after traumatic injury have been considered differently from elective procedures, given the unique features of trauma care in the United States, including delineation of services within the American College of Surgeons tiered trauma system.3 As far as we are aware, national changes in health care disparities within the setting of trauma care have not been examined for ACOs or non-ACOs in the current health reform era. In this context, we sought to examine the impact of Medicare ACO formation on post-treatment outcomes including in-hospital mortality, 90-day complications, and 90day readmissions for white and nonwhite beneficiaries treated for spinal fractures. Secondarily, we also evaluated the influence of ACOs on rates of surgical fixation in these populations. Spinal fractures were selected as a suitable clinical condition for the study, given previous publications documenting existing disparities in treatment and relatively high rates of postinjury morbidity and mortality.9,25e27 Based on prior literature, we hypothesized that ACO formation would not lead to meaningful improvements in health care disparities around treatment for spinal fractures.
Patients and methods We performed this study using national Medicare fee for service claims data for the years 2009-2014, obtained from the Centers for Medicare and Medicaid Services (CMS). We queried Medicare Provider Analysis and Review records to identify beneficiaries enrolled in part A and B plans at the time of treatment for a cervical, thoracic, or lumbar fracture (identified by International Classification of Disease-ninth revision (ICD-9) code 9. Patients younger than 65 y at the time of injury and those insured through Medicare-managed care plans were excluded from this analysis.
We abstracted linked CMS files for all beneficiaries meeting inclusion criteria to determine age at the time of treatment, biologic sex, and reported race/ethnicity. Race/ethnicity was recorded by the CMS as white, black (e.g., African-American), Hispanic, Asian, Native American, and other and dichotomized here as white or nonwhite because of the small numbers of nonwhite patients treated in ACOs. Clinical data obtained from CMS files included age-adjusted Charlson Comorbidity Score,28 location of spinal injury, and whether surgery was performed (defined using an ICD-9 coding algorithm available from the authors by request). Eligible surgical procedures consisted of decompression and/or instrumented fusion procedures. Patients who received stand-alone cement augmentation (e.g., kyphoplasty) performed without decompression or instrumented fusion were considered in the nonoperative cohort as these procedure can often be performed in an outpatient setting, do not require spinal instrumentation, and are not associated with substantial increases in health care costs. This approach to spine procedural classification is aligned with prior research in health care policy.29 Hospital-level data that were surveyed consisted of teaching status of the hospital where the patient was treated and hospital participation in an ACO. As outlined in prior investigations, we delineated ACO affiliation by matching the claims associated with fracture care to each of the Medicare Pioneer ACOs and Shared Savings Program participants identified by the CMS in ACO-specific files for the years 2012, 2013, and 2014.13,23 Beneficiaries treated in organizations that would form an ACO in the years 2012-2014 were considered as a single group in this investigation, with all other individuals assigned to the non-ACO cohort. Claims and beneficiary-level files were used to identify the outcomes of interest including in-hospital mortality, 90-day complications, and readmissions. Complications were defined in line with prior research using an ICD-9 coding system that captures major postsurgical events including surgical site infections, sepsis, shock, cardiopulmonary complications, neurologic events, renal complications, deep venous thrombosis, and pulmonary embolism.8 Readmissions were defined as an admission to the hospital for any cause within 90 d of discharge from the index admission associated with the spinal fracture.
Statistical testing The primary outcome in this analysis was in-hospital mortality after treatment for a spinal fracture with secondary consideration of 90-day complication and readmission events. Temporal changes in the use of surgical intervention for spinal fractures among whites and nonwhites were also evaluated. Patient race/ ethnicity (white versus nonwhite) was considered the primary predictor within the context of the environment of care (ACO versus non-ACO). Patient age, biologic sex, age-adjusted comorbidities, hospital teaching status, and the year of treatment were used as covariates based on our study’s conceptual model. Treatment year was classified as 2009-2011 (prehealth reform) and 2012-2014 (posthealth reform) based on the passage of the Affordable Care Act (2012).13,23 Beneficiaries treated in 2009-2011 by health care entities that would eventually organize as an ACO were evaluated as a single cohort in the prehealth reform period.
125
lipa et al spinal fracture disparities in acos
Table e Sociodemographic and clinical characteristics of all patients and nonwhite patients treated for spinal fractures in an Accountable Care Organization or non-ACO environment between 2009 and 2014.* Characteristic
All patients (n ¼ 245, 704) 2009-2011
P-value
2012-2014
Nonwhite patients (n ¼ 17,619) 2009-2011
P-value
2012-2014
Age (mean, SD)
80.7 (8.1)
80.8 (8.1)
0.02
79.2 (8.6)
79.7 (8.4)
<0.001
Female sex (%)
87,833 (71)
85,009 (70)
<0.001
6095 (70)
6202 (69)
0.17
Teaching hospital (%)
21,334 (17)
21,110 (17)
0.19
2000 (23)
2015 (23)
0.40
2364 (2)
2557 (2)
<0.001
146 (2)
207 (2)
0.003
11,076 (9)
10,084 (8)
<0.001
747 (9)
698 (8)
0.05
<0.001
5.0 (2.0)
5.0 (3.0)
ACO (%) Surgery (%) Age-adjusted Charlson Comorbidity Score (median, IQR)
5.0 (2.0)
5.0 (2.0)
<0.001
ACO ¼ Accountable Care Organization; SD ¼ standard deviation; IQR ¼ interquartile range. * Numbers are rounded.
Multivariable logistic regression analysis accounting for all confounders was used to determine the independent effect of patient race/ethnicity on the outcomes of interest. In all regression tests, nonwhites were compared with whites treated in non-ACOs between 2009 and 2011 as the referent. Disparities were considered to be present if the experience of nonwhites in ACOs was significantly different from that of whites treated in ACOs and/or whites receiving care in nonACOs in the periods under study. A priori, we defined significance for variables that demonstrated P < 0.05 and odds ratio (OR) with 95% confidence intervals (CIs) exclusive of 1.0 following multivariable testing. All analyses were performed using SAS v9.4 (SAS Institute; Cary, NC). Before initiation, this research received a determination of exempt status from our institutional review board and was also approved by the CMS.
Results In the period under study, we identified 245,704 patients who were treated for spinal fractures. Seventy percent of the cohort was female, and the median comorbidity score was 5.0 (interquartile range, 2.0). Overall, two percent of the cohort received care in an ACO (4921/245,704). Seven percent of the population (17,619/245,704) were nonwhite and, in parallel, two percent (353/17,619) of nonwhite patients were treated in ACOs. In 2012-2014, as compared with 2009-2011, the population of nonwhite patients was older, had a slightly larger number of comorbidities, and were marginally more likely to be treated in an ACO (Table). Given the size of our sample, all of these findings were significant, although absolute differences were small. The findings for nonwhite patients mirrored those of the general population, overall (Table). Across the entire cohort, the in-hospital mortality rate was 5% (n ¼ 11,138), whereas 90-day complications occurred in 11% (n ¼ 26,119) and readmissions were documented in 51% (n ¼ 124,726).
and 2012-2014. A comparable reduction in the rates of surgical fixation was appreciated in the nonwhite population as well, although this finding did not meet the threshold for significance (P ¼ 0.05). At baseline, nonwhites in non-ACOs had a significantly lower likelihood of receiving surgical intervention for a spinal fracture (OR 0.83; 95% CI 0.75, 0.91; Fig. 1) and this finding persisted in 2012-2014 (OR 0.79; 95% CI 0.72, 0.86). There was no significant difference in the use of surgery among nonwhites treated in ACOs during 2009-2011 (OR 0.89; 95% CI 0.46, 1.74) and no substantive change was encountered for this outcome over the course of 2012-2014 (OR 0.75; 95% CI 0.44, 1.28).
Changes in in-hospital mortality The mortality rate of the referent group (whites treated in non-ACOs) was 3.9% in 2009-2011. Nonwhites treated in nonACOs (mortality rate of 4.8%) had higher odds of in-hospital mortality during 2009-2011 (OR 1.25; 95% CI 1.12, 1.39; Fig. 2) and 2012-2014 (mortality rate of 6.3%; OR 1.68; 95% CI 1.53, 1.85). A significant difference in mortality was also encountered among nonwhites treated in ACOs in 2012-2014 (mortality rate of 6.3%; OR 1.65; 95% CI 1.04, 2.62). This trend, however, mirrored findings for whites treated in ACOs (mortality rate of 5.2%; OR 1.35; 95% CI 1.07, 1.70) and non-ACOs (mortality rate of 5.0%; OR 1.32; 95% CI 1.27, 1.37) over the same time period.
Changes in complications In 2009-2011, nonwhites treated in non-ACOs (OR 1.01; 95% CI 0.93, 1.09) and ACOs (OR 0.81; 95% CI 0.42, 1.58) did not demonstrate significant differences in the rate of 90-day complications as compared with whites (Fig. 3). In 20122014, however, the odds of complications among nonwhites treated in non-ACOs was significantly increased (OR 1.09; 95% CI 1.01, 1.17), whereas no such change was apparent for nonwhites in ACOs (OR 1.32; 95% CI 0.90, 1.95).
Changes in the use of surgical intervention for spinal fracture
Changes in readmissions
For the cohort as a whole, the use of surgical intervention for spinal fractures decreased from 9% to 8% between 2009-2011
Nonwhites treated in non-ACOs had lower odds of readmissions as compared with whites in both 2009-2011 (OR 0.92;
126
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
Fig. 1 e The adjusted odds of surgical intervention for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
95% CI 0.88, 0.97; Fig. 4) and 2012-2014 (OR 0.93; 95% CI 0.88, 0.97). Nonwhites in ACOs had significantly higher odds of readmission in 2009-2011 (OR 1.34; 95% CI 1.01, 1.80) which became insignificant over the course of the years 2012-2014 (OR 0.85; 95% CI 0.65, 1.09).
Discussion At the time of the implementation of the ACA, ACOs represented a cornerstone of the initiatives designed to reduce
Fig. 2 e The adjusted odds of in-hospital mortality for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
lipa et al spinal fracture disparities in acos
127
Fig. 3 e The adjusted odds of 90-day complications after treatment for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
health care expenditures while simultaneously improving efficiency and quality.14,16e19,24,29 The standardization of practice, integration of organizations and care pathways, and an emphasis on access and preventative care were advertised
as mechanisms that would not only lead to institutional success in the setting of accountable care but also would have a spillover effect in terms of reducing health care disparities.13,16 If changes in health care delivery associated with
Fig. 4 e The adjusted odds of 90-day readmissions after treatment for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
128
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
ACO implementation were to disproportionately impact minorities, however, or if disadvantaged populations were restricted to low-performing ACOs, these health reform efforts could culminate in an exacerbation of disparities. Such phenomena had previously been encountered among other well-intentioned CMS reform efforts, including managed care plans, pay for performance, and the centers of excellence program.20e22 Although early studies have indicated some success for ACOs in terms of reducing health care costs and moderating the use of preference sensitive surgeries,17,18,29 demonstrable improvements in health care quality around spine surgery have not been identified. Moreover, Schoenfeld et al. reported that the formation of ACOs did not lead to significant reductions in disparities regarding access to surgical care for racial and ethnic minorities, although this study largely involved elective interventions and was not restricted to spinal disorders.13 We sought to examine changes in the type and quality of care for spinal fractures among whites and nonwhites treated in ACOs and non-ACOs over a 6-year period encompassing the implementation of health care reform under the Affordable Care Act. Our reliance on Medicare claims data allows surveillance of patients across different health care systems and throughout various regions of the United States. Our methodology, including the means through which we delineate ACO enrollment and identify mortality, complication, and readmission events follow practices previously established in the literature.13,23,29 The rates of perioperative morbidity and mortality identified here are in line with prior reports for a Medicare age population, taking into account the size of the sample and events captured under our umbrella designation of complications.9 We found that disparities in the use of surgical fixation for spinal fractures were present in nonACOs over the period 2009-2014 but did not exist in the context of care provided through ACOs. Although, given the size of our sample, the point estimates for the two time periods are different (e.g., OR 0.89 in 2009-2011 versus 0.75 in 2012-2014), the overlapping 95% CIs for the two time frames (Fig. 1) provides reassurance of no measurable difference. Likewise, a disparity in the development of complications existed for nonwhites in non-ACOs that was not encountered among nonwhites receiving care in ACOs. Perhaps more importantly, an existing disparity in readmission rates for nonwhites in ACOs over 2009-2011 was eliminated in the period 2012-2014. Although elevations in in-hospital mortality across all cohorts in the period 2012-2014 are concerning for separate reasons, these trends were present in white and nonwhite beneficiaries in ACOs and non-ACOs and therefore do not meet our definition for a health care disparity. Our findings are similar to those previously expostulated for changes in cervical fracture care around health reform in Massachusetts.25 In this work, Schoenfeld et al documented no significant differences in the use of surgical interventions for cervical fractures before and after the implementation of health reform initiatives.25 In that effort, however, disparities among nonwhite patients persisted in terms of increased mortality, complications, and hospital length of stay. This was not the case in the present work, where the overall trend appeared to be one of the improvements with respect to
extant disparities in care, particularly the development of 90day complications and readmissions. Importantly, Massachusetts health reform efforts focused primarily on insurance expansion and not on promulgation of accountable care or the use of risk-based reimbursement.25,27 In addition, the results of the present study contrast markedly with glaring disparities in spine fracture care previously published from research using the National Trauma Databank.9 Viewed in this light then, our findings suggest promise for ACOs with respect to ameliorating disparities in spine fracture care. Standardization of care and integrated care delivery pathways occurring frequently in the setting of ACOs may eliminate some of the variations in clinical care that have previously potentiated health care disparities.2,3,30 Nonetheless, there remains room for improvement, especially as the rates of in-hospital mortality remain higher among whites and nonwhites treated in ACOs. As prior work showed an increased likelihood of complications and readmissions among patients receiving lumbar spine surgery in ACOs,23 a more holistic assessment of the reasons similar trends are not apparent in the context of fracture care is needed. Furthermore, although this analysis considered ACOs as a single group, there are invariably high and low performers within this category. As prior efforts outside the field of spine surgery have raised concerns for health care segregation in ACOs,14,24 where minorities are disproportionately served by lower performing and underfunded health care entities, a more granular assessment of the racial and ethnic composition of highand low-performing ACOs is warranted, along with determination of disparities within and between organizations. Evaluating the institutional culture of ACOs is also imperative, as recent work suggests that improved access and an ethic emphasizing equal care are both necessary for the elimination of disparities.8 As such characterizations are not possible, given our reliance on Medicare claims data, we recognize the potential for an ecological bias to confound our results. Additional limitations stem from the retrospective study design and the use of claims data collected for the purposes of payment and not designed for research. As a result, errors in coding at the provider level or during CMS data preparation could influence our determinations, including ACO attribution, and these cannot be adequately characterized or addressed methodologically. We are reassured to some extent in this matter, however, by the fact that the percentage of beneficiaries treated in ACO and non-ACOs mirrors estimates encountered in other works.13,23,29 Organizations electing to form ACOs may be fundamentally different from health care entities that did not join the Pioneer program or Medicare Shared Savings Program, and it is possible that our regression models are unable to fully account for this using the variables available through health care claims. Thus, there is the potential for residual confounding to influence the statistical models we used. Furthermore, because our study period ends only a few years beyond the implementation of the ACA, it is conceivable that ACOs may become more effective in mitigating disparities or their impact may be more discernible in the future. Given our reliance on claims data, we are not able to characterize the extent of spinal fractures, rationale for surgical intervention, or the factors considered in decision-
lipa et al spinal fracture disparities in acos
making for surgery. The appropriate use of surgical interventions similarly cannot be defined. We therefore recognize the possibility that selection and indication bias confound the results of our work. Further testing is necessary to validate the findings presented here and our determinations should not be considered translatable to other aspects of spine surgical care, trauma care, or the care of patients not insured through Medicare. Likewise, our category of nonwhite reflects an amalgam of racial and ethnic groups. This was necessary because of the small number of specific minority subgroups treated within ACOs in the study cohort. As a result, we recognize that our conclusions might not be applicable to any single class of nonwhite patients.2 Additional research involving the experiences of specific minority populations (e.g., African-Americans, Hispanics, Asians, and so forth) in ACOs is needed. Notwithstanding these limitations, this study presents important information for patients, providers, hospital systems, and policymakers. Our work reinforces the idealized concept that ACOs could help improve health care disparities for disadvantaged populations. There is also the potential that, as ACOs become more familiar with care integration and streamlined delivery of services, further improvements in disparities might be realized. This process could be aided, or even accelerated, were additional incentives placed on ACO performance in terms of providing equitable care or eliminating disparities. Ensuring adequate funding for those ACOs that serve a disproportionate number of minority patients also represents an important measure to ensure continued progress along the lines documented here. As we recognize that our determinations represent a single time-point in the course of ACO development, we believe that continued surveillance and cautious optimism regarding future benefits are supported at this time.
Acknowledgment Authors’ contributions: A.J.S., S.A.L., and D.J.S. had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. S.A.L., A.J.S., D.J.S., and M.B.H. contributed to study concept and design. D.J.S. and A.J.S. contributed to acquisition of data; S.A.L., A.J.S., J.A.B., D.J.S., and M.B.H. contributed to analysis and interpretation of data; S.A.L., J.A.B., D.J.S., and A.J.S. contributed to drafting of the manuscript; A.J.S. and M.B.H. contributed to critical revision of the manuscript for important intellectual content. A.J.S. and D.J.S. contributed to statistical analysis; A.J.S. obtained funding; S.A.L. and J.A.B. provided administrative, technical, or material support; A.J.S. and M.B.H. supervised the study.
Disclosure The research in this article was supported by the Centers for Medicare and Medicaid Services Office of Minority HealtheHealth Equity Data Access Program. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Centers for
129
Medicare and Medicaid Services or US Department of Health and Human Services.
references
1. Haider AH, Scott VK, Rehman KA, et al. Racial disparities in surgical care and outcomes in the United States: a comprehensive review of patient, provider, and systemic factors. J Am Coll Surg. 2013;216:482e492. 2. Schoenfeld AJ, Tipirneni R, Nelson JH, Carpenter JE, Iwashyna TJ. The influence of race and ethnicity on complications and mortality after orthopedic surgery: a systematic review of the literature. Med Care. 2014;52:842e851. 3. Pandya NK, Wustrack R, Metz L, Ward D. Current concepts in orthopaedic care disparities. J Am Acad Orthop Surg. 2018;26:828e832. 4. Schoenfeld AJ, Lurie JD, Zhao W, Bono CM. The effect of race on outcomes of surgical or nonsurgical treatment of patients in the Spine Patient Outcomes Research Trial (SPORT). Spine. 2012;37:1505e1515. 5. Skolasky RL, Thorpe RJ, Wegener ST, Riley LH. Complications and mortality in cervical spine surgery: racial differences. Spine. 2014;39:1506e1512. 6. Ravi P, Sood A, Schmid M, et al. Racial/ethnic disparities in perioperative outcomes of major procedures: results from the national surgical quality improvement program. Ann Surg. 2015;262:955e964. 7. Rangrass G, Ghaferi AA, Dimick JB. Explaining racial disparities in outcomes after cardiac surgery: the role of hospital quality. JAMA Surg. 2014;149:223e227. 8. Schoenfeld AJ, Jiang W, Harris MB, et al. Association between race and postoperative outcomes in a universally insured population versus patients in the state of California. Ann Surg. 2017;266:267e273. 9. Schoenfeld AJ, Belmont PJ, See AA, Bader JO, Bono CM. Patient demographics, insurance status, race, and ethnicity as predictors of morbidity and mortality after spine trauma: a study using the National Trauma Data Bank. Spine J. 2013;13:1766e1773. 10. Schoenfeld AJ, Sandhu T, Sturgeon DJ, Nwosu K, Bono CM. Examining healthcare segregation among racial and ethnic minorities receiving spine surgical procedures in the state of Florida. Spine. 2017;42:1917e1922. 11. Skolasky RL, Maggard AM, Thorpe RJ, Wegener ST, Riley LH. United States hospital admissions for lumbar spinal stenosis: racial and ethnic differences, 2000 through 2009. Spine. 2013;38:2272e2278. 12. Dy CJ, Lane JM, Pan TJ, Parks ML, Lyman S. Racial and socioeconomic disparities in hip fracture care. J Bone Joint Surg Am. 2016;98:858e865. 13. Schoenfeld AJ, Sturgeon DJ, Dimick JB, et al. Disparities in rates of surgical intervention among racial and ethnic minorities in Medicare accountable care organizations. Ann Surg. 2018;269:459e464. 14. Lewis VA, Fraze T, Fisher ES, Shortell SM, Colla CH. ACOs serving high proportions of racial and ethnic minorities lag in quality performance. Health Aff. 2017;36:57e66. 15. Rhoads KF, Patel MI, Ma Y, Schmidt LA. How do integrated health care systems address racial and ethnic disparities in colon cancer? J Clin Oncol. 2015;33:854e860. 16. Hartzman A, Rhodes K. Accountable care organizations and disparities. Health Aff. 2017;36:960. 17. Borza T, Oerline MK, Skolarus TA, et al. Association between hospital participation in Medicare shared Savings program
130
18.
19.
20.
21.
22.
23.
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
accountable care organizations and readmission following major surgery. Ann Surg. 2018;269:873e878. Herrel LA, Norton EC, Hawken SR, Ye Z, Hollenbeck BK, Miller DC. Early impact of Medicare accountable care organizations on cancer surgery outcomes. Cancer. 2016;122:2739e2746. Anderson RE, Ayanian JZ, Zaslavsky AM, McWilliams JM. Quality of care and racial disparities in Medicare among potential ACOs. J Gen Intern Med. 2014;29: 1296e1304. Ayanian JZ, Landon BE, Newhouse JP, Zaslavsky AM. Racial and ethnic disparities among enrollees in Medicare advantage plans. N Engl J Med. 2014;371:2288e2297. Weissman JS, Hasnain-Wynia R, Weinick RM, et al. Pay-forperformance programs to reduce racial/ethnic disparities: what might different designs achieve? J Health Care Poor Underserved. 2012;23:144e160. Nicholas LH, Dimick JB. Bariatric surgery in minority patients before and after implementation of a centers of excellence program. JAMA. 2013;310:1399e1400. Schoenfeld AJ, Sturgeon DJ, Blucher JA, Haider AH, Kang JD. Alterations in 90-day morbidity, mortality, and readmission rates following spine surgery in Medicare Accountable Care Organizations (2009e2014). Spine J. 2018;19:8e14.
24. Yasaitis LC, Pajerowski W, Polsky D, Werner RM. Physicians’ participation in ACOs is lower in places with vulnerable populations than in more affluent communities. Health Aff. 2016;35:1382e1390. 25. Schoenfeld AJ, Zhang D, Walley KC, Bono CM, Harris MB. The influence of race and hospital environment on the care of patients with cervical spine fractures. Spine J. 2016;16:602e607. 26. Harris MB, Reichmann WM, Bono CM, et al. Mortality in elderly patients after cervical spine fractures. J Bone Joint Surg Am. 2010;92:567e574. 27. Schoenfeld AJ, Wahlquist TC, Bono CM, Lehrich JL, Power RK, Harris MB. Changes in the care of patients with cervical spine fractures following health reform in Massachusetts. Injury. 2015;46:1545e1550. 28. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613e619. 29. Schoenfeld AJ, Sturgeon DJ, Harris MB, et al. Changes in the use of lumbar arthrodesis procedures within accountable care organizations. Spine. 2018. 30. Jain A, Menga E, Mesfin A. Outcomes following surgical management of cauda equina syndrome: does race matter? J Racial Ethn Health Disparities. 2018;5:287e292.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Do Medicare Accountable Care Organizations Reduce Disparities After Spinal Fracture? Shaina A. Lipa, MD,a Daniel J. Sturgeon, MS,b Justin A. Blucher, MS,a Mitchel B. Harris, MD,c and Andrew J. Schoenfeld, MD, MSca,b,* a
Department of Orthopaedic Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts b Department of Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts c Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts
article info
abstract
Article history:
Background: National changes in health care disparities within the setting of trauma care
Received 5 April 2019
have not been examined within Accountable Care Organizations (ACOs) or non-ACOs. We
Received in revised form
sought to examine the impact of ACOs on post-treatment outcomes (in-hospital mortality,
19 August 2019
90-day complications, and readmissions), as well as surgical intervention among whites
Accepted 5 September 2019
and nonwhites treated for spinal fractures.
Available online xxx
Materials and methods: We identified all beneficiaries treated for spinal fractures between 2009 and 2014 using national Medicare fee for service claims data. Claims were used to
Keywords:
identify sociodemographic and clinical criteria, receipt of surgery and in-hospital mortal-
Accountable care organziations
ity, 90-day complications, and readmissions. Multivariable logistic regression analysis ac-
Surgical disparities
counting for all confounders was used to determine the effect of race/ethnicity on
Trauma care
outcomes. Nonwhites were compared with whites treated in non-ACOs between 2009 and
Spinal fractures
2011 as the referent.
African Americans
Results: We identified 245,704 patients who were treated for spinal fractures. Two percent of the cohort received care in an ACO, whereas 7% were nonwhite. We found that disparities in the use of surgical fixation for spinal fractures were present in non-ACOs over the period 2009-2014 but did not exist in the context of care provided through ACOs (odds ratio [OR] 0.75; 95% confidence interval [CI] 0.44, 1.28). A disparity in the development of complications existed for nonwhites in non-ACOs (OR 1.09; 95% CI 1.01, 1.17) that was not encountered among nonwhites receiving care in ACOs (OR 1.32; 95% CI 0.90, 1.95). An existing disparity in readmission rates for nonwhites in ACOs over 2009-2011 (OR 1.34; 95% CI 1.01, 1.80) was eliminated in the period 2012-2014 (OR 0.85; 95% CI 0.65, 1.09). Conclusions: Our work reinforces the idea that ACOs could improve health care disparities among nonwhites. There is also the potential that as ACOs become more familiar with care integration and streamlined delivery of services, further improvements in disparities could be realized. ª 2019 Elsevier Inc. All rights reserved.
* Corresponding author. Department of Orthopaedic Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115. Tel.: þ330 329-2594; fax: þ617 278-6919. E-mail address: [email protected] (A.J. Schoenfeld). 0022-4804/$ e see front matter ª 2019 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jss.2019.09.003
124
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
Introduction Health care disparities after orthopedic and spine surgical interventions have been widely reported for racial and ethnic minorities in the past.1-8 Documented disparities in the orthopedic literature include reduced access to surgical services, health care segregation, higher rates of perioperative complications, and inferior outcomes in several respects.2,3,5,9e12 Recent health care reform efforts, such as Accountable Care Organizations (ACOs), have been proposed as initiatives that would not only reduce costs but also lead to reductions in health care disparities.13e19 Through incentives and accountability, it was believed that ACOs would lead the U.S. health care system to realize the “the triple aim”: better population health, better patient experience, and lower costs. Via actuation of these mechanisms, one of the highly touted benefits of ACO implementation included amelioration of health care disparities.15,16 In the past, however, other health reform efforts have been found to unintentionally aggravate existing disparities.20e22 Thus far, studies have shown mixed results regarding the impact of ACOs on surgical health care,17,18,23 including disparities in care among disadvantaged patient populations.13,19,24 For example, Schoenfeld et al. documented persistent differences in the rates of elective surgery, including hip replacement and lumbar spine procedures, among minorities treated in ACOs.13 In the past, surgical interventions after traumatic injury have been considered differently from elective procedures, given the unique features of trauma care in the United States, including delineation of services within the American College of Surgeons tiered trauma system.3 As far as we are aware, national changes in health care disparities within the setting of trauma care have not been examined for ACOs or non-ACOs in the current health reform era. In this context, we sought to examine the impact of Medicare ACO formation on post-treatment outcomes including in-hospital mortality, 90-day complications, and 90day readmissions for white and nonwhite beneficiaries treated for spinal fractures. Secondarily, we also evaluated the influence of ACOs on rates of surgical fixation in these populations. Spinal fractures were selected as a suitable clinical condition for the study, given previous publications documenting existing disparities in treatment and relatively high rates of postinjury morbidity and mortality.9,25e27 Based on prior literature, we hypothesized that ACO formation would not lead to meaningful improvements in health care disparities around treatment for spinal fractures.
Patients and methods We performed this study using national Medicare fee for service claims data for the years 2009-2014, obtained from the Centers for Medicare and Medicaid Services (CMS). We queried Medicare Provider Analysis and Review records to identify beneficiaries enrolled in part A and B plans at the time of treatment for a cervical, thoracic, or lumbar fracture (identified by International Classification of Disease-ninth revision (ICD-9) code 9. Patients younger than 65 y at the time of injury and those insured through Medicare-managed care plans were excluded from this analysis.
We abstracted linked CMS files for all beneficiaries meeting inclusion criteria to determine age at the time of treatment, biologic sex, and reported race/ethnicity. Race/ethnicity was recorded by the CMS as white, black (e.g., African-American), Hispanic, Asian, Native American, and other and dichotomized here as white or nonwhite because of the small numbers of nonwhite patients treated in ACOs. Clinical data obtained from CMS files included age-adjusted Charlson Comorbidity Score,28 location of spinal injury, and whether surgery was performed (defined using an ICD-9 coding algorithm available from the authors by request). Eligible surgical procedures consisted of decompression and/or instrumented fusion procedures. Patients who received stand-alone cement augmentation (e.g., kyphoplasty) performed without decompression or instrumented fusion were considered in the nonoperative cohort as these procedure can often be performed in an outpatient setting, do not require spinal instrumentation, and are not associated with substantial increases in health care costs. This approach to spine procedural classification is aligned with prior research in health care policy.29 Hospital-level data that were surveyed consisted of teaching status of the hospital where the patient was treated and hospital participation in an ACO. As outlined in prior investigations, we delineated ACO affiliation by matching the claims associated with fracture care to each of the Medicare Pioneer ACOs and Shared Savings Program participants identified by the CMS in ACO-specific files for the years 2012, 2013, and 2014.13,23 Beneficiaries treated in organizations that would form an ACO in the years 2012-2014 were considered as a single group in this investigation, with all other individuals assigned to the non-ACO cohort. Claims and beneficiary-level files were used to identify the outcomes of interest including in-hospital mortality, 90-day complications, and readmissions. Complications were defined in line with prior research using an ICD-9 coding system that captures major postsurgical events including surgical site infections, sepsis, shock, cardiopulmonary complications, neurologic events, renal complications, deep venous thrombosis, and pulmonary embolism.8 Readmissions were defined as an admission to the hospital for any cause within 90 d of discharge from the index admission associated with the spinal fracture.
Statistical testing The primary outcome in this analysis was in-hospital mortality after treatment for a spinal fracture with secondary consideration of 90-day complication and readmission events. Temporal changes in the use of surgical intervention for spinal fractures among whites and nonwhites were also evaluated. Patient race/ ethnicity (white versus nonwhite) was considered the primary predictor within the context of the environment of care (ACO versus non-ACO). Patient age, biologic sex, age-adjusted comorbidities, hospital teaching status, and the year of treatment were used as covariates based on our study’s conceptual model. Treatment year was classified as 2009-2011 (prehealth reform) and 2012-2014 (posthealth reform) based on the passage of the Affordable Care Act (2012).13,23 Beneficiaries treated in 2009-2011 by health care entities that would eventually organize as an ACO were evaluated as a single cohort in the prehealth reform period.
125
lipa et al spinal fracture disparities in acos
Table e Sociodemographic and clinical characteristics of all patients and nonwhite patients treated for spinal fractures in an Accountable Care Organization or non-ACO environment between 2009 and 2014.* Characteristic
All patients (n ¼ 245, 704) 2009-2011
P-value
2012-2014
Nonwhite patients (n ¼ 17,619) 2009-2011
P-value
2012-2014
Age (mean, SD)
80.7 (8.1)
80.8 (8.1)
0.02
79.2 (8.6)
79.7 (8.4)
<0.001
Female sex (%)
87,833 (71)
85,009 (70)
<0.001
6095 (70)
6202 (69)
0.17
Teaching hospital (%)
21,334 (17)
21,110 (17)
0.19
2000 (23)
2015 (23)
0.40
2364 (2)
2557 (2)
<0.001
146 (2)
207 (2)
0.003
11,076 (9)
10,084 (8)
<0.001
747 (9)
698 (8)
0.05
<0.001
5.0 (2.0)
5.0 (3.0)
ACO (%) Surgery (%) Age-adjusted Charlson Comorbidity Score (median, IQR)
5.0 (2.0)
5.0 (2.0)
<0.001
ACO ¼ Accountable Care Organization; SD ¼ standard deviation; IQR ¼ interquartile range. * Numbers are rounded.
Multivariable logistic regression analysis accounting for all confounders was used to determine the independent effect of patient race/ethnicity on the outcomes of interest. In all regression tests, nonwhites were compared with whites treated in non-ACOs between 2009 and 2011 as the referent. Disparities were considered to be present if the experience of nonwhites in ACOs was significantly different from that of whites treated in ACOs and/or whites receiving care in nonACOs in the periods under study. A priori, we defined significance for variables that demonstrated P < 0.05 and odds ratio (OR) with 95% confidence intervals (CIs) exclusive of 1.0 following multivariable testing. All analyses were performed using SAS v9.4 (SAS Institute; Cary, NC). Before initiation, this research received a determination of exempt status from our institutional review board and was also approved by the CMS.
Results In the period under study, we identified 245,704 patients who were treated for spinal fractures. Seventy percent of the cohort was female, and the median comorbidity score was 5.0 (interquartile range, 2.0). Overall, two percent of the cohort received care in an ACO (4921/245,704). Seven percent of the population (17,619/245,704) were nonwhite and, in parallel, two percent (353/17,619) of nonwhite patients were treated in ACOs. In 2012-2014, as compared with 2009-2011, the population of nonwhite patients was older, had a slightly larger number of comorbidities, and were marginally more likely to be treated in an ACO (Table). Given the size of our sample, all of these findings were significant, although absolute differences were small. The findings for nonwhite patients mirrored those of the general population, overall (Table). Across the entire cohort, the in-hospital mortality rate was 5% (n ¼ 11,138), whereas 90-day complications occurred in 11% (n ¼ 26,119) and readmissions were documented in 51% (n ¼ 124,726).
and 2012-2014. A comparable reduction in the rates of surgical fixation was appreciated in the nonwhite population as well, although this finding did not meet the threshold for significance (P ¼ 0.05). At baseline, nonwhites in non-ACOs had a significantly lower likelihood of receiving surgical intervention for a spinal fracture (OR 0.83; 95% CI 0.75, 0.91; Fig. 1) and this finding persisted in 2012-2014 (OR 0.79; 95% CI 0.72, 0.86). There was no significant difference in the use of surgery among nonwhites treated in ACOs during 2009-2011 (OR 0.89; 95% CI 0.46, 1.74) and no substantive change was encountered for this outcome over the course of 2012-2014 (OR 0.75; 95% CI 0.44, 1.28).
Changes in in-hospital mortality The mortality rate of the referent group (whites treated in non-ACOs) was 3.9% in 2009-2011. Nonwhites treated in nonACOs (mortality rate of 4.8%) had higher odds of in-hospital mortality during 2009-2011 (OR 1.25; 95% CI 1.12, 1.39; Fig. 2) and 2012-2014 (mortality rate of 6.3%; OR 1.68; 95% CI 1.53, 1.85). A significant difference in mortality was also encountered among nonwhites treated in ACOs in 2012-2014 (mortality rate of 6.3%; OR 1.65; 95% CI 1.04, 2.62). This trend, however, mirrored findings for whites treated in ACOs (mortality rate of 5.2%; OR 1.35; 95% CI 1.07, 1.70) and non-ACOs (mortality rate of 5.0%; OR 1.32; 95% CI 1.27, 1.37) over the same time period.
Changes in complications In 2009-2011, nonwhites treated in non-ACOs (OR 1.01; 95% CI 0.93, 1.09) and ACOs (OR 0.81; 95% CI 0.42, 1.58) did not demonstrate significant differences in the rate of 90-day complications as compared with whites (Fig. 3). In 20122014, however, the odds of complications among nonwhites treated in non-ACOs was significantly increased (OR 1.09; 95% CI 1.01, 1.17), whereas no such change was apparent for nonwhites in ACOs (OR 1.32; 95% CI 0.90, 1.95).
Changes in the use of surgical intervention for spinal fracture
Changes in readmissions
For the cohort as a whole, the use of surgical intervention for spinal fractures decreased from 9% to 8% between 2009-2011
Nonwhites treated in non-ACOs had lower odds of readmissions as compared with whites in both 2009-2011 (OR 0.92;
126
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
Fig. 1 e The adjusted odds of surgical intervention for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
95% CI 0.88, 0.97; Fig. 4) and 2012-2014 (OR 0.93; 95% CI 0.88, 0.97). Nonwhites in ACOs had significantly higher odds of readmission in 2009-2011 (OR 1.34; 95% CI 1.01, 1.80) which became insignificant over the course of the years 2012-2014 (OR 0.85; 95% CI 0.65, 1.09).
Discussion At the time of the implementation of the ACA, ACOs represented a cornerstone of the initiatives designed to reduce
Fig. 2 e The adjusted odds of in-hospital mortality for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
lipa et al spinal fracture disparities in acos
127
Fig. 3 e The adjusted odds of 90-day complications after treatment for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
health care expenditures while simultaneously improving efficiency and quality.14,16e19,24,29 The standardization of practice, integration of organizations and care pathways, and an emphasis on access and preventative care were advertised
as mechanisms that would not only lead to institutional success in the setting of accountable care but also would have a spillover effect in terms of reducing health care disparities.13,16 If changes in health care delivery associated with
Fig. 4 e The adjusted odds of 90-day readmissions after treatment for spinal fractures among whites and nonwhites treated in a future Accountable Care Organization (ACO; 2009-2011), an ACO (2012-2014), and non-ACO environments between 2009 and 2014.
128
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
ACO implementation were to disproportionately impact minorities, however, or if disadvantaged populations were restricted to low-performing ACOs, these health reform efforts could culminate in an exacerbation of disparities. Such phenomena had previously been encountered among other well-intentioned CMS reform efforts, including managed care plans, pay for performance, and the centers of excellence program.20e22 Although early studies have indicated some success for ACOs in terms of reducing health care costs and moderating the use of preference sensitive surgeries,17,18,29 demonstrable improvements in health care quality around spine surgery have not been identified. Moreover, Schoenfeld et al. reported that the formation of ACOs did not lead to significant reductions in disparities regarding access to surgical care for racial and ethnic minorities, although this study largely involved elective interventions and was not restricted to spinal disorders.13 We sought to examine changes in the type and quality of care for spinal fractures among whites and nonwhites treated in ACOs and non-ACOs over a 6-year period encompassing the implementation of health care reform under the Affordable Care Act. Our reliance on Medicare claims data allows surveillance of patients across different health care systems and throughout various regions of the United States. Our methodology, including the means through which we delineate ACO enrollment and identify mortality, complication, and readmission events follow practices previously established in the literature.13,23,29 The rates of perioperative morbidity and mortality identified here are in line with prior reports for a Medicare age population, taking into account the size of the sample and events captured under our umbrella designation of complications.9 We found that disparities in the use of surgical fixation for spinal fractures were present in nonACOs over the period 2009-2014 but did not exist in the context of care provided through ACOs. Although, given the size of our sample, the point estimates for the two time periods are different (e.g., OR 0.89 in 2009-2011 versus 0.75 in 2012-2014), the overlapping 95% CIs for the two time frames (Fig. 1) provides reassurance of no measurable difference. Likewise, a disparity in the development of complications existed for nonwhites in non-ACOs that was not encountered among nonwhites receiving care in ACOs. Perhaps more importantly, an existing disparity in readmission rates for nonwhites in ACOs over 2009-2011 was eliminated in the period 2012-2014. Although elevations in in-hospital mortality across all cohorts in the period 2012-2014 are concerning for separate reasons, these trends were present in white and nonwhite beneficiaries in ACOs and non-ACOs and therefore do not meet our definition for a health care disparity. Our findings are similar to those previously expostulated for changes in cervical fracture care around health reform in Massachusetts.25 In this work, Schoenfeld et al documented no significant differences in the use of surgical interventions for cervical fractures before and after the implementation of health reform initiatives.25 In that effort, however, disparities among nonwhite patients persisted in terms of increased mortality, complications, and hospital length of stay. This was not the case in the present work, where the overall trend appeared to be one of the improvements with respect to
extant disparities in care, particularly the development of 90day complications and readmissions. Importantly, Massachusetts health reform efforts focused primarily on insurance expansion and not on promulgation of accountable care or the use of risk-based reimbursement.25,27 In addition, the results of the present study contrast markedly with glaring disparities in spine fracture care previously published from research using the National Trauma Databank.9 Viewed in this light then, our findings suggest promise for ACOs with respect to ameliorating disparities in spine fracture care. Standardization of care and integrated care delivery pathways occurring frequently in the setting of ACOs may eliminate some of the variations in clinical care that have previously potentiated health care disparities.2,3,30 Nonetheless, there remains room for improvement, especially as the rates of in-hospital mortality remain higher among whites and nonwhites treated in ACOs. As prior work showed an increased likelihood of complications and readmissions among patients receiving lumbar spine surgery in ACOs,23 a more holistic assessment of the reasons similar trends are not apparent in the context of fracture care is needed. Furthermore, although this analysis considered ACOs as a single group, there are invariably high and low performers within this category. As prior efforts outside the field of spine surgery have raised concerns for health care segregation in ACOs,14,24 where minorities are disproportionately served by lower performing and underfunded health care entities, a more granular assessment of the racial and ethnic composition of highand low-performing ACOs is warranted, along with determination of disparities within and between organizations. Evaluating the institutional culture of ACOs is also imperative, as recent work suggests that improved access and an ethic emphasizing equal care are both necessary for the elimination of disparities.8 As such characterizations are not possible, given our reliance on Medicare claims data, we recognize the potential for an ecological bias to confound our results. Additional limitations stem from the retrospective study design and the use of claims data collected for the purposes of payment and not designed for research. As a result, errors in coding at the provider level or during CMS data preparation could influence our determinations, including ACO attribution, and these cannot be adequately characterized or addressed methodologically. We are reassured to some extent in this matter, however, by the fact that the percentage of beneficiaries treated in ACO and non-ACOs mirrors estimates encountered in other works.13,23,29 Organizations electing to form ACOs may be fundamentally different from health care entities that did not join the Pioneer program or Medicare Shared Savings Program, and it is possible that our regression models are unable to fully account for this using the variables available through health care claims. Thus, there is the potential for residual confounding to influence the statistical models we used. Furthermore, because our study period ends only a few years beyond the implementation of the ACA, it is conceivable that ACOs may become more effective in mitigating disparities or their impact may be more discernible in the future. Given our reliance on claims data, we are not able to characterize the extent of spinal fractures, rationale for surgical intervention, or the factors considered in decision-
lipa et al spinal fracture disparities in acos
making for surgery. The appropriate use of surgical interventions similarly cannot be defined. We therefore recognize the possibility that selection and indication bias confound the results of our work. Further testing is necessary to validate the findings presented here and our determinations should not be considered translatable to other aspects of spine surgical care, trauma care, or the care of patients not insured through Medicare. Likewise, our category of nonwhite reflects an amalgam of racial and ethnic groups. This was necessary because of the small number of specific minority subgroups treated within ACOs in the study cohort. As a result, we recognize that our conclusions might not be applicable to any single class of nonwhite patients.2 Additional research involving the experiences of specific minority populations (e.g., African-Americans, Hispanics, Asians, and so forth) in ACOs is needed. Notwithstanding these limitations, this study presents important information for patients, providers, hospital systems, and policymakers. Our work reinforces the idealized concept that ACOs could help improve health care disparities for disadvantaged populations. There is also the potential that, as ACOs become more familiar with care integration and streamlined delivery of services, further improvements in disparities might be realized. This process could be aided, or even accelerated, were additional incentives placed on ACO performance in terms of providing equitable care or eliminating disparities. Ensuring adequate funding for those ACOs that serve a disproportionate number of minority patients also represents an important measure to ensure continued progress along the lines documented here. As we recognize that our determinations represent a single time-point in the course of ACO development, we believe that continued surveillance and cautious optimism regarding future benefits are supported at this time.
Acknowledgment Authors’ contributions: A.J.S., S.A.L., and D.J.S. had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. S.A.L., A.J.S., D.J.S., and M.B.H. contributed to study concept and design. D.J.S. and A.J.S. contributed to acquisition of data; S.A.L., A.J.S., J.A.B., D.J.S., and M.B.H. contributed to analysis and interpretation of data; S.A.L., J.A.B., D.J.S., and A.J.S. contributed to drafting of the manuscript; A.J.S. and M.B.H. contributed to critical revision of the manuscript for important intellectual content. A.J.S. and D.J.S. contributed to statistical analysis; A.J.S. obtained funding; S.A.L. and J.A.B. provided administrative, technical, or material support; A.J.S. and M.B.H. supervised the study.
Disclosure The research in this article was supported by the Centers for Medicare and Medicaid Services Office of Minority HealtheHealth Equity Data Access Program. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Centers for
129
Medicare and Medicaid Services or US Department of Health and Human Services.
references
1. Haider AH, Scott VK, Rehman KA, et al. Racial disparities in surgical care and outcomes in the United States: a comprehensive review of patient, provider, and systemic factors. J Am Coll Surg. 2013;216:482e492. 2. Schoenfeld AJ, Tipirneni R, Nelson JH, Carpenter JE, Iwashyna TJ. The influence of race and ethnicity on complications and mortality after orthopedic surgery: a systematic review of the literature. Med Care. 2014;52:842e851. 3. Pandya NK, Wustrack R, Metz L, Ward D. Current concepts in orthopaedic care disparities. J Am Acad Orthop Surg. 2018;26:828e832. 4. Schoenfeld AJ, Lurie JD, Zhao W, Bono CM. The effect of race on outcomes of surgical or nonsurgical treatment of patients in the Spine Patient Outcomes Research Trial (SPORT). Spine. 2012;37:1505e1515. 5. Skolasky RL, Thorpe RJ, Wegener ST, Riley LH. Complications and mortality in cervical spine surgery: racial differences. Spine. 2014;39:1506e1512. 6. Ravi P, Sood A, Schmid M, et al. Racial/ethnic disparities in perioperative outcomes of major procedures: results from the national surgical quality improvement program. Ann Surg. 2015;262:955e964. 7. Rangrass G, Ghaferi AA, Dimick JB. Explaining racial disparities in outcomes after cardiac surgery: the role of hospital quality. JAMA Surg. 2014;149:223e227. 8. Schoenfeld AJ, Jiang W, Harris MB, et al. Association between race and postoperative outcomes in a universally insured population versus patients in the state of California. Ann Surg. 2017;266:267e273. 9. Schoenfeld AJ, Belmont PJ, See AA, Bader JO, Bono CM. Patient demographics, insurance status, race, and ethnicity as predictors of morbidity and mortality after spine trauma: a study using the National Trauma Data Bank. Spine J. 2013;13:1766e1773. 10. Schoenfeld AJ, Sandhu T, Sturgeon DJ, Nwosu K, Bono CM. Examining healthcare segregation among racial and ethnic minorities receiving spine surgical procedures in the state of Florida. Spine. 2017;42:1917e1922. 11. Skolasky RL, Maggard AM, Thorpe RJ, Wegener ST, Riley LH. United States hospital admissions for lumbar spinal stenosis: racial and ethnic differences, 2000 through 2009. Spine. 2013;38:2272e2278. 12. Dy CJ, Lane JM, Pan TJ, Parks ML, Lyman S. Racial and socioeconomic disparities in hip fracture care. J Bone Joint Surg Am. 2016;98:858e865. 13. Schoenfeld AJ, Sturgeon DJ, Dimick JB, et al. Disparities in rates of surgical intervention among racial and ethnic minorities in Medicare accountable care organizations. Ann Surg. 2018;269:459e464. 14. Lewis VA, Fraze T, Fisher ES, Shortell SM, Colla CH. ACOs serving high proportions of racial and ethnic minorities lag in quality performance. Health Aff. 2017;36:57e66. 15. Rhoads KF, Patel MI, Ma Y, Schmidt LA. How do integrated health care systems address racial and ethnic disparities in colon cancer? J Clin Oncol. 2015;33:854e860. 16. Hartzman A, Rhodes K. Accountable care organizations and disparities. Health Aff. 2017;36:960. 17. Borza T, Oerline MK, Skolarus TA, et al. Association between hospital participation in Medicare shared Savings program
130
18.
19.
20.
21.
22.
23.
j o u r n a l o f s u r g i c a l r e s e a r c h f e b r u a r y 2 0 2 0 ( 2 4 6 ) 1 2 3 e1 3 0
accountable care organizations and readmission following major surgery. Ann Surg. 2018;269:873e878. Herrel LA, Norton EC, Hawken SR, Ye Z, Hollenbeck BK, Miller DC. Early impact of Medicare accountable care organizations on cancer surgery outcomes. Cancer. 2016;122:2739e2746. Anderson RE, Ayanian JZ, Zaslavsky AM, McWilliams JM. Quality of care and racial disparities in Medicare among potential ACOs. J Gen Intern Med. 2014;29: 1296e1304. Ayanian JZ, Landon BE, Newhouse JP, Zaslavsky AM. Racial and ethnic disparities among enrollees in Medicare advantage plans. N Engl J Med. 2014;371:2288e2297. Weissman JS, Hasnain-Wynia R, Weinick RM, et al. Pay-forperformance programs to reduce racial/ethnic disparities: what might different designs achieve? J Health Care Poor Underserved. 2012;23:144e160. Nicholas LH, Dimick JB. Bariatric surgery in minority patients before and after implementation of a centers of excellence program. JAMA. 2013;310:1399e1400. Schoenfeld AJ, Sturgeon DJ, Blucher JA, Haider AH, Kang JD. Alterations in 90-day morbidity, mortality, and readmission rates following spine surgery in Medicare Accountable Care Organizations (2009e2014). Spine J. 2018;19:8e14.
24. Yasaitis LC, Pajerowski W, Polsky D, Werner RM. Physicians’ participation in ACOs is lower in places with vulnerable populations than in more affluent communities. Health Aff. 2016;35:1382e1390. 25. Schoenfeld AJ, Zhang D, Walley KC, Bono CM, Harris MB. The influence of race and hospital environment on the care of patients with cervical spine fractures. Spine J. 2016;16:602e607. 26. Harris MB, Reichmann WM, Bono CM, et al. Mortality in elderly patients after cervical spine fractures. J Bone Joint Surg Am. 2010;92:567e574. 27. Schoenfeld AJ, Wahlquist TC, Bono CM, Lehrich JL, Power RK, Harris MB. Changes in the care of patients with cervical spine fractures following health reform in Massachusetts. Injury. 2015;46:1545e1550. 28. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613e619. 29. Schoenfeld AJ, Sturgeon DJ, Harris MB, et al. Changes in the use of lumbar arthrodesis procedures within accountable care organizations. Spine. 2018. 30. Jain A, Menga E, Mesfin A. Outcomes following surgical management of cauda equina syndrome: does race matter? J Racial Ethn Health Disparities. 2018;5:287e292.