- Email: [email protected]
Medicaid Participation among Liver Transplant Candidates after the Affordable Care Act Medicaid Expansion
ORIGINAL SCIENTIFIC ARTICLE
Medicaid Participation among Liver Transplant Candidates after the Affordable Care Act Medicaid Expansion Dmitry Tumin, PhD, Eliza W Beal, MD, Khalid Mumtaz, MBBS, MSc, Don Hayes Jr, MD, MS, Joseph D Tobias, MD, Timothy M Pawlik, MD, MPH, PhD, FACS, W Kenneth Washburn, MD, FACS, Sylvester M Black, MD, PhD The 2014 Medicaid expansion in participating states increased insurance coverage among people with chronic health conditions, but its implications for access to surgical care remain unclear. We investigated how Medicaid expansion influenced the insurance status of candidates for liver transplantation (LT) and transplant center payor mix. STUDY DESIGN: Data on LT candidates aged 18 to 64 years, in 2012 to 2013 (pre-expansion) and 2014 to 2015 (post-expansion), were obtained from the United Network for Organ Sharing registry. Change between the 2 periods in the percent of LT candidates using Medicaid was compared between expansion and nonexpansion states. Multivariable logistic regression was used to determine how Medicaid expansion influenced individual LT candidates’ likelihood of using Medicaid insurance. RESULTS: The study included 33,017 LT candidates, of whom 29,666 had complete data for multivariable analysis. Medicaid enrollment increased by 4% after Medicaid expansion in participating states. One-quarter of the transplant centers in these states experienced 10% increase in the proportion of LT candidates using Medicaid insurance. Multivariable analysis confirmed that Medicaid expansion was associated with increased odds of LT candidates using Medicaid insurance (odds ratio 1.49; 95% CI 1.34, 1.66; p < 0.001). However, the absolute number and demographic characteristics of patients listed for LT did not change in Medicaid expansion states during the post-expansion period. CONCLUSIONS: Candidates for LT became more likely to use Medicaid after the 2014 Medicaid expansion policy came into effect. Enactment of this policy did not appear to increase access to LT or address socioeconomic and demographic disparities in access to the LT wait list. (J Am Coll Surg 2017;-:1e8. 2017 by the American College of Surgeons. Published by Elsevier Inc. All rights reserved.) BACKGROUND:
The Patient Protection and Affordable Care Act (ACA), enacted in 2010 and implemented in stages through 2014, profoundly reshaped health insurance provision in the US.1 The optional state-level expansion of eligibility for Medicaid public insurance was an ACA component that achieved significant reductions in the number of Americans without health insurance.1-3 Among Americans with chronic health conditions, Medicaid expansion was linked to a 5% reduction in the uninsured rate.4 This increase in insurance coverage suggests that Medicaid expansion may have increased health care use and facilitated improved access to primary and preventive care.4,5 Yet the benefits of Medicaid expansion for access to specialty and surgical care are less certain.3,6 In the case of solid organ transplantation, a treatment option of last resort for patients with end-stage organ failure, Medicaid
Disclosure Information: Nothing to disclose. Disclaimer: The data reported here have been supplied by the United Network for Organ Sharing as the contractor for the Organ Procurement and Transplantation Network (OPTN). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the OPTN or the US Government. Received February 13, 2017; Revised April 12, 2017; Accepted May 9, 2017. From the Department of Anesthesiology and Pain Medicine (Tumin, Tobias), the Section of Pulmonary Medicine (Hayes), and the Division of Transplantation (Washburn), Nationwide Children’s Hospital; and the Departments of Pediatrics (Tumin, Tobias), Surgery (Beal, Pawlik, Washburn, Black), Internal Medicine (Mumtaz, Hayes), and Anesthesiology (Tobias), the Division of Gastroenterology, Hepatology & Nutrition (Mumtaz), and the Comprehensive Transplant Center (Washburn, Black), The Ohio State University Wexner Medical Center, Columbus, OH. Correspondence address: Dmitry Tumin, PhD, The Ohio State University, Nationwide Children’s Hospital, 700 Children’s Dr, Columbus, OH 43205. email: [email protected]
ª 2017 by the American College of Surgeons. Published by Elsevier Inc. All rights reserved.
1
http://dx.doi.org/10.1016/j.jamcollsurg.2017.05.004 ISSN 1072-7515/17
2
Tumin et al
Medicaid Expansion and Liver Transplantation
Abbreviations and Acronyms
ACA ESLD LT MELD
¼ ¼ ¼ ¼
Patient Protection and Affordable Care Act end-stage liver disease liver transplantation Model for End-stage Liver Disease
expansion has been reported to be associated with increased heart transplant wait listing rates among African Americans with heart failure7; however, it did not increase wait listing rates among patients requiring kidney, liver, or lung transplantation.8 In the setting of liver transplantation (LT), expanded eligibility for Medicaid has complex implications for access to this procedure. On one hand, Medicaid enrollment of previously uninsured patients may broaden access to timely diagnosis of end-stage liver disease (ESLD), referral for LT, and transplant evaluation, potentially reducing disparities in LT wait listing.9,10 On the other hand, new enrollees in Medicaid may encounter existing barriers to being listed and ultimately undergoing LT. Only 1 in 6 patients with ESLD is ever placed on the LT wait list,11 and patients with Medicaid are less likely to be listed for LT than patients with commercial or Medicare insurance.12 Centers may be reluctant to list some patients with Medicaid insurance for LT if they lack appropriate social support or have a history of noncompliance with treatment. In turn, psychosocial limitations associated with Medicaid insurance participation may explain the worse post-transplantation outcomes seen in LT recipients with Medicaid insurance.13,14 As evinced by unchanged LT listing rates after Medicaid expansion,8 an increase in Medicaid enrollment among patients with ESLD may not necessarily lead to greater representation of patients with Medicaid among those listed for LT. Despite the lack of change in the number of patients listed for LT,8 we have recently demonstrated rapid shifts from commercial to Medicaid coverage in a cohort of LT recipients after Medicaid expansion.15 These data suggest that among patients referred for LT, a significant proportion of individuals with commercial insurance might prefer Medicaid coverage if it were available. For example, public insurance that is not tied to work participation may be preferable to employer-linked commercial insurance, given the low rate of employment among LT recipients.16 Therefore, we hypothesized that in states that expanded Medicaid eligibility, more adult LT candidates participated in Medicaid after Medicaid eligibility was expanded in 2014. We evaluated this hypothesis using center-specific and individual-level data from the United Network for Organ Sharing registry.
J Am Coll Surg
METHODS The study was deemed not human subjects research, and therefore exempt from review, by the IRB at Nationwide Children’s Hospital. The United Network for Organ Sharing registry was queried for patients listed for LT between January 2012 and December 2013 (pre-Medicaid expansion era), and between January 2014 and December 2015 (post-Medicaid expansion era).17 Patients younger than 18 years or older than 64 years at listing were excluded from the analysis, in order to focus on the population ineligible for children’s health insurance programs and age-based Medicare coverage. Patients whose insurance status at listing was not reported were also excluded. Patients with missing data on study covariates were excluded from the secondary analysis of individual insurance status. State participation in Medicaid expansion was determined according to whether each state expanded Medicaid eligibility as of January 1 2014.18 States that delayed Medicaid expansion until a later date in 2014 to 2015 (Alaska, Indiana, Michigan, New Hampshire, Pennsylvania) were excluded from the analysis. The annual numbers of patients listed for LT in the age range of 18 to 64 years, and the annual proportions of patients in this age range with Medicaid insurance, were plotted separately for states that did and did not participate in Medicaid expansion. Transplant centers were included in the analysis if the majority (>50%) of patients listed in the pre-Medicaid expansion era were aged 18 to 64 years. Centers averaging <10 LT listings per year of patients aged 18 to 64 years in the pre-Medicaid expansion era were excluded. The primary outcome was center-specific change in the percent of patients listed for LT, who had Medicaid insurance as opposed to any other insurance type. Transplant centers reported the projected primary source of payment for LT at the time of listing using the online UNet system, as previously described.19 The percentage of LT candidates with Medicaid was determined separately in the pre-expansion (2012 to 2013) and post-expansion (2014 to 2015) periods; and the change in the percent of LT candidates with Medicaid between these periods was calculated for each center. Centers’ change in the percent of LT candidates with Medicaid was compared between Medicaid expansion and nonexpansion states using a Wilcoxon rank-sum test, with a 95% CI for the difference in medians calculated using the Hodges-Lehmann method. By comparing this outcome between states that did and did not participate in Medicaid expansion, we aimed to exclude the influences of contemporary changes in liver disease treatment and LT allocation policy, whose effects were shared across all states.
Vol.
-,
No.
-, -
2017
Tumin et al
The secondary outcome was individual patients’ insurance status at listing for LT. This outcome was modeled using difference-in-difference logistic regression, which compared change from the pre- to post-policy period between a group exposed to the policy (patients listed in Medicaid expansion states) and a group not exposed to the policy (patients in nonparticipating states).20 The model included the period of listing (pre- vs post-expansion), place of listing (participating vs nonparticipating state), and the interaction between these 2 factors. The interaction term tested whether change in LT candidates’ likelihood of having Medicaid insurance was significantly different between participating and nonparticipating states. The individual-level model controlled for clinical and demographic factors likely related to participation in Medicaid, including patient age, sex, and race; educational attainment; employment before transplant; rural residence (>50% rural population in the patient’s county of residence as of the 2010 Census); indication for LT; history of dialysis; history of diabetes; and initial Model for End-stage Liver Disease (MELD) score. A centerlevel random intercept was included to account for time-invariant differences among centers in LT candidates’ likelihood of having Medicaid insurance; and standard errors were adjusted for the clustering of patients within listing centers. The area under the receiver operating characteristic curve was used as a summary measure of model fit. To account for patients’ travel to out-ofstate centers, this model was refitted with only patients listed for LT in their state of permanent residence. Data analysis was performed in Stata/IC 13.1 (StataCorp LP), and 2-sided p < 0.05 was considered statistically significant.
RESULTS There were 37,419 patients aged 18 to 64 years listed for LT between 2012 and 2015. Of these cases, 82 were excluded due to missing insurance information; 4,149 were excluded due to being listed in a state that implemented late Medicaid expansion; and 171 were excluded due to being listed in a very small (average <10 listings/ year) or primarily pediatric program. For the primary analysis of center-specific payor mix, the sample included 33,017 listings representing 96 transplant centers. For the secondary multivariable analysis of individual patients’ insurance at the time of listing, the sample was limited to 29,666 cases with complete data on covariates. Study sample characteristics and the extent of missing data on each covariate are summarized in Table 1. Liver transplantation candidate characteristics are compared between pre- and post-expansion periods in eTable 1.
Medicaid Expansion and Liver Transplantation
3
The annual number of LT listings and the proportion involving patients with Medicaid are contrasted between Medicaid expansion and nonexpansion states in Figures 1 and 2. Comparing 2013 to 2014 in Medicaid expansion states, the number of listings involving patients ages 18 to 64 years remained constant (4,739 to 4,732), yet the proportion of patients with Medicaid insurance increased from 21% to 25%, and this level was sustained in 2015 (26%). Despite the rapid and sustained increase in Medicaid insurance use among LT candidates, the number of LT listings in Medicaid expansion states declined from 2014 to 2015 (Fig. 1). Figure 3 illustrates center-specific changes in the payor mix between the preexpansion (2012 to 2013) and post-expansion (2014 to 2015) periods. Among the 53 centers located in Medicaid expansion states, the median center experienced a 4.4% increase in the proportion of LT candidates aged 18 to 64 years, who had Medicaid insurance (interquartile range 0.5%, þ10.3%). By contrast, among 42 centers in nonexpansion states, the median center experienced a 0.8% decline in the proportion of patients with Medicaid in this age range (interquartile range 3.5%, þ2.6%). The Wilcoxon rank-sum test confirmed that centerspecific increases in LT candidates’ Medicaid enrollment were greater in participating states than in nonparticipating states (p ¼ 0.003; 95% CI of difference in medians, 2%, 7%). The individual-level analysis used a difference-indifference approach to estimate the association between Medicaid expansion and LT candidates’ insurance status (Table 2). Consistent with the center-level analysis, this model found that in Medicaid expansion states, LT candidates became more likely to use Medicaid in the 2014 to 2015 period compared with the 2012 to 2013 period (odds ratio 1.49; 95% CI 1.34, 1.66; p < 0.001). The interaction between post-expansion period and state nonparticipation in Medicaid expansion implied that this increase in Medicaid enrollment was completely attenuated in nonparticipating states. In these states, LT candidates’ odds of Medicaid participation did not significantly change between 2012 to 2013 and 2014 to 2015 (odds ratio 1.49 0.61 ¼ 0.91; 95% CI 0.78, 1.05; p ¼ 0.178). Results were similar when restricting the analysis to 24,294 patients listed for LT in their state of permanent residence (eTable 2; odds ratio of Medicaid expansion in participating states 1.49; 95% CI 1.33, 1.65; p < 0.001). Demographic characteristics associated with greater likelihood of Medicaid use (younger age, female sex, non-white race, lower educational attainment, lack of pre-transplant employment, and rural residence) did not substantially change after Medicaid expansion in participating states (eTable 1).
4
Tumin et al
Medicaid Expansion and Liver Transplantation
J Am Coll Surg
Table 1. Characteristics of Adults Aged 18 to 64 Years Listed for Liver Transplantation in 2012 to 2015 (n ¼ 33,017) Variable
Cases missing data, n
State participated in Medicaid expansion, n (%)* Period of listing for liver transplant, n (%) 2012e2013 (pre-expansion) 2014e2015 (post-expansion) Patient age, y, mean (SD) Female patient, n (%) Patient race/ethnicity, n (%) Non-Hispanic white Non-Hispanic black Other Patient educational attainment, n (%) High school or less Some college Four-year college Patient working for income, n (%) Rural residence, n (%) Etiology of liver disease, n (%) Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes, n (%) History of dialysis, n (%) Initial MELD score, mean (SD)
0
Data
18,535 (56)
0
0 0 0
16,633 16,384 53 11,684
(50) (50) (10) (35)
22,537 (68) 3,120 (9) 7,360 (22) 2,077
679 924 0
138 0 6
14,945 8,304 7,691 7,571 3,879
(48) (27) (25) (23) (12)
9,499 1,406 3,081 3,413 7,953 3,297 4,418 8,542 2,703 18.7
(29) (4) (9) (10) (24) (10) (13) (26) (8) (9.7)
Figure 1. Annual number of patients aged 18 to 64 years listed for liver transplantation (LT), by state participation in the 2014 Medicaid expansion.
participation increased among LT candidates; the number of LT listings did not increase, and the demographic profile of LT candidates remained substantively similar to that in the pre-expansion period. Among transplant centers in Medicaid expansion states, one-quarter of centers saw their proportion of LT candidates with Medicaid increase by 10% or more. Taken together, our findings suggest that Medicaid expansion altered the insurance status of LT candidates without enabling broader access to LT. Barriers to LT listing among patients with ESLD remain formidable, and even among patients placed on the wait list, fewer than 1 in 12 ultimately receive LT.11 Patients must receive a timely diagnosis, undergo clinical and psychosocial evaluation, and receive a referral for LT before being placed on the wait list.14 Recent studies suggest that listing for LT may be less likely among patients
*State of listing center. MELD, Model for End-stage Liver Disease.
DISCUSSION For low-income Americans with chronic health conditions, the ACA had a cascading impact on health insurance coverage, access to care, and clinical outcomes.4 The ACA’s impact tended to be greatest for health insurance coverage; changes in health care access attributable to ACA implementation were relatively modest, and the policy impact on health status or clinical outcomes was often absent.1,3 In the case of ESLD, recent evidence suggests that the ACA Medicaid expansion transformed the insurance coverage of patients undergoing LT,15 without increasing access to LT listing.8 This study elaborates on the implications of Medicaid expansion for patients on the LT wait list. In Medicaid expansion states, Medicaid
Figure 2. Annual proportion of patients aged 18 to 64 years listed for liver transplantation (LT) with Medicaid insurance, by state participation in the 2014 Medicaid expansion.
Vol.
-,
No.
-, -
2017
Tumin et al
Figure 3. Center-specific change in proportion of patients aged 18 to 64 years listed for liver transplantation with Medicaid insurance, between 2012 to 2013 and 2014 to 2015, by state participation in the 2014 Medicaid expansion. (A) Centers in Medicaid expansion states. (B) Centers in non-expansion states.
who do not understand the LT evaluation and listing process,21 or live further away from a transplant center.22 Financial instability, and particularly the lack of insurance, is another major obstacle to wait listing for LT.23 In a survey of LT programs, 69% reported encountering inadequate or unstable health insurance as a social barrier to listing in 10% of more of patients undergoing LT evaluation.24 Conceptually, broader provision of public health insurance (Medicaid) might have reduced disparities in access to the wait list by enhancing the transplant candidacy of low-income patients enrolling in Medicaid after a period of un- or under-insurance. Yet, we found that the demographic and socioeconomic profile of LT candidates did not substantially change in Medicaid expansion states after implementation of this policy. Specifically, there were no statistically significant changes in the race/ethnicity, educational attainment, or employment history of LT candidates in these states, while the proportion of LT candidates from rural areas increased by only 1%. Therefore, Medicaid expansion in participating states could not be linked to perceptible change in the demographic profile of patients who gain access to the LT wait list. Although our study adds to the evidence that Medicaid expansion did not increase access to LT, this policy is clearly reshaping the insurance profile of patients referred for transplantation. Patients with ESLD typically experience disruption of paid work that endures long after
Medicaid Expansion and Liver Transplantation
5
recovery from LT,16,25 jeopardizing their access to employer-sponsored commercial insurance policies. Although the ACA included provisions for improving access to individual policies through creation of marketplace exchanges, not all states cooperated in setting up exchanges, and insurers’ participation as well as the prices of policies sold on exchanges have been highly variable. Therefore, it is unsurprising that patients with advanced disease, requiring complex surgical intervention and lifelong costs of post-transplant care, would enroll in a public plan.15 Similarly, in the setting of kidney transplantation, where Medicare insurance is available to all patients, 60% of adult kidney transplant recipients use Medicare to pay for the transplant.26 These findings suggest that high levels of Medicaid uptake are to be expected among patients with ESLD after Medicaid expansion. Increased Medicaid uptake among LT candidates and recipients poses a dilemma for transplant centers,13 which are evaluated on risk-adjusted patient outcomes, and flagged or penalized when these outcomes fall short of expected values.27 Although data on LT outcomes since Medicaid expansion remain limited, analyses of LT outcomes on the eve of Medicaid expansion and in historical cohorts consistently suggest that Medicaid participation is associated with increased risk of graft failure and mortality.13,28 Much of this association may be due to confounding of Medicaid participation with liver disease progression, comorbidities, socioeconomic disadvantage, noncompliance, and other psychosocial characteristics that are the true predictors of worse LT outcomes. This confounding may also explain why patients with Medicaid were less likely to be listed for LT, compared with patients with commercial or Medicare insurance, in a study completed before passage of the ACA.12 Nevertheless, it is plausible that LT recipients with Medicaid insurance will continue exhibiting worse post-transplant outcomes after Medicaid expansion.13 Therefore, revising the risk adjustment of LT quality metrics to include insurance type and other aspects of socioeconomic status would be an especially timely endeavor, considering the increasing proportion of LT candidates using Medicaid, and the geographic disparity in this increase. Our study used both center- and individual-level analyses to establish increased Medicaid participation among LT candidates in Medicaid expansion states, yet our conclusions are constrained by some aspects of the study data. Most importantly, in the absence of a national ESLD registry, true rates of LT listing among patients with ESLD remain elusive.11,12 Furthermore, our coverage of the post-expansion period was limited to 2 years of data. Yet the increase in Medicaid enrollment among LT candidates in expansion states was immediate
6
Tumin et al
J Am Coll Surg
Medicaid Expansion and Liver Transplantation
Table 2. Multivariable Logistic Regression of Medicaid Insurance Coverage at the Time of Listing for Liver Transplantation among Adult Candidates Aged 18 to 64 Years (n ¼ 29,666) Variable*
State participation in Medicaid expansion State expanded Medicaid State did not expand Medicaid Period of listing for liver transplant 2012e2013 (pre-expansion) 2014e2015 (post-expansion) State and period interaction Medicaid non-expansion post-expansion period Patient age, y Female patient Patient race/ethnicity Non-Hispanic white Non-Hispanic black Other Patient educational attainment High school or less Some college Four-year college Patient working for income Rural residence Etiology of liver disease Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes History of dialysis Initial MELD score
Odds ratio
95% CIy
p Valuey
ref. 0.64
0.46, 0.88
0.007
ref. 1.49
1.34, 1.66
<0.001
0.61 0.96 1.30
0.51, 0.73 0.95, 0.96 1.21, 1.40
<0.001 <0.001 <0.001
ref. 1.81 1.80
1.60, 2.05 1.60, 2.03
<0.001 <0.001
ref. 0.59 0.37 0.20 1.17
0.54, 0.32, 0.17, 1.04,
0.64 0.42 0.23 1.32
<0.001 <0.001 <0.001 0.009
ref. 0.91 0.50 0.52 0.97 1.06 0.46 0.87 0.82 1.00
0.77, 0.42, 0.45, 0.88, 0.91, 0.39, 0.80, 0.69, 0.99,
1.08 0.59 0.60 1.07 1.24 0.55 0.94 0.97 1.00
0.280 <0.001 <0.001 0.575 0.460 <0.001 <0.001 0.018 0.671
*Model includes center-level random intercept. Area under receiver operating characteristic curve ¼ 0.76. y Model standard errors adjusted for clustering of patients within listing centers. MELD, Model for End-stage Liver Disease.
in 2014 and sustained in 2015; the total number of LT candidates began declining in Medicaid expansion states during the post-expansion period. The former finding was consistent with previous evidence that newly eligible adults rapidly enrolled in Medicaid after eligibility criteria were expanded.1,2 Still, among the most recent Medicaid enrollees, the process of transplant evaluation and listing may have concluded after the end of our study period. Therefore, monitoring the ACA’s impact on LT access will remain an important component of future research in this area, even as attention turns toward the impact of Medicaid expansion on post-transplant outcomes.9 A further limitation was the lack of information on the timing and reason for Medicaid enrollment
among newly eligible adults. Specifically, we could not delineate which patients enrolled in Medicaid for the express purpose of meeting financial criteria for LT listing. Last, our secondary analysis was limited by missing data on patient characteristics, particularly educational attainment and rural as compared with urban residence. Although these missing data may have biased point estimates in the multivariable analysis, we found consistent results supporting increased Medicaid enrollment among LT candidates across all Medicaid expansion states (Fig. 2), at transplant centers located in Medicaid expansion states (Fig. 3), and among individual LT candidates, according to state of listing (Table 2) or state of residence (eTable 2).
Vol.
-,
No.
-, -
2017
Tumin et al
CONCLUSIONS The 2014 implementation of the ACA Medicaid expansion in participating states significantly increased the proportion of LT candidates using Medicaid insurance. This increase was unequal across transplant centers, with some centers seeing increases of 10 or more points in the percentage of LT candidates aged 18 to 64 years who had Medicaid insurance. Despite significant increases in Medicaid participation among LT candidates in Medicaid expansion states, we found no evidence that Medicaid expansion resulted in broader access to LT wait listing. Future policy or practice interventions may target specific factors contributing to disparities in listing for LT, such as differences in patient understanding of the LT evaluation and listing process, or differences in health care infrastructure investment, that influence patients’ geographic access to transplant centers. Addressing these sources of variability in LT evaluation and listing may increase access to LT in socioeconomically disadvantaged and medically underserved groups. Author Contributions Study conception and design: Tumin, Black Acquisition of data: Tumin, Beal, Hayes Analysis and interpretation of data: Tumin, Beal, Mumtaz, Hayes, Tobias, Pawlik, Washburn, Black Drafting of manuscript: Tumin Critical revision: Beal, Mumtaz, Hayes, Tobias, Pawlik, Washburn, Black REFERENCES 1. French MT, Homer J, Gumus G, Hickling L. Key provisions of the Patient Protection and Affordable Care Act (ACA): a systematic review and presentation of early research findings. Health Serv Res 2016;51:1735e1771. 2. Courtemanche C, Marton J, Ukert B, et al. Early impacts of the Affordable Care Act on health insurance coverage in Medicaid expansion and non-expansion states. J Policy Anal Manage 2017;36:178e210. 3. Sommers BD, Blendon RJ, Orav EJ, Epstein AM. Changes in utilization and health among low-income adults after Medicaid expansion or expanded private insurance. JAMA Intern Med 2016;176:1501e1509. 4. Torres H, Poorman E, Tadepalli U, et al. Coverage and access for Americans with chronic disease under the Affordable Care Act: a quasi-experimental study. Ann Intern Med 2017;166: 472e479. 5. Simon K, Soni A, Cawley J. The impact of health insurance on preventive care and health behaviors: evidence from the first two years of the ACA Medicaid expansions. J Policy Anal Manage 2017;36:390e417. 6. Puls MW. 2015 ACS Governors Survey: Surgeons describe the ACA’s effects on access to care. Bull Am Coll Surg 2016;101: 19e24.
Medicaid Expansion and Liver Transplantation
7
7. Breathett K, Allen LA, Helmkamp L, et al. The Affordable Care Act Medicaid expansion correlated with increased heart transplant listings in African-Americans but not Hispanics or Caucasians. JACC Heart Fail 2017;5:136e147. 8. Oliveira GH, Al-Kindi SG, Simon DI. Implementation of the Affordable Care Act and solid-organ transplantation listings in the United States. JAMA Cardiol 2016;1: 737e738. 9. Schlansky B, Shachar C. Implications of expanded Medicaid eligibility for patient outcomes after liver transplantation: Caveat emptor. Liver Transpl 2016;22: 1062e1064. 10. Mathur AK, Ashby VB, Fuller DS, et al. Variation in access to the liver transplant waiting list in the United States. Transplantation 2014;98:94e99. 11. Goldberg DS, French B, Sahota G, et al. Use of populationbased data to demonstrate how waitlist-based metrics overestimate geographic disparities in access to liver transplant care. Am J Transplant 2016;16:2903e2911. 12. Bryce CL, Angus DC, Arnold RM, et al. Sociodemographic differences in early access to liver transplantation services. Am J Transplant 2009;9:2092e2101. 13. DuBay DA, MacLennan PA, Reed RD, et al. Insurance type and solid organ transplantation outcomes: A historical perspective on how Medicaid expansion might impact transplantation outcomes. J Am Coll Surg 2016;223:611e620. 14. Axelrod DA, Lentine KL. Improving access to liver care across the continuum of care: opportunities and challenges. Am J Transplant 2016;16:2777e2778. 15. Tumin D, Hayes D Jr, Washburn WK, et al. Medicaid enrollment after liver transplantation: Effects of Medicaid expansion. Liver Transpl 2016;22:1075e1084. 16. Huda A, Newcomer R, Harrington C, et al. High rate of unemployment after liver transplantation: analysis of the United Network for Organ Sharing database. Liver Transpl 2012;18: 89e99. 17. United Network for Organ Sharing/Organ Procurement and Transplantation Network Standard Transplant Analysis and Research Database. Available at: http://optn.transplant.hrsa. gov/data/. Accessed March 8, 2016. 18. Kaiser Family Foundation. Status of state action on the Medicaid expansion decision. Available at: http://kff.org/medicaid/stateindicator/state-activity-around-expanding-medicaid-underthe-affordable-care-act/. Accessed November 2, 2015. 19. Dickinson DM, Bryant PC, Williams MC, et al. Transplant data: sources, collection, and caveats. Am J Transplant 2004; 4[Suppl 9]:13e26. 20. Dimick JB, Ryan AM. Methods for evaluating changes in health care policy: the difference-in-differences approach. JAMA 2014;312:2401e2402. 21. Wilder JM, Oloruntoba OO, Muir AJ, Moylan CA. Role of patient factors, preferences, and distrust in health care and access to liver transplantation and organ donation. Liver Transpl 2016;22:895e905. 22. Goldberg DS, French B, Forde KA, et al. Association of distance from a transplant center with access to waitlist placement, receipt of liver transplantation, and survival among US veterans. JAMA 2014;311:1234e1243. 23. Laurentine KA, Bramstedt KA. Too poor for transplant: finance and insurance issues in transplant ethics. Prog Transplant 2010;20:178e185.
8
Tumin et al
Medicaid Expansion and Liver Transplantation
24. Flattau A, Olaywi M, Gaglio PJ, et al. Social barriers to listing for adult liver transplantation: their prevalence and association with program characteristics. Liver Transpl 2011;17:1167e1175. 25. Beal EW, Tumin D, Mumtaz K, et al. Factors contributing to employment patterns after liver transplantation. Clin Transplant 2017 [Epub ahead of print]. 26. Hart A, Smith JM, Skeans MA, et al. OPTN/SRTR 2015 Annual Data Report: Kidney. Am J Transplant 2017;17[Suppl 1]:21e116.
J Am Coll Surg
27. Pelletier RP, Henry ML. Program specific reports: friend or foe? dThe intended and unintended consequences of Scientific Registry of Transplant Recipient program specific reports. Curr Transpl Rep 2014;1:86e90. 28. Yoo HY, Thuluvath PJ. Outcome of liver transplantation in adult recipients: influence of neighborhood income, education, and insurance. Liver Transpl 2004;10: 235e243.
Vol. -,
No. -, -
Variable
2,350 (26) 52 (10) 3,257 (36)
6,064 (65) 779 (8) 2,530 (27)
5,897 (64) 787 (9) 2,478 (27)
4,365 2,410 2,180 2,176 635
(49) (27) (24) (24) (7)
4,294 2,265 2,194 2,171 698
(49) (26) (25) (24) (8)
3,058 344 844 725 2,251 927 1,224 2,383 720 18.1
(33) (4) (9) (8) (24) (10) (13) (26) (8) (9.8)
2,508 370 842 974 2,463 848 1,157 2,364 816 18.7
(27) (4) (9) (11) (27) (9) (13) (26) (9) (10.2)
<0.001 <0.001 0.094 0.775
Non-expansion states 2012e2013 (n ¼ 7,260) 2014e2015 (n ¼ 7,222)
969 (13) 53 (9) 2,604 (36)
900 (12) 53 (10) 2,601 (36)
5,376 (75) 775 (11) 1,109 (15)
5,200 (72) 779 (11) 1,243 (17)
3,118 1,839 1,627 1,670 1,280
(47) (27) (25) (24) (18)
3,168 1,790 1,690 1,554 1,266
(48) (27) (25) (22) (18)
2,086 361 694 744 1,572 806 997 1,811 515 18.6
(29) (5) (10) (10) (22) (11) (14) (25) (7) (9.0)
1,797 331 701 970 1,667 716 1,040 1,984 652 19.4
(25) (5) (10) (13) (23) (10) (14) (28) (9) (9.7)
0.244
*Unpaired t-test for continuous variables and chi-square test for categorical variables. MELD, Model for End-stage Liver Disease.
0.548 0.019 <0.001
0.547 0.002 <0.001
p Value*
0.112 0.181 0.854 0.005
0.378
0.017 0.691 <0.001
0.001 <0.001 <0.001
Medicaid Expansion and Liver Transplantation
1,956 (21) 53 (10) 3,222 (34)
p Value*
Tumin et al
Medicaid insurance, n (%) Patient age, y, mean (SD) Female patient, n (%) Patient race/ethnicity, n (%) Non-Hispanic white Non-Hispanic black Other Patient educational attainment, n (%) High school or less Some college Four-year college Patient working for income Rural residence Etiology of liver disease, n (%) Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes, n (%) History of dialysis, n (%) Initial MELD score, mean (SD)
Medicaid expansion states 2012e2013 (n ¼ 9,373) 2014e2015 (n¼ 9,162)
2017
eTable 1. Characteristics of Adults Aged 18 to 64 Years Listed for Liver Transplantation before (2012 to 2013) and after (2014 to 2015) Medicaid Expansion, According to State Participation in Medicaid Expansion (n ¼ 33,017)
8.e1
8.e2
Tumin et al
J Am Coll Surg
Medicaid Expansion and Liver Transplantation
eTable 2. Multivariable Logistic Regression of Medicaid Insurance Coverage at the Time of Listing for Liver Transplantation among Adult Candidates Aged 18 to 64 Years Listed in Their State of Permanent Residence (n ¼ 24,294) Variable*
State participation in Medicaid expansion State expanded Medicaid State did not expand Medicaid Period of listing for liver transplant 2012e2013 (pre-expansion) 2014e2015 (post-expansion) State and period interaction Medicaid non-expansion post-expansion period Patient age, y Female patient Patient race/ethnicity Non-Hispanic white Non-Hispanic black Other Patient educational attainment High school or less Some college Four-year college Patient working for income Rural residence Etiology of liver disease Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes History of dialysis Initial MELD score
Odds ratio
95% CIy
ref. 0.63
0.46, 0.86
0.004
ref. 1.49
1.33, 1.65
<0.001
0.60 0.96 1.32
0.49, 0.72 0.95, 0.97 1.23, 1.42
<0.001 <0.001 <0.001
ref. 1.73 1.80
1.54, 1.94 1.60, 2.02
<0.001 <0.001
ref. 0.59 0.39 0.20 1.15
0.54, 0.34, 0.17, 1.02,
0.65 0.44 0.23 1.31
<0.001 <0.001 <0.001 0.027
ref. 0.96 0.53 0.53 0.98 1.03 0.47 0.87 0.80 1.00
0.81, 0.45, 0.46, 0.88, 0.88, 0.39, 0.80, 0.67, 0.99,
1.14 0.63 0.61 1.08 1.21 0.57 0.95 0.96 1.00
0.640 <0.001 <0.001 0.663 0.732 <0.001 <0.001 0.015 0.299
*Model includes center-level random intercept. Area under receiver operating characteristic curve ¼ 0.75. y Model standard errors adjusted for clustering of patients within listing centers. MELD, Model for End-stage Liver Disease.
p Valuey
Medicaid Participation among Liver Transplant Candidates after the Affordable Care Act Medicaid Expansion Dmitry Tumin, PhD, Eliza W Beal, MD, Khalid Mumtaz, MBBS, MSc, Don Hayes Jr, MD, MS, Joseph D Tobias, MD, Timothy M Pawlik, MD, MPH, PhD, FACS, W Kenneth Washburn, MD, FACS, Sylvester M Black, MD, PhD The 2014 Medicaid expansion in participating states increased insurance coverage among people with chronic health conditions, but its implications for access to surgical care remain unclear. We investigated how Medicaid expansion influenced the insurance status of candidates for liver transplantation (LT) and transplant center payor mix. STUDY DESIGN: Data on LT candidates aged 18 to 64 years, in 2012 to 2013 (pre-expansion) and 2014 to 2015 (post-expansion), were obtained from the United Network for Organ Sharing registry. Change between the 2 periods in the percent of LT candidates using Medicaid was compared between expansion and nonexpansion states. Multivariable logistic regression was used to determine how Medicaid expansion influenced individual LT candidates’ likelihood of using Medicaid insurance. RESULTS: The study included 33,017 LT candidates, of whom 29,666 had complete data for multivariable analysis. Medicaid enrollment increased by 4% after Medicaid expansion in participating states. One-quarter of the transplant centers in these states experienced 10% increase in the proportion of LT candidates using Medicaid insurance. Multivariable analysis confirmed that Medicaid expansion was associated with increased odds of LT candidates using Medicaid insurance (odds ratio 1.49; 95% CI 1.34, 1.66; p < 0.001). However, the absolute number and demographic characteristics of patients listed for LT did not change in Medicaid expansion states during the post-expansion period. CONCLUSIONS: Candidates for LT became more likely to use Medicaid after the 2014 Medicaid expansion policy came into effect. Enactment of this policy did not appear to increase access to LT or address socioeconomic and demographic disparities in access to the LT wait list. (J Am Coll Surg 2017;-:1e8. 2017 by the American College of Surgeons. Published by Elsevier Inc. All rights reserved.) BACKGROUND:
The Patient Protection and Affordable Care Act (ACA), enacted in 2010 and implemented in stages through 2014, profoundly reshaped health insurance provision in the US.1 The optional state-level expansion of eligibility for Medicaid public insurance was an ACA component that achieved significant reductions in the number of Americans without health insurance.1-3 Among Americans with chronic health conditions, Medicaid expansion was linked to a 5% reduction in the uninsured rate.4 This increase in insurance coverage suggests that Medicaid expansion may have increased health care use and facilitated improved access to primary and preventive care.4,5 Yet the benefits of Medicaid expansion for access to specialty and surgical care are less certain.3,6 In the case of solid organ transplantation, a treatment option of last resort for patients with end-stage organ failure, Medicaid
Disclosure Information: Nothing to disclose. Disclaimer: The data reported here have been supplied by the United Network for Organ Sharing as the contractor for the Organ Procurement and Transplantation Network (OPTN). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the OPTN or the US Government. Received February 13, 2017; Revised April 12, 2017; Accepted May 9, 2017. From the Department of Anesthesiology and Pain Medicine (Tumin, Tobias), the Section of Pulmonary Medicine (Hayes), and the Division of Transplantation (Washburn), Nationwide Children’s Hospital; and the Departments of Pediatrics (Tumin, Tobias), Surgery (Beal, Pawlik, Washburn, Black), Internal Medicine (Mumtaz, Hayes), and Anesthesiology (Tobias), the Division of Gastroenterology, Hepatology & Nutrition (Mumtaz), and the Comprehensive Transplant Center (Washburn, Black), The Ohio State University Wexner Medical Center, Columbus, OH. Correspondence address: Dmitry Tumin, PhD, The Ohio State University, Nationwide Children’s Hospital, 700 Children’s Dr, Columbus, OH 43205. email: [email protected]
ª 2017 by the American College of Surgeons. Published by Elsevier Inc. All rights reserved.
1
http://dx.doi.org/10.1016/j.jamcollsurg.2017.05.004 ISSN 1072-7515/17
2
Tumin et al
Medicaid Expansion and Liver Transplantation
Abbreviations and Acronyms
ACA ESLD LT MELD
¼ ¼ ¼ ¼
Patient Protection and Affordable Care Act end-stage liver disease liver transplantation Model for End-stage Liver Disease
expansion has been reported to be associated with increased heart transplant wait listing rates among African Americans with heart failure7; however, it did not increase wait listing rates among patients requiring kidney, liver, or lung transplantation.8 In the setting of liver transplantation (LT), expanded eligibility for Medicaid has complex implications for access to this procedure. On one hand, Medicaid enrollment of previously uninsured patients may broaden access to timely diagnosis of end-stage liver disease (ESLD), referral for LT, and transplant evaluation, potentially reducing disparities in LT wait listing.9,10 On the other hand, new enrollees in Medicaid may encounter existing barriers to being listed and ultimately undergoing LT. Only 1 in 6 patients with ESLD is ever placed on the LT wait list,11 and patients with Medicaid are less likely to be listed for LT than patients with commercial or Medicare insurance.12 Centers may be reluctant to list some patients with Medicaid insurance for LT if they lack appropriate social support or have a history of noncompliance with treatment. In turn, psychosocial limitations associated with Medicaid insurance participation may explain the worse post-transplantation outcomes seen in LT recipients with Medicaid insurance.13,14 As evinced by unchanged LT listing rates after Medicaid expansion,8 an increase in Medicaid enrollment among patients with ESLD may not necessarily lead to greater representation of patients with Medicaid among those listed for LT. Despite the lack of change in the number of patients listed for LT,8 we have recently demonstrated rapid shifts from commercial to Medicaid coverage in a cohort of LT recipients after Medicaid expansion.15 These data suggest that among patients referred for LT, a significant proportion of individuals with commercial insurance might prefer Medicaid coverage if it were available. For example, public insurance that is not tied to work participation may be preferable to employer-linked commercial insurance, given the low rate of employment among LT recipients.16 Therefore, we hypothesized that in states that expanded Medicaid eligibility, more adult LT candidates participated in Medicaid after Medicaid eligibility was expanded in 2014. We evaluated this hypothesis using center-specific and individual-level data from the United Network for Organ Sharing registry.
J Am Coll Surg
METHODS The study was deemed not human subjects research, and therefore exempt from review, by the IRB at Nationwide Children’s Hospital. The United Network for Organ Sharing registry was queried for patients listed for LT between January 2012 and December 2013 (pre-Medicaid expansion era), and between January 2014 and December 2015 (post-Medicaid expansion era).17 Patients younger than 18 years or older than 64 years at listing were excluded from the analysis, in order to focus on the population ineligible for children’s health insurance programs and age-based Medicare coverage. Patients whose insurance status at listing was not reported were also excluded. Patients with missing data on study covariates were excluded from the secondary analysis of individual insurance status. State participation in Medicaid expansion was determined according to whether each state expanded Medicaid eligibility as of January 1 2014.18 States that delayed Medicaid expansion until a later date in 2014 to 2015 (Alaska, Indiana, Michigan, New Hampshire, Pennsylvania) were excluded from the analysis. The annual numbers of patients listed for LT in the age range of 18 to 64 years, and the annual proportions of patients in this age range with Medicaid insurance, were plotted separately for states that did and did not participate in Medicaid expansion. Transplant centers were included in the analysis if the majority (>50%) of patients listed in the pre-Medicaid expansion era were aged 18 to 64 years. Centers averaging <10 LT listings per year of patients aged 18 to 64 years in the pre-Medicaid expansion era were excluded. The primary outcome was center-specific change in the percent of patients listed for LT, who had Medicaid insurance as opposed to any other insurance type. Transplant centers reported the projected primary source of payment for LT at the time of listing using the online UNet system, as previously described.19 The percentage of LT candidates with Medicaid was determined separately in the pre-expansion (2012 to 2013) and post-expansion (2014 to 2015) periods; and the change in the percent of LT candidates with Medicaid between these periods was calculated for each center. Centers’ change in the percent of LT candidates with Medicaid was compared between Medicaid expansion and nonexpansion states using a Wilcoxon rank-sum test, with a 95% CI for the difference in medians calculated using the Hodges-Lehmann method. By comparing this outcome between states that did and did not participate in Medicaid expansion, we aimed to exclude the influences of contemporary changes in liver disease treatment and LT allocation policy, whose effects were shared across all states.
Vol.
-,
No.
-, -
2017
Tumin et al
The secondary outcome was individual patients’ insurance status at listing for LT. This outcome was modeled using difference-in-difference logistic regression, which compared change from the pre- to post-policy period between a group exposed to the policy (patients listed in Medicaid expansion states) and a group not exposed to the policy (patients in nonparticipating states).20 The model included the period of listing (pre- vs post-expansion), place of listing (participating vs nonparticipating state), and the interaction between these 2 factors. The interaction term tested whether change in LT candidates’ likelihood of having Medicaid insurance was significantly different between participating and nonparticipating states. The individual-level model controlled for clinical and demographic factors likely related to participation in Medicaid, including patient age, sex, and race; educational attainment; employment before transplant; rural residence (>50% rural population in the patient’s county of residence as of the 2010 Census); indication for LT; history of dialysis; history of diabetes; and initial Model for End-stage Liver Disease (MELD) score. A centerlevel random intercept was included to account for time-invariant differences among centers in LT candidates’ likelihood of having Medicaid insurance; and standard errors were adjusted for the clustering of patients within listing centers. The area under the receiver operating characteristic curve was used as a summary measure of model fit. To account for patients’ travel to out-ofstate centers, this model was refitted with only patients listed for LT in their state of permanent residence. Data analysis was performed in Stata/IC 13.1 (StataCorp LP), and 2-sided p < 0.05 was considered statistically significant.
RESULTS There were 37,419 patients aged 18 to 64 years listed for LT between 2012 and 2015. Of these cases, 82 were excluded due to missing insurance information; 4,149 were excluded due to being listed in a state that implemented late Medicaid expansion; and 171 were excluded due to being listed in a very small (average <10 listings/ year) or primarily pediatric program. For the primary analysis of center-specific payor mix, the sample included 33,017 listings representing 96 transplant centers. For the secondary multivariable analysis of individual patients’ insurance at the time of listing, the sample was limited to 29,666 cases with complete data on covariates. Study sample characteristics and the extent of missing data on each covariate are summarized in Table 1. Liver transplantation candidate characteristics are compared between pre- and post-expansion periods in eTable 1.
Medicaid Expansion and Liver Transplantation
3
The annual number of LT listings and the proportion involving patients with Medicaid are contrasted between Medicaid expansion and nonexpansion states in Figures 1 and 2. Comparing 2013 to 2014 in Medicaid expansion states, the number of listings involving patients ages 18 to 64 years remained constant (4,739 to 4,732), yet the proportion of patients with Medicaid insurance increased from 21% to 25%, and this level was sustained in 2015 (26%). Despite the rapid and sustained increase in Medicaid insurance use among LT candidates, the number of LT listings in Medicaid expansion states declined from 2014 to 2015 (Fig. 1). Figure 3 illustrates center-specific changes in the payor mix between the preexpansion (2012 to 2013) and post-expansion (2014 to 2015) periods. Among the 53 centers located in Medicaid expansion states, the median center experienced a 4.4% increase in the proportion of LT candidates aged 18 to 64 years, who had Medicaid insurance (interquartile range 0.5%, þ10.3%). By contrast, among 42 centers in nonexpansion states, the median center experienced a 0.8% decline in the proportion of patients with Medicaid in this age range (interquartile range 3.5%, þ2.6%). The Wilcoxon rank-sum test confirmed that centerspecific increases in LT candidates’ Medicaid enrollment were greater in participating states than in nonparticipating states (p ¼ 0.003; 95% CI of difference in medians, 2%, 7%). The individual-level analysis used a difference-indifference approach to estimate the association between Medicaid expansion and LT candidates’ insurance status (Table 2). Consistent with the center-level analysis, this model found that in Medicaid expansion states, LT candidates became more likely to use Medicaid in the 2014 to 2015 period compared with the 2012 to 2013 period (odds ratio 1.49; 95% CI 1.34, 1.66; p < 0.001). The interaction between post-expansion period and state nonparticipation in Medicaid expansion implied that this increase in Medicaid enrollment was completely attenuated in nonparticipating states. In these states, LT candidates’ odds of Medicaid participation did not significantly change between 2012 to 2013 and 2014 to 2015 (odds ratio 1.49 0.61 ¼ 0.91; 95% CI 0.78, 1.05; p ¼ 0.178). Results were similar when restricting the analysis to 24,294 patients listed for LT in their state of permanent residence (eTable 2; odds ratio of Medicaid expansion in participating states 1.49; 95% CI 1.33, 1.65; p < 0.001). Demographic characteristics associated with greater likelihood of Medicaid use (younger age, female sex, non-white race, lower educational attainment, lack of pre-transplant employment, and rural residence) did not substantially change after Medicaid expansion in participating states (eTable 1).
4
Tumin et al
Medicaid Expansion and Liver Transplantation
J Am Coll Surg
Table 1. Characteristics of Adults Aged 18 to 64 Years Listed for Liver Transplantation in 2012 to 2015 (n ¼ 33,017) Variable
Cases missing data, n
State participated in Medicaid expansion, n (%)* Period of listing for liver transplant, n (%) 2012e2013 (pre-expansion) 2014e2015 (post-expansion) Patient age, y, mean (SD) Female patient, n (%) Patient race/ethnicity, n (%) Non-Hispanic white Non-Hispanic black Other Patient educational attainment, n (%) High school or less Some college Four-year college Patient working for income, n (%) Rural residence, n (%) Etiology of liver disease, n (%) Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes, n (%) History of dialysis, n (%) Initial MELD score, mean (SD)
0
Data
18,535 (56)
0
0 0 0
16,633 16,384 53 11,684
(50) (50) (10) (35)
22,537 (68) 3,120 (9) 7,360 (22) 2,077
679 924 0
138 0 6
14,945 8,304 7,691 7,571 3,879
(48) (27) (25) (23) (12)
9,499 1,406 3,081 3,413 7,953 3,297 4,418 8,542 2,703 18.7
(29) (4) (9) (10) (24) (10) (13) (26) (8) (9.7)
Figure 1. Annual number of patients aged 18 to 64 years listed for liver transplantation (LT), by state participation in the 2014 Medicaid expansion.
participation increased among LT candidates; the number of LT listings did not increase, and the demographic profile of LT candidates remained substantively similar to that in the pre-expansion period. Among transplant centers in Medicaid expansion states, one-quarter of centers saw their proportion of LT candidates with Medicaid increase by 10% or more. Taken together, our findings suggest that Medicaid expansion altered the insurance status of LT candidates without enabling broader access to LT. Barriers to LT listing among patients with ESLD remain formidable, and even among patients placed on the wait list, fewer than 1 in 12 ultimately receive LT.11 Patients must receive a timely diagnosis, undergo clinical and psychosocial evaluation, and receive a referral for LT before being placed on the wait list.14 Recent studies suggest that listing for LT may be less likely among patients
*State of listing center. MELD, Model for End-stage Liver Disease.
DISCUSSION For low-income Americans with chronic health conditions, the ACA had a cascading impact on health insurance coverage, access to care, and clinical outcomes.4 The ACA’s impact tended to be greatest for health insurance coverage; changes in health care access attributable to ACA implementation were relatively modest, and the policy impact on health status or clinical outcomes was often absent.1,3 In the case of ESLD, recent evidence suggests that the ACA Medicaid expansion transformed the insurance coverage of patients undergoing LT,15 without increasing access to LT listing.8 This study elaborates on the implications of Medicaid expansion for patients on the LT wait list. In Medicaid expansion states, Medicaid
Figure 2. Annual proportion of patients aged 18 to 64 years listed for liver transplantation (LT) with Medicaid insurance, by state participation in the 2014 Medicaid expansion.
Vol.
-,
No.
-, -
2017
Tumin et al
Figure 3. Center-specific change in proportion of patients aged 18 to 64 years listed for liver transplantation with Medicaid insurance, between 2012 to 2013 and 2014 to 2015, by state participation in the 2014 Medicaid expansion. (A) Centers in Medicaid expansion states. (B) Centers in non-expansion states.
who do not understand the LT evaluation and listing process,21 or live further away from a transplant center.22 Financial instability, and particularly the lack of insurance, is another major obstacle to wait listing for LT.23 In a survey of LT programs, 69% reported encountering inadequate or unstable health insurance as a social barrier to listing in 10% of more of patients undergoing LT evaluation.24 Conceptually, broader provision of public health insurance (Medicaid) might have reduced disparities in access to the wait list by enhancing the transplant candidacy of low-income patients enrolling in Medicaid after a period of un- or under-insurance. Yet, we found that the demographic and socioeconomic profile of LT candidates did not substantially change in Medicaid expansion states after implementation of this policy. Specifically, there were no statistically significant changes in the race/ethnicity, educational attainment, or employment history of LT candidates in these states, while the proportion of LT candidates from rural areas increased by only 1%. Therefore, Medicaid expansion in participating states could not be linked to perceptible change in the demographic profile of patients who gain access to the LT wait list. Although our study adds to the evidence that Medicaid expansion did not increase access to LT, this policy is clearly reshaping the insurance profile of patients referred for transplantation. Patients with ESLD typically experience disruption of paid work that endures long after
Medicaid Expansion and Liver Transplantation
5
recovery from LT,16,25 jeopardizing their access to employer-sponsored commercial insurance policies. Although the ACA included provisions for improving access to individual policies through creation of marketplace exchanges, not all states cooperated in setting up exchanges, and insurers’ participation as well as the prices of policies sold on exchanges have been highly variable. Therefore, it is unsurprising that patients with advanced disease, requiring complex surgical intervention and lifelong costs of post-transplant care, would enroll in a public plan.15 Similarly, in the setting of kidney transplantation, where Medicare insurance is available to all patients, 60% of adult kidney transplant recipients use Medicare to pay for the transplant.26 These findings suggest that high levels of Medicaid uptake are to be expected among patients with ESLD after Medicaid expansion. Increased Medicaid uptake among LT candidates and recipients poses a dilemma for transplant centers,13 which are evaluated on risk-adjusted patient outcomes, and flagged or penalized when these outcomes fall short of expected values.27 Although data on LT outcomes since Medicaid expansion remain limited, analyses of LT outcomes on the eve of Medicaid expansion and in historical cohorts consistently suggest that Medicaid participation is associated with increased risk of graft failure and mortality.13,28 Much of this association may be due to confounding of Medicaid participation with liver disease progression, comorbidities, socioeconomic disadvantage, noncompliance, and other psychosocial characteristics that are the true predictors of worse LT outcomes. This confounding may also explain why patients with Medicaid were less likely to be listed for LT, compared with patients with commercial or Medicare insurance, in a study completed before passage of the ACA.12 Nevertheless, it is plausible that LT recipients with Medicaid insurance will continue exhibiting worse post-transplant outcomes after Medicaid expansion.13 Therefore, revising the risk adjustment of LT quality metrics to include insurance type and other aspects of socioeconomic status would be an especially timely endeavor, considering the increasing proportion of LT candidates using Medicaid, and the geographic disparity in this increase. Our study used both center- and individual-level analyses to establish increased Medicaid participation among LT candidates in Medicaid expansion states, yet our conclusions are constrained by some aspects of the study data. Most importantly, in the absence of a national ESLD registry, true rates of LT listing among patients with ESLD remain elusive.11,12 Furthermore, our coverage of the post-expansion period was limited to 2 years of data. Yet the increase in Medicaid enrollment among LT candidates in expansion states was immediate
6
Tumin et al
J Am Coll Surg
Medicaid Expansion and Liver Transplantation
Table 2. Multivariable Logistic Regression of Medicaid Insurance Coverage at the Time of Listing for Liver Transplantation among Adult Candidates Aged 18 to 64 Years (n ¼ 29,666) Variable*
State participation in Medicaid expansion State expanded Medicaid State did not expand Medicaid Period of listing for liver transplant 2012e2013 (pre-expansion) 2014e2015 (post-expansion) State and period interaction Medicaid non-expansion post-expansion period Patient age, y Female patient Patient race/ethnicity Non-Hispanic white Non-Hispanic black Other Patient educational attainment High school or less Some college Four-year college Patient working for income Rural residence Etiology of liver disease Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes History of dialysis Initial MELD score
Odds ratio
95% CIy
p Valuey
ref. 0.64
0.46, 0.88
0.007
ref. 1.49
1.34, 1.66
<0.001
0.61 0.96 1.30
0.51, 0.73 0.95, 0.96 1.21, 1.40
<0.001 <0.001 <0.001
ref. 1.81 1.80
1.60, 2.05 1.60, 2.03
<0.001 <0.001
ref. 0.59 0.37 0.20 1.17
0.54, 0.32, 0.17, 1.04,
0.64 0.42 0.23 1.32
<0.001 <0.001 <0.001 0.009
ref. 0.91 0.50 0.52 0.97 1.06 0.46 0.87 0.82 1.00
0.77, 0.42, 0.45, 0.88, 0.91, 0.39, 0.80, 0.69, 0.99,
1.08 0.59 0.60 1.07 1.24 0.55 0.94 0.97 1.00
0.280 <0.001 <0.001 0.575 0.460 <0.001 <0.001 0.018 0.671
*Model includes center-level random intercept. Area under receiver operating characteristic curve ¼ 0.76. y Model standard errors adjusted for clustering of patients within listing centers. MELD, Model for End-stage Liver Disease.
in 2014 and sustained in 2015; the total number of LT candidates began declining in Medicaid expansion states during the post-expansion period. The former finding was consistent with previous evidence that newly eligible adults rapidly enrolled in Medicaid after eligibility criteria were expanded.1,2 Still, among the most recent Medicaid enrollees, the process of transplant evaluation and listing may have concluded after the end of our study period. Therefore, monitoring the ACA’s impact on LT access will remain an important component of future research in this area, even as attention turns toward the impact of Medicaid expansion on post-transplant outcomes.9 A further limitation was the lack of information on the timing and reason for Medicaid enrollment
among newly eligible adults. Specifically, we could not delineate which patients enrolled in Medicaid for the express purpose of meeting financial criteria for LT listing. Last, our secondary analysis was limited by missing data on patient characteristics, particularly educational attainment and rural as compared with urban residence. Although these missing data may have biased point estimates in the multivariable analysis, we found consistent results supporting increased Medicaid enrollment among LT candidates across all Medicaid expansion states (Fig. 2), at transplant centers located in Medicaid expansion states (Fig. 3), and among individual LT candidates, according to state of listing (Table 2) or state of residence (eTable 2).
Vol.
-,
No.
-, -
2017
Tumin et al
CONCLUSIONS The 2014 implementation of the ACA Medicaid expansion in participating states significantly increased the proportion of LT candidates using Medicaid insurance. This increase was unequal across transplant centers, with some centers seeing increases of 10 or more points in the percentage of LT candidates aged 18 to 64 years who had Medicaid insurance. Despite significant increases in Medicaid participation among LT candidates in Medicaid expansion states, we found no evidence that Medicaid expansion resulted in broader access to LT wait listing. Future policy or practice interventions may target specific factors contributing to disparities in listing for LT, such as differences in patient understanding of the LT evaluation and listing process, or differences in health care infrastructure investment, that influence patients’ geographic access to transplant centers. Addressing these sources of variability in LT evaluation and listing may increase access to LT in socioeconomically disadvantaged and medically underserved groups. Author Contributions Study conception and design: Tumin, Black Acquisition of data: Tumin, Beal, Hayes Analysis and interpretation of data: Tumin, Beal, Mumtaz, Hayes, Tobias, Pawlik, Washburn, Black Drafting of manuscript: Tumin Critical revision: Beal, Mumtaz, Hayes, Tobias, Pawlik, Washburn, Black REFERENCES 1. French MT, Homer J, Gumus G, Hickling L. Key provisions of the Patient Protection and Affordable Care Act (ACA): a systematic review and presentation of early research findings. Health Serv Res 2016;51:1735e1771. 2. Courtemanche C, Marton J, Ukert B, et al. Early impacts of the Affordable Care Act on health insurance coverage in Medicaid expansion and non-expansion states. J Policy Anal Manage 2017;36:178e210. 3. Sommers BD, Blendon RJ, Orav EJ, Epstein AM. Changes in utilization and health among low-income adults after Medicaid expansion or expanded private insurance. JAMA Intern Med 2016;176:1501e1509. 4. Torres H, Poorman E, Tadepalli U, et al. Coverage and access for Americans with chronic disease under the Affordable Care Act: a quasi-experimental study. Ann Intern Med 2017;166: 472e479. 5. Simon K, Soni A, Cawley J. The impact of health insurance on preventive care and health behaviors: evidence from the first two years of the ACA Medicaid expansions. J Policy Anal Manage 2017;36:390e417. 6. Puls MW. 2015 ACS Governors Survey: Surgeons describe the ACA’s effects on access to care. Bull Am Coll Surg 2016;101: 19e24.
Medicaid Expansion and Liver Transplantation
7
7. Breathett K, Allen LA, Helmkamp L, et al. The Affordable Care Act Medicaid expansion correlated with increased heart transplant listings in African-Americans but not Hispanics or Caucasians. JACC Heart Fail 2017;5:136e147. 8. Oliveira GH, Al-Kindi SG, Simon DI. Implementation of the Affordable Care Act and solid-organ transplantation listings in the United States. JAMA Cardiol 2016;1: 737e738. 9. Schlansky B, Shachar C. Implications of expanded Medicaid eligibility for patient outcomes after liver transplantation: Caveat emptor. Liver Transpl 2016;22: 1062e1064. 10. Mathur AK, Ashby VB, Fuller DS, et al. Variation in access to the liver transplant waiting list in the United States. Transplantation 2014;98:94e99. 11. Goldberg DS, French B, Sahota G, et al. Use of populationbased data to demonstrate how waitlist-based metrics overestimate geographic disparities in access to liver transplant care. Am J Transplant 2016;16:2903e2911. 12. Bryce CL, Angus DC, Arnold RM, et al. Sociodemographic differences in early access to liver transplantation services. Am J Transplant 2009;9:2092e2101. 13. DuBay DA, MacLennan PA, Reed RD, et al. Insurance type and solid organ transplantation outcomes: A historical perspective on how Medicaid expansion might impact transplantation outcomes. J Am Coll Surg 2016;223:611e620. 14. Axelrod DA, Lentine KL. Improving access to liver care across the continuum of care: opportunities and challenges. Am J Transplant 2016;16:2777e2778. 15. Tumin D, Hayes D Jr, Washburn WK, et al. Medicaid enrollment after liver transplantation: Effects of Medicaid expansion. Liver Transpl 2016;22:1075e1084. 16. Huda A, Newcomer R, Harrington C, et al. High rate of unemployment after liver transplantation: analysis of the United Network for Organ Sharing database. Liver Transpl 2012;18: 89e99. 17. United Network for Organ Sharing/Organ Procurement and Transplantation Network Standard Transplant Analysis and Research Database. Available at: http://optn.transplant.hrsa. gov/data/. Accessed March 8, 2016. 18. Kaiser Family Foundation. Status of state action on the Medicaid expansion decision. Available at: http://kff.org/medicaid/stateindicator/state-activity-around-expanding-medicaid-underthe-affordable-care-act/. Accessed November 2, 2015. 19. Dickinson DM, Bryant PC, Williams MC, et al. Transplant data: sources, collection, and caveats. Am J Transplant 2004; 4[Suppl 9]:13e26. 20. Dimick JB, Ryan AM. Methods for evaluating changes in health care policy: the difference-in-differences approach. JAMA 2014;312:2401e2402. 21. Wilder JM, Oloruntoba OO, Muir AJ, Moylan CA. Role of patient factors, preferences, and distrust in health care and access to liver transplantation and organ donation. Liver Transpl 2016;22:895e905. 22. Goldberg DS, French B, Forde KA, et al. Association of distance from a transplant center with access to waitlist placement, receipt of liver transplantation, and survival among US veterans. JAMA 2014;311:1234e1243. 23. Laurentine KA, Bramstedt KA. Too poor for transplant: finance and insurance issues in transplant ethics. Prog Transplant 2010;20:178e185.
8
Tumin et al
Medicaid Expansion and Liver Transplantation
24. Flattau A, Olaywi M, Gaglio PJ, et al. Social barriers to listing for adult liver transplantation: their prevalence and association with program characteristics. Liver Transpl 2011;17:1167e1175. 25. Beal EW, Tumin D, Mumtaz K, et al. Factors contributing to employment patterns after liver transplantation. Clin Transplant 2017 [Epub ahead of print]. 26. Hart A, Smith JM, Skeans MA, et al. OPTN/SRTR 2015 Annual Data Report: Kidney. Am J Transplant 2017;17[Suppl 1]:21e116.
J Am Coll Surg
27. Pelletier RP, Henry ML. Program specific reports: friend or foe? dThe intended and unintended consequences of Scientific Registry of Transplant Recipient program specific reports. Curr Transpl Rep 2014;1:86e90. 28. Yoo HY, Thuluvath PJ. Outcome of liver transplantation in adult recipients: influence of neighborhood income, education, and insurance. Liver Transpl 2004;10: 235e243.
Vol. -,
No. -, -
Variable
2,350 (26) 52 (10) 3,257 (36)
6,064 (65) 779 (8) 2,530 (27)
5,897 (64) 787 (9) 2,478 (27)
4,365 2,410 2,180 2,176 635
(49) (27) (24) (24) (7)
4,294 2,265 2,194 2,171 698
(49) (26) (25) (24) (8)
3,058 344 844 725 2,251 927 1,224 2,383 720 18.1
(33) (4) (9) (8) (24) (10) (13) (26) (8) (9.8)
2,508 370 842 974 2,463 848 1,157 2,364 816 18.7
(27) (4) (9) (11) (27) (9) (13) (26) (9) (10.2)
<0.001 <0.001 0.094 0.775
Non-expansion states 2012e2013 (n ¼ 7,260) 2014e2015 (n ¼ 7,222)
969 (13) 53 (9) 2,604 (36)
900 (12) 53 (10) 2,601 (36)
5,376 (75) 775 (11) 1,109 (15)
5,200 (72) 779 (11) 1,243 (17)
3,118 1,839 1,627 1,670 1,280
(47) (27) (25) (24) (18)
3,168 1,790 1,690 1,554 1,266
(48) (27) (25) (22) (18)
2,086 361 694 744 1,572 806 997 1,811 515 18.6
(29) (5) (10) (10) (22) (11) (14) (25) (7) (9.0)
1,797 331 701 970 1,667 716 1,040 1,984 652 19.4
(25) (5) (10) (13) (23) (10) (14) (28) (9) (9.7)
0.244
*Unpaired t-test for continuous variables and chi-square test for categorical variables. MELD, Model for End-stage Liver Disease.
0.548 0.019 <0.001
0.547 0.002 <0.001
p Value*
0.112 0.181 0.854 0.005
0.378
0.017 0.691 <0.001
0.001 <0.001 <0.001
Medicaid Expansion and Liver Transplantation
1,956 (21) 53 (10) 3,222 (34)
p Value*
Tumin et al
Medicaid insurance, n (%) Patient age, y, mean (SD) Female patient, n (%) Patient race/ethnicity, n (%) Non-Hispanic white Non-Hispanic black Other Patient educational attainment, n (%) High school or less Some college Four-year college Patient working for income Rural residence Etiology of liver disease, n (%) Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes, n (%) History of dialysis, n (%) Initial MELD score, mean (SD)
Medicaid expansion states 2012e2013 (n ¼ 9,373) 2014e2015 (n¼ 9,162)
2017
eTable 1. Characteristics of Adults Aged 18 to 64 Years Listed for Liver Transplantation before (2012 to 2013) and after (2014 to 2015) Medicaid Expansion, According to State Participation in Medicaid Expansion (n ¼ 33,017)
8.e1
8.e2
Tumin et al
J Am Coll Surg
Medicaid Expansion and Liver Transplantation
eTable 2. Multivariable Logistic Regression of Medicaid Insurance Coverage at the Time of Listing for Liver Transplantation among Adult Candidates Aged 18 to 64 Years Listed in Their State of Permanent Residence (n ¼ 24,294) Variable*
State participation in Medicaid expansion State expanded Medicaid State did not expand Medicaid Period of listing for liver transplant 2012e2013 (pre-expansion) 2014e2015 (post-expansion) State and period interaction Medicaid non-expansion post-expansion period Patient age, y Female patient Patient race/ethnicity Non-Hispanic white Non-Hispanic black Other Patient educational attainment High school or less Some college Four-year college Patient working for income Rural residence Etiology of liver disease Viral Cryptogenic Autoimmune Nonalcoholic steatohepatitis Alcoholic liver disease Hepatocellular carcinoma Other History of diabetes History of dialysis Initial MELD score
Odds ratio
95% CIy
ref. 0.63
0.46, 0.86
0.004
ref. 1.49
1.33, 1.65
<0.001
0.60 0.96 1.32
0.49, 0.72 0.95, 0.97 1.23, 1.42
<0.001 <0.001 <0.001
ref. 1.73 1.80
1.54, 1.94 1.60, 2.02
<0.001 <0.001
ref. 0.59 0.39 0.20 1.15
0.54, 0.34, 0.17, 1.02,
0.65 0.44 0.23 1.31
<0.001 <0.001 <0.001 0.027
ref. 0.96 0.53 0.53 0.98 1.03 0.47 0.87 0.80 1.00
0.81, 0.45, 0.46, 0.88, 0.88, 0.39, 0.80, 0.67, 0.99,
1.14 0.63 0.61 1.08 1.21 0.57 0.95 0.96 1.00
0.640 <0.001 <0.001 0.663 0.732 <0.001 <0.001 0.015 0.299
*Model includes center-level random intercept. Area under receiver operating characteristic curve ¼ 0.75. y Model standard errors adjusted for clustering of patients within listing centers. MELD, Model for End-stage Liver Disease.
p Valuey