- Email: [email protected]
Impact of Medicaid Expansion on Cardiac Surgery Volume and Outcomes
Impact of Medicaid Expansion on Cardiac Surgery Volume and Outcomes
Department of Surgery, University of Virginia, Charlottesville, Virginia; Virginia Cardiac Services Quality Initiative, Falls Church, Virginia; Department of Cardiac Surgery, University of Michigan, Ann Arbor, Michigan; Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative, Ann Arbor, Michigan; and Department of Cardiac Surgery, Inova Heart and Vascular Institute, Falls Church, Virginia
Background. Thirty-one states approved Medicaid expansion after implementation of the Affordable Care Act. The objective of this study was to evaluate the effect of Medicaid expansion on cardiac surgery volume and outcomes comparing one state that expanded to one that did not. Methods. Data from the Virginia (nonexpansion state) Cardiac Services Quality Initiative and the Michigan (expanded Medicaid, April 2014) Society of Thoracic and Cardiovascular Surgeons Quality Collaborative were analyzed to identify uninsured and Medicaid patients undergoing coronary bypass graft or valve operations, or both. Demographics, operative details, predicted risk scores, and morbidity and mortality rates, stratified by state and compared across era (preexpansion: 18 months before vs postexpansion: 18 months after), were analyzed. Results. In Virginia, there were no differences in volume between eras, whereas in Michigan, there was a significant increase in Medicaid volume (54.4% [558 of
1,026] vs 84.1% [954 of 1,135], p < 0.001) and a corresponding decrease in uninsured volume. In Virginia Medicaid patients, there were no differences in predicted risk of morbidity or mortality or postoperative major morbidities. In Michigan Medicaid patients, a significant decrease in predicted risk of morbidity or mortality (11.9% [8.1% to 20.0%] vs 11.1% [7.7% to 17.9%], p [ 0.02) and morbidities (18.3% [102 of 558] vs 13.2% [126 of 954], p [ 0.008) was identified. Postexpansion was associated with a decreased risk-adjusted rate of major morbidity (odds ratio, 0.69; 95% confidence interval, 0.51 to 0.91; p [ 0.01) in Michigan Medicaid patients. Conclusions. Medicaid expansion was associated with fewer uninsured cardiac surgery patients and improved predicted risk scores and morbidity rates. In addition to improving health care financing, Medicaid expansion may positively affect patient care and outcomes.
T
were deemed “unconstitutionally coercive” [3, 4]. State governments therefore became responsible for deciding whether to expand Medicaid: 31 states (plus the District of Columbia) decided to do so, recognizing the financial implications and contentious political environment [5, 6]. The effect of primary payer status on access to medical care, affordability, and quality has long been debated. Some studies concluded that payer status is not associated with outcomes, instead identifying risk factors and comorbidities as important predictors [7, 8], whereas others have identified a significant relationship between the two [9, 10]. In a series of publications using the Nationwide Inpatient Sample database, Medicaid payer status was associated with increased rates of riskadjusted mortality and worse outcomes. Patients undergoing coronary artery bypass grafting (CABG) and cardiac valve operations had longer lengths of stay and
he Patient Protection and Affordable Care Act (ACA) was signed into law on March 23, 2010, in an effort to provide affordable, quality health care to all Americans [1]. A major pillar of the comprehensive reform package was the expansion of Medicaid, which would provide insurance coverage to all Americans aged younger than 65 years earning less than 138% of the federal poverty level (FPL) [2]. Although the constitutionality of the ACA was upheld by the United States Supreme Court in National Federation of Independent Business (NFIB) vs Sebelius, provisions requiring states to expand Medicaid
Accepted for publication March 27, 2017. Presented at the Fifty-Third Annual Meeting of The Society of Thoracic Surgeons, Houston, TX, Jan 21–25, 2017. Address correspondence to Dr Kron, PO Box 800679, Charlottesville, VA 22908; email: [email protected].
Ó 2017 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2017;104:1251–8) Ó 2017 by The Society of Thoracic Surgeons
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2017.03.079
ADULT CARDIAC
Eric J. Charles, MD, Lily E. Johnston, MD, MPH, Morley A. Herbert, PhD, J. Hunter Mehaffey, MD, Kenan W. Yount, MD, MBA, Donald S. Likosky, PhD, Patricia F. Theurer, BSN, Clifford E. Fonner, BA, Jeffrey B. Rich, MD, Alan M. Speir, MD, Gorav Ailawadi, MD, Richard L. Prager, MD, and Irving L. Kron, MD, for the Investigators for the Virginia Cardiac Services Quality Initiative and the Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative
1252
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
ADULT CARDIAC
Abbreviations and Acronyms ACA AVR BMI CABG CI FPL MSTCVS-QC
= = = = = = =
MV MVR OR PROM PROMM
= = = = =
STS VCSQI
= =
Affordable Care Act aortic valve replacement body mass index coronary artery bypass grafting confidence interval Federal poverty level Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative mitral valve mitral valve replacement odds ratio predicted risk of mortality predicted risk of morbidity or mortality The Society of Thoracic Surgeons Virginia Cardiac Services Quality Initiative
higher costs if they had Medicaid compared with those with Medicare, private insurance, or no insurance [11, 12]. Considering that these analyses were completed before the ACA, the purpose of this study was to determine the effect of Medicaid expansion on improving access to health care and outcomes, using validated cardiac surgery databases from two states that made independent and different decisions regarding Medicaid expansion. We hypothesized that expansion of Medicaid in Michigan would decrease uninsured cardiac surgery volume and improve outcomes compared with Virginia, a state that declined expansion. Improving the understanding of the effects of Medicaid expansion on the delivery of health care and patient outcomes may help state governments make informed policy decisions.
Patients and Methods Study Design This study was exempt from Institutional Review Board approval at both institutions. We analyzed prospectively collected data from the Virginia Cardiac Services Quality Initiative (VCSQI) and the Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative (MSTCVS-QC), both of which contain The Society of Thoracic Surgeons (STS) Adult Cardiac Surgery data. VCSQI consists of 18 cardiac surgery member sites in Virginia, and MSTCVS-QC captures all 33 nonfederal hospitals performing adult cardiac surgery in Michigan. STS data are abstracted at each participating institution by trained coordinators and consist of demographic, payer status, risk factor, perioperative, morbidity, death, discharge, and readmission data on patients aged older than 18. The STS database is recognized as one of the leading national registry initiatives focused on improving surgical outcomes and health care quality [13].
Ann Thorac Surg 2017;104:1251–8
After implementation of the ACA, Michigan expanded Medicaid in April 2014 to provide coverage for all adults making less than 138% of FPL [14]. Conversely, the Virginia government decided not to expand Medicaid and continued to provide traditional Medicaid coverage for pregnant women and children up to 143% of FPL and disabled adults up to 80% of FPL [15]. Analyses were completed on patients undergoing cardiac operations during the 18 months before the start of Medicaid expansion in Michigan (October 2012 to April 2014, preexpansion era) and during the 18 months thereafter (April 2014 to September 2015, postexpansion era).
Patient Selection This study included Medicaid and uninsured patients aged 18 years or older who underwent CABG or valve operations in Virginia (VCSQI) and Michigan (MSTCVSQC) between October 2012 and September 2015. We excluded patients without risk scores, including predicted risk of morbidity or mortality (PROMM) and predicted risk of mortality, which facilitates the normalization of observed-to-expected outcomes across institutions.
Variables The primary end point for this study was rate of major morbidity (reoperation, stroke, kidney failure, deep sternal wound infection, or ventilator support >24 hours) within 30 days postoperatively. Death was not chosen as the primary outcome because event rates are historically low after CABG and valve operations. The primary exposure variable was era.
Statistical Analysis Univariate analysis was performed simultaneously on data from VCSQI and MSTCVS-QC comparing patient demographics, case specifics, predicted risk scores, and outcomes between eras. Analyses were completed on three cohorts: Medicaid patients only (excluding dualeligible Medicare/Medicaid patients), uninsured patients only, and combined Medicaid and uninsured patients. The Student t test and Mann-Whitney U test were used for normally and nonnormally distributed continuous variables, respectively. The c2 test or Fisher exact test were used for categoric variables. Unadjusted and adjusted postexpansion odds ratios for morbidity and death were calculated using multivariable logistic regression. Odds of major morbidity were adjusted for PROMM, and odds of death within 30 days were adjusted for predicted risk of mortality. The a level for significance was 0.05. Statistical analyses were performed with Stata 14 software (StataCorp LP, College Station, TX) and SAS 9.4 software (SAS Institute, Inc, Cary, NC).
Results Characteristics of Cardiac Surgery Patients A total of 1,183 uninsured or Medicaid patients underwent CABG or cardiac valve operations, or both, between
Ann Thorac Surg 2017;104:1251–8
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
1253
p ¼ 0.006) and predicted risk of mortality (0.8% [0.5% to 1.5%] vs 1.0% [0.6% to 1.9%], p ¼ 0.004; Table 3).
October 2012 and September 2015 in Virginia, whereas in Michigan, there were 2,161 patients (Fig 1). There were no differences in Medicaid (24.5% [153 of 624] vs 28.1% [157 of 559], p ¼ 0.20) or uninsured volume (75.5% [471 of 624] vs 71.9% [402 of 559], p ¼ 0.20) between eras in Virginia; whereas in Michigan, there was a significant increase in Medicaid volume (54.4% [558 of 1026] vs 84.1% [954 of 1135], p < 0.001) and a corresponding decrease in uninsured volume (45.6% [468 of 1026] vs 15.9% [181 of 1135], p < 0.001). In all three cohorts of both states, there were no differences in patient age, sex, or body mass index between eras, except for the median age of uninsured patients in Michigan (55 vs 58 years, p ¼ 0.02; Table 1). In Virginia and Michigan, CABG was the most common operation performed, followed by single valve replacement. There were no differences in types of operation, elective, urgent, or emergent case status, or reoperation status between eras in either state (Table 2).
In Virginia, major morbidity and 30-day mortality rates were not significantly different between eras for Medicaid patients, uninsured patients, or combined (Table 4). In Michigan Medicaid patients, a significant decrease in the frequency of postoperative major morbidity was identified (18.3% [102 of 558] vs 13.2% [126 of 954], p ¼ 0.008). This decrease was also observed in the Michigan combined cohort (18.3% [188 of 1,026] vs 15.1% [171 of 1,135], p ¼ 0.04). There were no significant differences in any of the three Michigan cohorts in 30-day mortality rates between eras, although the increase in death in postexpansion uninsured patients neared significance (1.1% [5 of 468] vs 3.3% [6 of 181], p ¼ 0.05; Table 4).
Predicted Risk Scores
Medicaid Expansion and Outcomes
In Virginia Medicaid patients, there were no differences between eras in PROMM or predicted risk of mortality. In Michigan, however, postexpansion Medicaid patients had significantly lower PROMM (11.9% [8.1% to 20.0%] vs 11.1% [7.7% to 17.9%], p ¼ 0.02) and lower predicted risk of mortality (1.0% [0.5% to 1.9%] vs 0.8% [0.5% to 1.6%], p ¼ 0.06; Table 3). Uninsured patients in Virginia had similar rates of PROMM and predicted risk of mortality regardless of era, whereas in Michigan, a significant increase was identified in uninsured patients in both PROMM (10.8% [7.3% to 18.4%] vs 13.5% [8.2% to 21.5%],
In a multivariable logistic regression model, era was not associated with unadjusted or adjusted rates of 30-day major morbidity or mortality in all three cohorts in Virginia (Table 5). In Michigan Medicaid patients, however, postexpansion era was associated with a decreased risk-adjusted rate of postoperative major morbidity (odds ratio [OR], 0.69; 95% confidence interval [CI], 0.51 to 0.91; p ¼ 0.01). This association between postexpansion era and risk-adjusted major morbidity rates was also observed in the Michigan combined cohort (OR, 0.84; 95% CI, 0.77 to 0.92; p < 0.001). In addition, risk-adjusted death increased
Postoperative Morbidity and Mortality Rates ADULT CARDIAC
Fig 1. Distribution of Medicaid and uninsured cardiac surgery patients in Virginia and Michigan. (CABG ¼ coronary artery bypass grafting.)
1254
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
Ann Thorac Surg 2017;104:1251–8
Table 1. Patient Demographics Virginia a
ADULT CARDIAC
Variable
Medicaid Patients, No. Age, y Female BMI, kg/m2 Uninsured Patients, No. Age, y Female BMI, kg/m2 Combined Patients, No. Age, y Female BMI, kg/m2 a
Michigan
Preexpansion
Postexpansion
p Value
Preexpansion
Postexpansion
p Value
153 56 (50–61) 62 (40.5) 29.0 (25.1–33.8)
157 57 (51–62) 58 (36.9) 29.5 (24.5–35.0)
0.35 0.52 0.90
558 56 (50–61) 192 (34.4) 29.3 (25.3–34.3)
954 56 (49–60) 328 (34.4) 29.4 (25.0–34.5)
0.74 0.99 0.56
471 56 (50–61) 136 (28.9) 28.7 (25.7–32.4)
402 56 (50–61) 121 (30.1) 28.9 (24.8–33.4)
0.79 0.69 0.70
468 55 (50–60) 107 (22.9) 29.3 (25.6–33.5)
181 58 (53–61) 41 (22.7) 28.1 (25.0–32.9)
0.02 0.95 0.07
624 56 (50–61) 198 (31.7) 28.8 (25.5–32.9)
559 56 (50–61) 179 (32.0) 29.0 (24.8–33.8)
0.43 0.91 0.82
1,026 56 (50–60) 299 (29.1) 29.3 (25.5–33.8)
1,135 56 (50–60) 369 (32.5) 29.2 (25.0–34.4)
0.51 0.09 0.78
Continuous variables are presented as median (interquartile range) and categoric variables as number (%) or as indicated.
BMI ¼ body mass index.
in Michigan in postexpansion uninsured patients (OR, 3.94; 95% CI, 0.98 to 15.86; p ¼ 0.05; Table 5).
Comment When the ACA was signed into law, it included provisions requiring states to expand Medicaid coverage to all Americans with incomes less than 138% of FPL. Before the ACA, traditional Medicaid insurance principally covered pregnant women, children, and disabled adults, leaving a large number of low-income Americans without health insurance. In 2012 however, the federal mandate to expand Medicaid was deemed unconstitutional by the Supreme Court, and thus, the decision fell to individual states. Although federal incentives were provided to encourage Medicaid expansion, only 31 states and the District of Columbia decided to do so. The financial implications of Medicaid expansion have dominated state government debates and their decisions, with little contemporary data to support or refute the patient benefits of expanding Medicaid. Using two statewide quality databases, the present study identified a significant benefit of Medicaid expansion in terms of decreased uninsured surgical volume and improved outcomes. In a nonexpansion state (Virginia), no differences were identified in uninsured volume, preoperative risk scores, or postoperative morbidity or mortality rates. However, in a state that expanded Medicaid (Michigan), uninsured cardiac surgery volume decreased by 60% over an 18-month period, with a 70% increase in Medicaidinsured volume. After expansion, postoperative rates of major morbidity significantly decreased in Medicaid patients after adjusting for preoperative risk. The observed decrease in uninsured volume in Michigan aligns with predictions that Medicaid expansion
under the ACA would lower the number of uninsured people by half [6]. During the same time period in Virginia, no changes in Medicaid or uninsured volume were identified. According to the Congressional Budget Office, the number of people enrolled in Medicaid is estimated to increase by 12 million people per year between now and 2024 [16]. These changes in uninsured and Medicaid volume will have significant effects on hospital reimbursements for uncompensated care under the Medicare and Medicaid Disproportionate Share Hospital program. Although lower on average than private insurance reimbursement, Medicaid reimbursement is financially more advantageous for hospitals than providing uncompensated care. Under the ACA, Disproportionate Share Hospital payments are being reduced by up to 75%, meaning hospitals will recoup significantly less of the cost for providing care to the uninsured [6]. Within the cardiac surgery population, Medicaid expansion appears to significantly increase the number of insured patients. In addition, the present study identified a relationship between Medicaid expansion and lower rates of postoperative morbidity, which may also decrease total costs of care. Using the American College of Surgeons National Surgical Quality Improvement Program, Ko and colleagues [17] identified that patients with postoperative complications had higher readmission rates and overall costs of care. When extrapolated to other surgical populations and patient care in general, these changes as a result of Medicaid expansion may have significant financial implications for hospitals. Aside from monetary benefits, patients may benefit from improvements in preoperative health status. No changes in predicted risk scores were identified in Virginia. In Michigan, however, postexpansion Medicaid patients had a significant decrease in PROMM and lower
Ann Thorac Surg 2017;104:1251–8
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
1255
Table 2. Operative Case Specifics Virginia Variable
a
a
Preexpansion
Postexpansion
153
157
p Value
Preexpansion
Postexpansion
558
954
0.61 99 15 19 9 4 7
(64.7) (9.8) (12.4) (5.9) (2.6) (4.6)
105 14 12 5 2 5
(66.9) (8.9) (7.6) (3.2) (1.3) (3.2)
0.61 409 53 29 21 5 41
(73.3) (9.5) (5.2) (3.8) (0.9) (7.3)
716 99 47 32 12 48
(75.1) (10.4) (4.9) (3.4) (1.3) (5.0)
0.59 71 (46.4) 79 (51.6) 3 (2.0)
64 (40.8) 89 (56.7) 4 (2.5)
141 (92.2) 9 (5.9) 3 (2.0)
148 (94.3) 8 (5.1) 1 (0.6)
471
402
0.41 219 (39.3) 326 (58.4) 13 (2.3)
373 (39.1) 547 (57.3) 34 (3.6)
534 (95.7) 21 (3.8) 2 (0.4)
917 (96.1) 36 (3.8) 1 (0.1)
0.56
0.41
468
181
0.96 379 40 17 11 14 10
(80.5) (8.5) (3.6) (2.3) (3.0) (2.1)
319 33 12 8 9 11
(79.4) (8.2) (3.0) (2.0) (2.2) (2.7)
0.55 370 34 15 15 4 30
(79.1) (7.3) (3.2) (3.2) (0.9) (6.4)
146 10 3 6 1 15
(80.7) (5.5) (1.7) (3.3) (0.6) (8.3)
0.15 102 (21.7) 334 (71.2) 33 (7.0)
103 (25.6) 281 (69.9) 18 (4.5)
456 (96.8) 14 (3.0) 1 (0.2)
389 (96.8) 12 (3.0) 1 (0.2)
624
559
0.30 103 (22.0) 341 (72.9) 24 (5.1)
30 (16.6) 142 (78.5) 9 (5.0)
456 (97.4) 11 (2.4) 1 (0.2)
176 (97.2) 5 (2.8) 0 (0)
0.99
0.79
1,026
1,135
0.75 478 55 36 20 18 17
(76.6) (8.8) (5.8) (3.2) (2.9) (2.7)
424 47 24 13 11 16
(75.8) (8.4) (4.3) (2.3) (2.0) (2.9)
0.82 779 87 44 36 9 71
(75.9) (8.5) (4.3) (3.5) (0.9) (6.9)
862 109 50 38 13 63
(75.9) (9.6) (4.4) (3.3) (1.2) (5.6)
0.29 173 (27.8) 413 (66.4) 36 (5.8)
167 (29.9) 370 (66.2) 22 (3.9)
597 (95.7) 23 (3.7) 4 (0.6)
537 (96.1) 20 (3.6) 2 (0.4)
p Value
0.11 322 (31.3) 667 (65.0) 37 (3.6)
403 (35.5) 689 (60.7) 43 (3.8)
990 (96.5) 32 (3.1) 3 (0.3)
1,093 (96.3) 41 (3.6) 1 (0.1)
0.79
0.44
Data are presented as number (%) or as indicated.
AVR ¼ aortic valve replacement; CABG ¼ coronary artery bypass graft operation; mitral valve repair with or without CABG.
MVR ¼ mitral valve replacement;
MV repair CABG ¼
ADULT CARDIAC
Medicaid Patients, No. Case mix CABG only AVR only MVR only AVR þ CABG MVR þ CABG MV repair CABG Case status Elective Urgent Emergent Reoperation status First operation First reoperation Reoperation (2) Uninsured Patients, No. Case mix CABG only AVR only MVR only AVR þ CABG MVR þ CABG MV repair CABG Case status Elective Urgent Emergent Reoperation status First operation First reoperation Reoperation (2) Combined Patients, No. Case mix CABG only AVR only MVR only AVR þ CABG MVR þ CABG MV repair CABG Case status Elective Urgent Emergent Reoperation status First operation First reoperation Reoperation (2)
Michigan
1256
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
Ann Thorac Surg 2017;104:1251–8
Table 3. Preoperative Predicted Risk Scores Virginia a
ADULT CARDIAC
Variable
Medicaid Patients, No. PROMM, % PROM, % Uninsured Patients, No. PROMM, % PROM, % Combined Patients, No. PROMM, % PROM, % a
Michigan
Preexpansion
Postexpansion
p Value
Preexpansion
Postexpansion
p Value
153 14.0 (8.7–24.9) 1.1 (0.6–2.3)
157 13.2 (8.7–21.3) 0.9 (0.6–2.5)
0.59 0.85
558 11.9 (8.1–20.0) 1.0 (0.5–1.9)
954 11.1 (7.7–17.9) 0.8 (0.5–1.6)
0.02 0.06
471 12.4 (8.1–20.9) 0.9 (0.5–1.8)
402 12.0 (8.3–18.6) 0.8 (0.5–1.6)
0.56 0.37
468 10.8 (7.3–18.4) 0.8 (0.5–1.5)
181 13.5 (8.2–21.5) 1.0 (0.6–1.9)
0.006 0.004
624 12.6 (8.2–21.7) 0.9 (0.5–1.9)
559 12.4 (8.4–19.1) 0.9 (0.5–1.8)
0.50 0.46
1,026 11.6 (7.6–18.9) 0.9 (0.5–1.7)
1,135 11.3 (7.8–18.4) 0.9 (0.5–1.7)
0.53 0.78
Data are presented as median (interquartile range).
PROM ¼ predicted risk of mortality;
PROMM ¼ predicted risk of morbidity or mortality.
predicted risk of mortality. It is difficult to determine how Medicaid insurance may affect preoperative optimization from the present study, but the findings may reflect increased access to primary care. Using National Health Interview Survey data, Wherry and Miller [18] identified an association between Medicaid expansion and increased health care utilization and more frequent diagnosis of health problems. In a single-state review of Medicaid expansion, Benitez and colleagues [19] found lower rates of unmet medical needs and fewer people without routine access to health care. Proper diagnosis and medical management of comorbidities may result in lower preoperative risk scores. Surprisingly, no significant differences in the elective, urgent, or emergent case status were identified in either state. This lack of difference in Michigan may indicate that changes observed may not reflect increased access to
preoperative care but rather a shift of younger, betterinformed, low-risk patients from being uninsured to having Medicaid. This shift in relatively healthy patients subsequently reduces the risk pool in the postexpansion Medicaid population. If Medicaid expansion does lead to improved preoperative optimization, then an associated improvement in outcomes would be expected. In the current study, postexpansion Medicaid patients in Michigan had significantly lower rates of postoperative morbidity. For health care providers, reimbursement rates are important but not nearly as important as the effect of policy interventions on actual patient outcomes. In Michigan, postexpansion Medicaid patients were 30% less likely to experience a major postoperative complication (adjusted OR, 0.69). In addition to Medicaid insurance reimbursements and higher Disproportionate Share
Table 4. Morbidity and Mortality Outcomes Virginia a
Variable
Medicaid Patients, No. Major morbidityb 30-day mortality Uninsured Patients, No. Major morbidity 30-day mortality Combined Patients, No. Major morbidity 30-day mortality
Michigan
Preexpansion
Postexpansion
p Value
Preexpansion
Postexpansion
p Value
153 23 (15.0) 4 (2.6)
157 21 (13.4) 3 (1.9)
0.68 0.68
558 102 (18.3) 14 (2.5)
954 126 (13.2) 15 (1.6)
0.008 0.20
471 61 (13.0) 8 (1.7)
402 51 (12.7) 10 (2.5)
0.91 0.41
468 71 (15.2) 5 (1.1)
181 26 (14.4) 6 (3.3)
0.80 0.05
624 84 (13.5) 12 (1.9)
559 72 (12.9) 13 (2.3)
0.77 0.63
1,026 188 (18.3) 19 (1.9)
1,135 171 (15.1) 21 (1.9)
0.04 1.0
a b Data are presented as number (%) or as indicated. Major morbidity includes reoperation, stroke, kidney failure, deep sternal wound infection, and ventilator support >24 hours within 30 days of operation.
Ann Thorac Surg 2017;104:1251–8
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
1257
Table 5. Unadjusted and Adjusted Postexpansion Odds Ratios for Major Morbidity and 30-Day Mortality Virginia Variable
a
Odds Ratio (95% CI)
p Value
Odds Ratio (95% CI)
0.87 (0.46–1.65) 0.73 (0.16–3.30)
0.7 0.7
0.68 (0.52–0.89) 0.62 (0.30–1.30)
0.004 0.20
0.92 (0.47–1.78) 0.63 (0.13–3.02)
0.8 0.6
0.69 (0.51–0.91) 0.61 (0.30–1.27)
0.01 0.19
0.98 (0.66–1.45) 1.48 (0.58–3.78)
0.9 0.4
0.94 (0.58–1.53) 3.18 (0.96–10.54)
0.80 0.06
0.97 (0.63–1.51) 1.30 (0.48–3.57)
0.9 0.6
0.83 (0.49–1.42) 3.94 (0.98–15.86)
0.50 0.05
0.92 (0.70–1.22) 1.16 (0.62–2.18)
0.6 0.7
0.86 (0.79–0.93) 1.14 (0.95–1.37)
<0.001 0.16
0.89 (0.70–1.19) 1.02 (0.53–1.98)
0.4 0.9
0.84 (0.77–0.92) 1.10 (0.91–1.32)
<0.001 0.35
Major morbidity adjusted for predicted risk of morbidity or mortality.
b
p Value
Mortality adjusted for predicted risk of mortality.
CI ¼ confidence interval.
Hospital payments, states that expand Medicaid may see improvements in access to care, preoperative patient optimization, and surgical outcomes. Although the focus was to evaluate the effect of expanding Medicaid, some findings in the uninsured patient population warrant discussion. In Michigan, the uninsured cardiac surgery volume dropped by 60%, but there was no change in Virginia. The Michigan patients who were still uninsured after Medicaid expansion were older, had a higher PROMM, a higher predicted risk of mortality, and a greater 30-day mortality rate. Because Medicaid expansion appears to benefit a significant number of patients, there remains a vulnerable population of uninsured patients who have higher predicted risk scores and rates of morbidity and mortality. An important next step will be to identify patients who did not qualify under traditional Medicaid but gained coverage after expansion and compare them to postexpansion uninsured patients, paying special attention to social determinants of health (ie, education level, employment status, and housing conditions). This analysis will help determine whether the uninsured actually qualify for expanded Medicaid but have barriers to enrollment or whether they are indeed sicker because they cannot access primary care. The current study has several limitations. Considering that insurance status is only one factor related to the delivery of health care, no definitive or causal relationships between Medicaid expansion and outcomes can be
determined; rather, these are inferences that should be used to direct further health policy implementation and research. The data set did not allow for postexpansion Medicaid patients to be separated by whether or not they qualified under the traditional system or only after expansion. This study may not reflect changes nationwide, considering it is an analysis of cardiac surgery patients in only two states with differently sized populations. The findings are limited by the relatively short interval of follow-up and lack of complementary percutaneous coronary intervention data. Considering that Medicaid expansion is a fairly recent policy change, repeat analyses will be necessary to validate these findings and assess the effect on patient access to both surgery and interventional cardiology. Using two statewide databases that collect validated and audited STS cardiac surgery data, the present study identified an association between Medicaid expansion and decreased volume of uninsured cardiac surgery patients, lower preoperative predicted risk scores, and improved postoperative outcomes. Considering that reimbursements for uncompensated care will continue to decline, it appears that Medicaid expansion may improve health care financing and positively affect patient care and outcomes. Continuous evaluation of health policy changes such as Medicaid expansion is necessary to expedite advancements in the delivery of high-quality, cost-effective health care.
ADULT CARDIAC
Medicaid Unadjusted Major morbiditya 30-day mortalityb Adjusted Major morbidity 30-day mortality Uninsured Unadjusted Major morbidity 30-day mortality Adjusted Major morbidity 30-day mortality Combined Unadjusted Major morbidity 30-day mortality Adjusted Major morbidity 30-day mortality
Michigan
ADULT CARDIAC
1258
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
The authors extend special appreciation to Jenna Charles for her invaluable assistance in navigating the intricacies of the Patient Protection and Affordable Care Act. This work was supported by grants UM1-HL-088925 and T32-HL-007849 (I.L.K.) from the National Heart, Lung, and Blood Institute and R01-HS-022535 (D.S.L.) from the Agency for Healthcare Research and Quality. Support for MSTCVS Quality Collaborative is provided by Blue Cross and Blue Shield of Michigan and Blue Care Network as part of the BCBSM Value Partnerships program.
Audio Discussion: Audio of the discussion that followed the presentation of this paper at the STS Annual Meeting can be accessed in the online version of this article [http://dx.doi.org/10.1016/j.athoracsur.2017.03. 079] on http://www.annalsthoracicsurgery.org.
References 1. U.S. Department of Health and Human Services. Key features of the Affordable Care Act by year. Available at http:// www.hhs.gov/healthcare/facts-and-features/key-features-ofaca-by-year/index.html. Accessed August 13, 2016. 2. Medicaid.com. Centers for Medicare & Medicaid Services. Affordable Care Act eligibility, Medicaid expansion. Available at https://www.medicaid.gov/affordable-care-act/ eligibility/index.html. Accessed August 13, 2016. 3. The Henry J. Kaiser Family Foundation. A guide to the Supreme Court’s decision on the ACA’s Medicaid expansion. Available at https://kaiserfamilyfoundation.files.wordpress. com/2013/01/8347.pdf. Accessed August 13, 2016. 4. U.S. Supreme Court. National Federation of Independent Business et al. v. Sebelius, Secretary of Health and Human Services, et al. Available at https://www.supremecourt.gov/ opinions/11pdf/11-393c3a2.pdf. Accessed August 15, 2016. 5. The Henry J. Kaiser Family Foundation. Status of state action on the Medicaid expansion decision. Available at http://kff. org/health-reform/state-indicator/state-activity-aroundexpanding-medicaid-under-the-affordable-care-act/?current Timeframe¼0&sortModel¼%7B%22colId%22:%22Location %22,%22sort%22:%22asc%22%7D. Accessed August 15, 2016. 6. Graves JA. Medicaid expansion opt-outs and uncompensated care. N Engl J Med 2012;367:2365–7.
Ann Thorac Surg 2017;104:1251–8
7. Polanco A, Breglio AM, Itagaki S, Goldstone AB, Chikwe J. Does payer status impact clinical outcomes after cardiac surgery? A propensity analysis. Heart Surg Forum 2012;15: E262–7. 8. Mell MW, Baker LC. Payer status, preoperative surveillance, and rupture of abdominal aortic aneurysms in the US Medicare population. Ann Vasc Surg 2014;28:1378–83. 9. Adepoju L, Wanjiku S, Brown M, et al. Effect of insurance payer status on the surgical treatment of early stage breast cancer: data analysis from a single health system. JAMA Surg 2013;148:570–2. 10. Boxer LK, Dimick JB, Wainess RM, et al. Payer status is related to differences in access and outcomes of abdominal aortic aneurysm repair in the United States. Surgery 2003;134:142–5. 11. Lapar DJ, Bhamidipati CM, Walters DM, et al. Primary payer status affects outcomes for cardiac valve operations. J Am Coll Surg 2011;212:759–67. 12. LaPar DJ, Stukenborg GJ, Guyer RA, et al. Primary payer status is associated with mortality and resource utilization for coronary artery bypass grafting. Circulation 2012;126(11 Suppl 1):S132–9. 13. Jacobs JP, Shahian DM, Prager RL, et al. The Society of Thoracic Surgeons National Database 2016 Annual Report. Ann Thorac Surg 2016;102:1790–7. 14. Ayanian JZ. Michigan’s approach to Medicaid expansion and reform. N Engl J Med 2013;369:1773–5. 15. Virginia Department of Social Services Medicaid Fact Sheet #45. Available at https://www.dss.virginia.gov/files/division/ bp/medical_assistance/forms/all_other/d032-03-0452-03-eng. pdf. Accessed December 14, 2016. 16. Congressional Budget Office. Updated estimates of the insurance coverage provisions of the Affordable Care Act. Available at https://www.cbo.gov/sites/default/files/cbo files/attachments/45010-breakout-AppendixB.pdf. Accessed December 7, 2016. 17. Lawson EH, Hall BL, Louie R, et al. Association between occurrence of a postoperative complication and readmission: implications for quality improvement and cost savings. Ann Surg 2013;258:10–8. 18. Wherry LR, Miller S. Early coverage, access, utilization, and health effects associated with the Affordable Care Act Medicaid expansions: a quasi-experimental study. Ann Intern Med 2016;164:795–803. 19. Benitez JA, Creel L, Jennings J. Kentucky’s Medicaid expansion showing early promise on coverage and access to care. Health Aff (Millwood) 2016;35:528–34.
Department of Surgery, University of Virginia, Charlottesville, Virginia; Virginia Cardiac Services Quality Initiative, Falls Church, Virginia; Department of Cardiac Surgery, University of Michigan, Ann Arbor, Michigan; Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative, Ann Arbor, Michigan; and Department of Cardiac Surgery, Inova Heart and Vascular Institute, Falls Church, Virginia
Background. Thirty-one states approved Medicaid expansion after implementation of the Affordable Care Act. The objective of this study was to evaluate the effect of Medicaid expansion on cardiac surgery volume and outcomes comparing one state that expanded to one that did not. Methods. Data from the Virginia (nonexpansion state) Cardiac Services Quality Initiative and the Michigan (expanded Medicaid, April 2014) Society of Thoracic and Cardiovascular Surgeons Quality Collaborative were analyzed to identify uninsured and Medicaid patients undergoing coronary bypass graft or valve operations, or both. Demographics, operative details, predicted risk scores, and morbidity and mortality rates, stratified by state and compared across era (preexpansion: 18 months before vs postexpansion: 18 months after), were analyzed. Results. In Virginia, there were no differences in volume between eras, whereas in Michigan, there was a significant increase in Medicaid volume (54.4% [558 of
1,026] vs 84.1% [954 of 1,135], p < 0.001) and a corresponding decrease in uninsured volume. In Virginia Medicaid patients, there were no differences in predicted risk of morbidity or mortality or postoperative major morbidities. In Michigan Medicaid patients, a significant decrease in predicted risk of morbidity or mortality (11.9% [8.1% to 20.0%] vs 11.1% [7.7% to 17.9%], p [ 0.02) and morbidities (18.3% [102 of 558] vs 13.2% [126 of 954], p [ 0.008) was identified. Postexpansion was associated with a decreased risk-adjusted rate of major morbidity (odds ratio, 0.69; 95% confidence interval, 0.51 to 0.91; p [ 0.01) in Michigan Medicaid patients. Conclusions. Medicaid expansion was associated with fewer uninsured cardiac surgery patients and improved predicted risk scores and morbidity rates. In addition to improving health care financing, Medicaid expansion may positively affect patient care and outcomes.
T
were deemed “unconstitutionally coercive” [3, 4]. State governments therefore became responsible for deciding whether to expand Medicaid: 31 states (plus the District of Columbia) decided to do so, recognizing the financial implications and contentious political environment [5, 6]. The effect of primary payer status on access to medical care, affordability, and quality has long been debated. Some studies concluded that payer status is not associated with outcomes, instead identifying risk factors and comorbidities as important predictors [7, 8], whereas others have identified a significant relationship between the two [9, 10]. In a series of publications using the Nationwide Inpatient Sample database, Medicaid payer status was associated with increased rates of riskadjusted mortality and worse outcomes. Patients undergoing coronary artery bypass grafting (CABG) and cardiac valve operations had longer lengths of stay and
he Patient Protection and Affordable Care Act (ACA) was signed into law on March 23, 2010, in an effort to provide affordable, quality health care to all Americans [1]. A major pillar of the comprehensive reform package was the expansion of Medicaid, which would provide insurance coverage to all Americans aged younger than 65 years earning less than 138% of the federal poverty level (FPL) [2]. Although the constitutionality of the ACA was upheld by the United States Supreme Court in National Federation of Independent Business (NFIB) vs Sebelius, provisions requiring states to expand Medicaid
Accepted for publication March 27, 2017. Presented at the Fifty-Third Annual Meeting of The Society of Thoracic Surgeons, Houston, TX, Jan 21–25, 2017. Address correspondence to Dr Kron, PO Box 800679, Charlottesville, VA 22908; email: [email protected].
Ó 2017 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2017;104:1251–8) Ó 2017 by The Society of Thoracic Surgeons
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2017.03.079
ADULT CARDIAC
Eric J. Charles, MD, Lily E. Johnston, MD, MPH, Morley A. Herbert, PhD, J. Hunter Mehaffey, MD, Kenan W. Yount, MD, MBA, Donald S. Likosky, PhD, Patricia F. Theurer, BSN, Clifford E. Fonner, BA, Jeffrey B. Rich, MD, Alan M. Speir, MD, Gorav Ailawadi, MD, Richard L. Prager, MD, and Irving L. Kron, MD, for the Investigators for the Virginia Cardiac Services Quality Initiative and the Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative
1252
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
ADULT CARDIAC
Abbreviations and Acronyms ACA AVR BMI CABG CI FPL MSTCVS-QC
= = = = = = =
MV MVR OR PROM PROMM
= = = = =
STS VCSQI
= =
Affordable Care Act aortic valve replacement body mass index coronary artery bypass grafting confidence interval Federal poverty level Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative mitral valve mitral valve replacement odds ratio predicted risk of mortality predicted risk of morbidity or mortality The Society of Thoracic Surgeons Virginia Cardiac Services Quality Initiative
higher costs if they had Medicaid compared with those with Medicare, private insurance, or no insurance [11, 12]. Considering that these analyses were completed before the ACA, the purpose of this study was to determine the effect of Medicaid expansion on improving access to health care and outcomes, using validated cardiac surgery databases from two states that made independent and different decisions regarding Medicaid expansion. We hypothesized that expansion of Medicaid in Michigan would decrease uninsured cardiac surgery volume and improve outcomes compared with Virginia, a state that declined expansion. Improving the understanding of the effects of Medicaid expansion on the delivery of health care and patient outcomes may help state governments make informed policy decisions.
Patients and Methods Study Design This study was exempt from Institutional Review Board approval at both institutions. We analyzed prospectively collected data from the Virginia Cardiac Services Quality Initiative (VCSQI) and the Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative (MSTCVS-QC), both of which contain The Society of Thoracic Surgeons (STS) Adult Cardiac Surgery data. VCSQI consists of 18 cardiac surgery member sites in Virginia, and MSTCVS-QC captures all 33 nonfederal hospitals performing adult cardiac surgery in Michigan. STS data are abstracted at each participating institution by trained coordinators and consist of demographic, payer status, risk factor, perioperative, morbidity, death, discharge, and readmission data on patients aged older than 18. The STS database is recognized as one of the leading national registry initiatives focused on improving surgical outcomes and health care quality [13].
Ann Thorac Surg 2017;104:1251–8
After implementation of the ACA, Michigan expanded Medicaid in April 2014 to provide coverage for all adults making less than 138% of FPL [14]. Conversely, the Virginia government decided not to expand Medicaid and continued to provide traditional Medicaid coverage for pregnant women and children up to 143% of FPL and disabled adults up to 80% of FPL [15]. Analyses were completed on patients undergoing cardiac operations during the 18 months before the start of Medicaid expansion in Michigan (October 2012 to April 2014, preexpansion era) and during the 18 months thereafter (April 2014 to September 2015, postexpansion era).
Patient Selection This study included Medicaid and uninsured patients aged 18 years or older who underwent CABG or valve operations in Virginia (VCSQI) and Michigan (MSTCVSQC) between October 2012 and September 2015. We excluded patients without risk scores, including predicted risk of morbidity or mortality (PROMM) and predicted risk of mortality, which facilitates the normalization of observed-to-expected outcomes across institutions.
Variables The primary end point for this study was rate of major morbidity (reoperation, stroke, kidney failure, deep sternal wound infection, or ventilator support >24 hours) within 30 days postoperatively. Death was not chosen as the primary outcome because event rates are historically low after CABG and valve operations. The primary exposure variable was era.
Statistical Analysis Univariate analysis was performed simultaneously on data from VCSQI and MSTCVS-QC comparing patient demographics, case specifics, predicted risk scores, and outcomes between eras. Analyses were completed on three cohorts: Medicaid patients only (excluding dualeligible Medicare/Medicaid patients), uninsured patients only, and combined Medicaid and uninsured patients. The Student t test and Mann-Whitney U test were used for normally and nonnormally distributed continuous variables, respectively. The c2 test or Fisher exact test were used for categoric variables. Unadjusted and adjusted postexpansion odds ratios for morbidity and death were calculated using multivariable logistic regression. Odds of major morbidity were adjusted for PROMM, and odds of death within 30 days were adjusted for predicted risk of mortality. The a level for significance was 0.05. Statistical analyses were performed with Stata 14 software (StataCorp LP, College Station, TX) and SAS 9.4 software (SAS Institute, Inc, Cary, NC).
Results Characteristics of Cardiac Surgery Patients A total of 1,183 uninsured or Medicaid patients underwent CABG or cardiac valve operations, or both, between
Ann Thorac Surg 2017;104:1251–8
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
1253
p ¼ 0.006) and predicted risk of mortality (0.8% [0.5% to 1.5%] vs 1.0% [0.6% to 1.9%], p ¼ 0.004; Table 3).
October 2012 and September 2015 in Virginia, whereas in Michigan, there were 2,161 patients (Fig 1). There were no differences in Medicaid (24.5% [153 of 624] vs 28.1% [157 of 559], p ¼ 0.20) or uninsured volume (75.5% [471 of 624] vs 71.9% [402 of 559], p ¼ 0.20) between eras in Virginia; whereas in Michigan, there was a significant increase in Medicaid volume (54.4% [558 of 1026] vs 84.1% [954 of 1135], p < 0.001) and a corresponding decrease in uninsured volume (45.6% [468 of 1026] vs 15.9% [181 of 1135], p < 0.001). In all three cohorts of both states, there were no differences in patient age, sex, or body mass index between eras, except for the median age of uninsured patients in Michigan (55 vs 58 years, p ¼ 0.02; Table 1). In Virginia and Michigan, CABG was the most common operation performed, followed by single valve replacement. There were no differences in types of operation, elective, urgent, or emergent case status, or reoperation status between eras in either state (Table 2).
In Virginia, major morbidity and 30-day mortality rates were not significantly different between eras for Medicaid patients, uninsured patients, or combined (Table 4). In Michigan Medicaid patients, a significant decrease in the frequency of postoperative major morbidity was identified (18.3% [102 of 558] vs 13.2% [126 of 954], p ¼ 0.008). This decrease was also observed in the Michigan combined cohort (18.3% [188 of 1,026] vs 15.1% [171 of 1,135], p ¼ 0.04). There were no significant differences in any of the three Michigan cohorts in 30-day mortality rates between eras, although the increase in death in postexpansion uninsured patients neared significance (1.1% [5 of 468] vs 3.3% [6 of 181], p ¼ 0.05; Table 4).
Predicted Risk Scores
Medicaid Expansion and Outcomes
In Virginia Medicaid patients, there were no differences between eras in PROMM or predicted risk of mortality. In Michigan, however, postexpansion Medicaid patients had significantly lower PROMM (11.9% [8.1% to 20.0%] vs 11.1% [7.7% to 17.9%], p ¼ 0.02) and lower predicted risk of mortality (1.0% [0.5% to 1.9%] vs 0.8% [0.5% to 1.6%], p ¼ 0.06; Table 3). Uninsured patients in Virginia had similar rates of PROMM and predicted risk of mortality regardless of era, whereas in Michigan, a significant increase was identified in uninsured patients in both PROMM (10.8% [7.3% to 18.4%] vs 13.5% [8.2% to 21.5%],
In a multivariable logistic regression model, era was not associated with unadjusted or adjusted rates of 30-day major morbidity or mortality in all three cohorts in Virginia (Table 5). In Michigan Medicaid patients, however, postexpansion era was associated with a decreased risk-adjusted rate of postoperative major morbidity (odds ratio [OR], 0.69; 95% confidence interval [CI], 0.51 to 0.91; p ¼ 0.01). This association between postexpansion era and risk-adjusted major morbidity rates was also observed in the Michigan combined cohort (OR, 0.84; 95% CI, 0.77 to 0.92; p < 0.001). In addition, risk-adjusted death increased
Postoperative Morbidity and Mortality Rates ADULT CARDIAC
Fig 1. Distribution of Medicaid and uninsured cardiac surgery patients in Virginia and Michigan. (CABG ¼ coronary artery bypass grafting.)
1254
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
Ann Thorac Surg 2017;104:1251–8
Table 1. Patient Demographics Virginia a
ADULT CARDIAC
Variable
Medicaid Patients, No. Age, y Female BMI, kg/m2 Uninsured Patients, No. Age, y Female BMI, kg/m2 Combined Patients, No. Age, y Female BMI, kg/m2 a
Michigan
Preexpansion
Postexpansion
p Value
Preexpansion
Postexpansion
p Value
153 56 (50–61) 62 (40.5) 29.0 (25.1–33.8)
157 57 (51–62) 58 (36.9) 29.5 (24.5–35.0)
0.35 0.52 0.90
558 56 (50–61) 192 (34.4) 29.3 (25.3–34.3)
954 56 (49–60) 328 (34.4) 29.4 (25.0–34.5)
0.74 0.99 0.56
471 56 (50–61) 136 (28.9) 28.7 (25.7–32.4)
402 56 (50–61) 121 (30.1) 28.9 (24.8–33.4)
0.79 0.69 0.70
468 55 (50–60) 107 (22.9) 29.3 (25.6–33.5)
181 58 (53–61) 41 (22.7) 28.1 (25.0–32.9)
0.02 0.95 0.07
624 56 (50–61) 198 (31.7) 28.8 (25.5–32.9)
559 56 (50–61) 179 (32.0) 29.0 (24.8–33.8)
0.43 0.91 0.82
1,026 56 (50–60) 299 (29.1) 29.3 (25.5–33.8)
1,135 56 (50–60) 369 (32.5) 29.2 (25.0–34.4)
0.51 0.09 0.78
Continuous variables are presented as median (interquartile range) and categoric variables as number (%) or as indicated.
BMI ¼ body mass index.
in Michigan in postexpansion uninsured patients (OR, 3.94; 95% CI, 0.98 to 15.86; p ¼ 0.05; Table 5).
Comment When the ACA was signed into law, it included provisions requiring states to expand Medicaid coverage to all Americans with incomes less than 138% of FPL. Before the ACA, traditional Medicaid insurance principally covered pregnant women, children, and disabled adults, leaving a large number of low-income Americans without health insurance. In 2012 however, the federal mandate to expand Medicaid was deemed unconstitutional by the Supreme Court, and thus, the decision fell to individual states. Although federal incentives were provided to encourage Medicaid expansion, only 31 states and the District of Columbia decided to do so. The financial implications of Medicaid expansion have dominated state government debates and their decisions, with little contemporary data to support or refute the patient benefits of expanding Medicaid. Using two statewide quality databases, the present study identified a significant benefit of Medicaid expansion in terms of decreased uninsured surgical volume and improved outcomes. In a nonexpansion state (Virginia), no differences were identified in uninsured volume, preoperative risk scores, or postoperative morbidity or mortality rates. However, in a state that expanded Medicaid (Michigan), uninsured cardiac surgery volume decreased by 60% over an 18-month period, with a 70% increase in Medicaidinsured volume. After expansion, postoperative rates of major morbidity significantly decreased in Medicaid patients after adjusting for preoperative risk. The observed decrease in uninsured volume in Michigan aligns with predictions that Medicaid expansion
under the ACA would lower the number of uninsured people by half [6]. During the same time period in Virginia, no changes in Medicaid or uninsured volume were identified. According to the Congressional Budget Office, the number of people enrolled in Medicaid is estimated to increase by 12 million people per year between now and 2024 [16]. These changes in uninsured and Medicaid volume will have significant effects on hospital reimbursements for uncompensated care under the Medicare and Medicaid Disproportionate Share Hospital program. Although lower on average than private insurance reimbursement, Medicaid reimbursement is financially more advantageous for hospitals than providing uncompensated care. Under the ACA, Disproportionate Share Hospital payments are being reduced by up to 75%, meaning hospitals will recoup significantly less of the cost for providing care to the uninsured [6]. Within the cardiac surgery population, Medicaid expansion appears to significantly increase the number of insured patients. In addition, the present study identified a relationship between Medicaid expansion and lower rates of postoperative morbidity, which may also decrease total costs of care. Using the American College of Surgeons National Surgical Quality Improvement Program, Ko and colleagues [17] identified that patients with postoperative complications had higher readmission rates and overall costs of care. When extrapolated to other surgical populations and patient care in general, these changes as a result of Medicaid expansion may have significant financial implications for hospitals. Aside from monetary benefits, patients may benefit from improvements in preoperative health status. No changes in predicted risk scores were identified in Virginia. In Michigan, however, postexpansion Medicaid patients had a significant decrease in PROMM and lower
Ann Thorac Surg 2017;104:1251–8
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
1255
Table 2. Operative Case Specifics Virginia Variable
a
a
Preexpansion
Postexpansion
153
157
p Value
Preexpansion
Postexpansion
558
954
0.61 99 15 19 9 4 7
(64.7) (9.8) (12.4) (5.9) (2.6) (4.6)
105 14 12 5 2 5
(66.9) (8.9) (7.6) (3.2) (1.3) (3.2)
0.61 409 53 29 21 5 41
(73.3) (9.5) (5.2) (3.8) (0.9) (7.3)
716 99 47 32 12 48
(75.1) (10.4) (4.9) (3.4) (1.3) (5.0)
0.59 71 (46.4) 79 (51.6) 3 (2.0)
64 (40.8) 89 (56.7) 4 (2.5)
141 (92.2) 9 (5.9) 3 (2.0)
148 (94.3) 8 (5.1) 1 (0.6)
471
402
0.41 219 (39.3) 326 (58.4) 13 (2.3)
373 (39.1) 547 (57.3) 34 (3.6)
534 (95.7) 21 (3.8) 2 (0.4)
917 (96.1) 36 (3.8) 1 (0.1)
0.56
0.41
468
181
0.96 379 40 17 11 14 10
(80.5) (8.5) (3.6) (2.3) (3.0) (2.1)
319 33 12 8 9 11
(79.4) (8.2) (3.0) (2.0) (2.2) (2.7)
0.55 370 34 15 15 4 30
(79.1) (7.3) (3.2) (3.2) (0.9) (6.4)
146 10 3 6 1 15
(80.7) (5.5) (1.7) (3.3) (0.6) (8.3)
0.15 102 (21.7) 334 (71.2) 33 (7.0)
103 (25.6) 281 (69.9) 18 (4.5)
456 (96.8) 14 (3.0) 1 (0.2)
389 (96.8) 12 (3.0) 1 (0.2)
624
559
0.30 103 (22.0) 341 (72.9) 24 (5.1)
30 (16.6) 142 (78.5) 9 (5.0)
456 (97.4) 11 (2.4) 1 (0.2)
176 (97.2) 5 (2.8) 0 (0)
0.99
0.79
1,026
1,135
0.75 478 55 36 20 18 17
(76.6) (8.8) (5.8) (3.2) (2.9) (2.7)
424 47 24 13 11 16
(75.8) (8.4) (4.3) (2.3) (2.0) (2.9)
0.82 779 87 44 36 9 71
(75.9) (8.5) (4.3) (3.5) (0.9) (6.9)
862 109 50 38 13 63
(75.9) (9.6) (4.4) (3.3) (1.2) (5.6)
0.29 173 (27.8) 413 (66.4) 36 (5.8)
167 (29.9) 370 (66.2) 22 (3.9)
597 (95.7) 23 (3.7) 4 (0.6)
537 (96.1) 20 (3.6) 2 (0.4)
p Value
0.11 322 (31.3) 667 (65.0) 37 (3.6)
403 (35.5) 689 (60.7) 43 (3.8)
990 (96.5) 32 (3.1) 3 (0.3)
1,093 (96.3) 41 (3.6) 1 (0.1)
0.79
0.44
Data are presented as number (%) or as indicated.
AVR ¼ aortic valve replacement; CABG ¼ coronary artery bypass graft operation; mitral valve repair with or without CABG.
MVR ¼ mitral valve replacement;
MV repair CABG ¼
ADULT CARDIAC
Medicaid Patients, No. Case mix CABG only AVR only MVR only AVR þ CABG MVR þ CABG MV repair CABG Case status Elective Urgent Emergent Reoperation status First operation First reoperation Reoperation (2) Uninsured Patients, No. Case mix CABG only AVR only MVR only AVR þ CABG MVR þ CABG MV repair CABG Case status Elective Urgent Emergent Reoperation status First operation First reoperation Reoperation (2) Combined Patients, No. Case mix CABG only AVR only MVR only AVR þ CABG MVR þ CABG MV repair CABG Case status Elective Urgent Emergent Reoperation status First operation First reoperation Reoperation (2)
Michigan
1256
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
Ann Thorac Surg 2017;104:1251–8
Table 3. Preoperative Predicted Risk Scores Virginia a
ADULT CARDIAC
Variable
Medicaid Patients, No. PROMM, % PROM, % Uninsured Patients, No. PROMM, % PROM, % Combined Patients, No. PROMM, % PROM, % a
Michigan
Preexpansion
Postexpansion
p Value
Preexpansion
Postexpansion
p Value
153 14.0 (8.7–24.9) 1.1 (0.6–2.3)
157 13.2 (8.7–21.3) 0.9 (0.6–2.5)
0.59 0.85
558 11.9 (8.1–20.0) 1.0 (0.5–1.9)
954 11.1 (7.7–17.9) 0.8 (0.5–1.6)
0.02 0.06
471 12.4 (8.1–20.9) 0.9 (0.5–1.8)
402 12.0 (8.3–18.6) 0.8 (0.5–1.6)
0.56 0.37
468 10.8 (7.3–18.4) 0.8 (0.5–1.5)
181 13.5 (8.2–21.5) 1.0 (0.6–1.9)
0.006 0.004
624 12.6 (8.2–21.7) 0.9 (0.5–1.9)
559 12.4 (8.4–19.1) 0.9 (0.5–1.8)
0.50 0.46
1,026 11.6 (7.6–18.9) 0.9 (0.5–1.7)
1,135 11.3 (7.8–18.4) 0.9 (0.5–1.7)
0.53 0.78
Data are presented as median (interquartile range).
PROM ¼ predicted risk of mortality;
PROMM ¼ predicted risk of morbidity or mortality.
predicted risk of mortality. It is difficult to determine how Medicaid insurance may affect preoperative optimization from the present study, but the findings may reflect increased access to primary care. Using National Health Interview Survey data, Wherry and Miller [18] identified an association between Medicaid expansion and increased health care utilization and more frequent diagnosis of health problems. In a single-state review of Medicaid expansion, Benitez and colleagues [19] found lower rates of unmet medical needs and fewer people without routine access to health care. Proper diagnosis and medical management of comorbidities may result in lower preoperative risk scores. Surprisingly, no significant differences in the elective, urgent, or emergent case status were identified in either state. This lack of difference in Michigan may indicate that changes observed may not reflect increased access to
preoperative care but rather a shift of younger, betterinformed, low-risk patients from being uninsured to having Medicaid. This shift in relatively healthy patients subsequently reduces the risk pool in the postexpansion Medicaid population. If Medicaid expansion does lead to improved preoperative optimization, then an associated improvement in outcomes would be expected. In the current study, postexpansion Medicaid patients in Michigan had significantly lower rates of postoperative morbidity. For health care providers, reimbursement rates are important but not nearly as important as the effect of policy interventions on actual patient outcomes. In Michigan, postexpansion Medicaid patients were 30% less likely to experience a major postoperative complication (adjusted OR, 0.69). In addition to Medicaid insurance reimbursements and higher Disproportionate Share
Table 4. Morbidity and Mortality Outcomes Virginia a
Variable
Medicaid Patients, No. Major morbidityb 30-day mortality Uninsured Patients, No. Major morbidity 30-day mortality Combined Patients, No. Major morbidity 30-day mortality
Michigan
Preexpansion
Postexpansion
p Value
Preexpansion
Postexpansion
p Value
153 23 (15.0) 4 (2.6)
157 21 (13.4) 3 (1.9)
0.68 0.68
558 102 (18.3) 14 (2.5)
954 126 (13.2) 15 (1.6)
0.008 0.20
471 61 (13.0) 8 (1.7)
402 51 (12.7) 10 (2.5)
0.91 0.41
468 71 (15.2) 5 (1.1)
181 26 (14.4) 6 (3.3)
0.80 0.05
624 84 (13.5) 12 (1.9)
559 72 (12.9) 13 (2.3)
0.77 0.63
1,026 188 (18.3) 19 (1.9)
1,135 171 (15.1) 21 (1.9)
0.04 1.0
a b Data are presented as number (%) or as indicated. Major morbidity includes reoperation, stroke, kidney failure, deep sternal wound infection, and ventilator support >24 hours within 30 days of operation.
Ann Thorac Surg 2017;104:1251–8
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
1257
Table 5. Unadjusted and Adjusted Postexpansion Odds Ratios for Major Morbidity and 30-Day Mortality Virginia Variable
a
Odds Ratio (95% CI)
p Value
Odds Ratio (95% CI)
0.87 (0.46–1.65) 0.73 (0.16–3.30)
0.7 0.7
0.68 (0.52–0.89) 0.62 (0.30–1.30)
0.004 0.20
0.92 (0.47–1.78) 0.63 (0.13–3.02)
0.8 0.6
0.69 (0.51–0.91) 0.61 (0.30–1.27)
0.01 0.19
0.98 (0.66–1.45) 1.48 (0.58–3.78)
0.9 0.4
0.94 (0.58–1.53) 3.18 (0.96–10.54)
0.80 0.06
0.97 (0.63–1.51) 1.30 (0.48–3.57)
0.9 0.6
0.83 (0.49–1.42) 3.94 (0.98–15.86)
0.50 0.05
0.92 (0.70–1.22) 1.16 (0.62–2.18)
0.6 0.7
0.86 (0.79–0.93) 1.14 (0.95–1.37)
<0.001 0.16
0.89 (0.70–1.19) 1.02 (0.53–1.98)
0.4 0.9
0.84 (0.77–0.92) 1.10 (0.91–1.32)
<0.001 0.35
Major morbidity adjusted for predicted risk of morbidity or mortality.
b
p Value
Mortality adjusted for predicted risk of mortality.
CI ¼ confidence interval.
Hospital payments, states that expand Medicaid may see improvements in access to care, preoperative patient optimization, and surgical outcomes. Although the focus was to evaluate the effect of expanding Medicaid, some findings in the uninsured patient population warrant discussion. In Michigan, the uninsured cardiac surgery volume dropped by 60%, but there was no change in Virginia. The Michigan patients who were still uninsured after Medicaid expansion were older, had a higher PROMM, a higher predicted risk of mortality, and a greater 30-day mortality rate. Because Medicaid expansion appears to benefit a significant number of patients, there remains a vulnerable population of uninsured patients who have higher predicted risk scores and rates of morbidity and mortality. An important next step will be to identify patients who did not qualify under traditional Medicaid but gained coverage after expansion and compare them to postexpansion uninsured patients, paying special attention to social determinants of health (ie, education level, employment status, and housing conditions). This analysis will help determine whether the uninsured actually qualify for expanded Medicaid but have barriers to enrollment or whether they are indeed sicker because they cannot access primary care. The current study has several limitations. Considering that insurance status is only one factor related to the delivery of health care, no definitive or causal relationships between Medicaid expansion and outcomes can be
determined; rather, these are inferences that should be used to direct further health policy implementation and research. The data set did not allow for postexpansion Medicaid patients to be separated by whether or not they qualified under the traditional system or only after expansion. This study may not reflect changes nationwide, considering it is an analysis of cardiac surgery patients in only two states with differently sized populations. The findings are limited by the relatively short interval of follow-up and lack of complementary percutaneous coronary intervention data. Considering that Medicaid expansion is a fairly recent policy change, repeat analyses will be necessary to validate these findings and assess the effect on patient access to both surgery and interventional cardiology. Using two statewide databases that collect validated and audited STS cardiac surgery data, the present study identified an association between Medicaid expansion and decreased volume of uninsured cardiac surgery patients, lower preoperative predicted risk scores, and improved postoperative outcomes. Considering that reimbursements for uncompensated care will continue to decline, it appears that Medicaid expansion may improve health care financing and positively affect patient care and outcomes. Continuous evaluation of health policy changes such as Medicaid expansion is necessary to expedite advancements in the delivery of high-quality, cost-effective health care.
ADULT CARDIAC
Medicaid Unadjusted Major morbiditya 30-day mortalityb Adjusted Major morbidity 30-day mortality Uninsured Unadjusted Major morbidity 30-day mortality Adjusted Major morbidity 30-day mortality Combined Unadjusted Major morbidity 30-day mortality Adjusted Major morbidity 30-day mortality
Michigan
ADULT CARDIAC
1258
CHARLES ET AL IMPACT OF MEDICAID EXPANSION
The authors extend special appreciation to Jenna Charles for her invaluable assistance in navigating the intricacies of the Patient Protection and Affordable Care Act. This work was supported by grants UM1-HL-088925 and T32-HL-007849 (I.L.K.) from the National Heart, Lung, and Blood Institute and R01-HS-022535 (D.S.L.) from the Agency for Healthcare Research and Quality. Support for MSTCVS Quality Collaborative is provided by Blue Cross and Blue Shield of Michigan and Blue Care Network as part of the BCBSM Value Partnerships program.
Audio Discussion: Audio of the discussion that followed the presentation of this paper at the STS Annual Meeting can be accessed in the online version of this article [http://dx.doi.org/10.1016/j.athoracsur.2017.03. 079] on http://www.annalsthoracicsurgery.org.
References 1. U.S. Department of Health and Human Services. Key features of the Affordable Care Act by year. Available at http:// www.hhs.gov/healthcare/facts-and-features/key-features-ofaca-by-year/index.html. Accessed August 13, 2016. 2. Medicaid.com. Centers for Medicare & Medicaid Services. Affordable Care Act eligibility, Medicaid expansion. Available at https://www.medicaid.gov/affordable-care-act/ eligibility/index.html. Accessed August 13, 2016. 3. The Henry J. Kaiser Family Foundation. A guide to the Supreme Court’s decision on the ACA’s Medicaid expansion. Available at https://kaiserfamilyfoundation.files.wordpress. com/2013/01/8347.pdf. Accessed August 13, 2016. 4. U.S. Supreme Court. National Federation of Independent Business et al. v. Sebelius, Secretary of Health and Human Services, et al. Available at https://www.supremecourt.gov/ opinions/11pdf/11-393c3a2.pdf. Accessed August 15, 2016. 5. The Henry J. Kaiser Family Foundation. Status of state action on the Medicaid expansion decision. Available at http://kff. org/health-reform/state-indicator/state-activity-aroundexpanding-medicaid-under-the-affordable-care-act/?current Timeframe¼0&sortModel¼%7B%22colId%22:%22Location %22,%22sort%22:%22asc%22%7D. Accessed August 15, 2016. 6. Graves JA. Medicaid expansion opt-outs and uncompensated care. N Engl J Med 2012;367:2365–7.
Ann Thorac Surg 2017;104:1251–8
7. Polanco A, Breglio AM, Itagaki S, Goldstone AB, Chikwe J. Does payer status impact clinical outcomes after cardiac surgery? A propensity analysis. Heart Surg Forum 2012;15: E262–7. 8. Mell MW, Baker LC. Payer status, preoperative surveillance, and rupture of abdominal aortic aneurysms in the US Medicare population. Ann Vasc Surg 2014;28:1378–83. 9. Adepoju L, Wanjiku S, Brown M, et al. Effect of insurance payer status on the surgical treatment of early stage breast cancer: data analysis from a single health system. JAMA Surg 2013;148:570–2. 10. Boxer LK, Dimick JB, Wainess RM, et al. Payer status is related to differences in access and outcomes of abdominal aortic aneurysm repair in the United States. Surgery 2003;134:142–5. 11. Lapar DJ, Bhamidipati CM, Walters DM, et al. Primary payer status affects outcomes for cardiac valve operations. J Am Coll Surg 2011;212:759–67. 12. LaPar DJ, Stukenborg GJ, Guyer RA, et al. Primary payer status is associated with mortality and resource utilization for coronary artery bypass grafting. Circulation 2012;126(11 Suppl 1):S132–9. 13. Jacobs JP, Shahian DM, Prager RL, et al. The Society of Thoracic Surgeons National Database 2016 Annual Report. Ann Thorac Surg 2016;102:1790–7. 14. Ayanian JZ. Michigan’s approach to Medicaid expansion and reform. N Engl J Med 2013;369:1773–5. 15. Virginia Department of Social Services Medicaid Fact Sheet #45. Available at https://www.dss.virginia.gov/files/division/ bp/medical_assistance/forms/all_other/d032-03-0452-03-eng. pdf. Accessed December 14, 2016. 16. Congressional Budget Office. Updated estimates of the insurance coverage provisions of the Affordable Care Act. Available at https://www.cbo.gov/sites/default/files/cbo files/attachments/45010-breakout-AppendixB.pdf. Accessed December 7, 2016. 17. Lawson EH, Hall BL, Louie R, et al. Association between occurrence of a postoperative complication and readmission: implications for quality improvement and cost savings. Ann Surg 2013;258:10–8. 18. Wherry LR, Miller S. Early coverage, access, utilization, and health effects associated with the Affordable Care Act Medicaid expansions: a quasi-experimental study. Ann Intern Med 2016;164:795–803. 19. Benitez JA, Creel L, Jennings J. Kentucky’s Medicaid expansion showing early promise on coverage and access to care. Health Aff (Millwood) 2016;35:528–34.