Abstract
Background
Health insurance coverage changes for many patients after liver transplantation, but the implications of this change on long-term outcomes are unclear.
Aims
To assess post-transplant patient and graft survival according to change in insurance coverage within 1 year of transplantation.
Methods
We queried the United Network for Organ Sharing for patients between ages 18–64 years undergoing liver transplantation in 2002–2016. Patients surviving > 1 year were categorized by insurance coverage at transplantation and the 1-year transplant anniversary. Multivariable Cox regression characterized the association between coverage pattern and long-term patient or graft survival.
Results
Among 34,487 patients in the analysis, insurance coverage patterns included continuous private coverage (58%), continuous public coverage (29%), private to public transition (8%) and public to private transition (4%). In multivariable analysis of patient survival, continuous public insurance (HR 1.29, CI 1.22, 1.37, p < 0.001), private to public transition (HR 1.17, CI 1.07, 1.28, p < 0.001), and public to private transition (HR 1.14, CI 1.00, 1.29, p = 0.044), were associated with greater mortality hazard, compared to continuous private coverage. After disaggregating public coverage by source, mortality hazard was highest for patients transitioning from private insurance to Medicaid (HR vs. continuous private coverage = 1.32; 95% CI 1.14, 1.52; p < 0.001). Similar differences by insurance category were found for death-censored graft failure.
Conclusion
Post-transplant transition to public insurance coverage is associated with higher risk of adverse outcomes when compared to retaining private coverage.
Keywords: Liver transplantation, Graft survival, Human, Medicaid, Survival, Insurance
Introduction
Liver transplantation (LT) is indicated for eligible patients with end-stage liver disease (ESLD), yet access to LT remains limited by the shortage of donor organs, barriers to timely diagnosis of ESLD and referral for evaluation, and significant financial burden of lifetime post-transplant care [1–6]. In the United States (US), means-tested Medicaid public insurance supports access to LT among low-income patients with ESLD, whereas Medicare public insurance supports LT access for both elderly and disabled patients [7]. Yet, public insurance coverage—particularly Medicaid coverage—has been persistently associated with increased risk of graft failure and lower patient survival after LT [8–10]. The association between public insurance coverage and worse LT outcomes has been partially attributed to higher disease severity at time of transplant (MELD) and longer wait time to transplant [9, 11–13] Patients with public insurance also tend to have lower educational attainment and less social support, both of which are associated with adverse health outcomes [8, 14]. Nevertheless, disparities in LT outcomes persist for Medicaid and Medicare patients even after adjusting for other risk factors such as age and creatinine, as well as other measures of socioeconomic status (income and education) and transplant-related factors such as center expertise [8].
In recent years, expansions of Medicaid eligibility related to the Patient Protection and Affordable Care Act (ACA) have fostered greater Medicaid participation among LT candidates, as well as LT recipients who had previously been covered by private insurance [15, 16]. Increasing public insurance enrollment after LT compounds a pattern previously noted in a single-center survey, where up to 52% of patients changed insurance after LT [17]. The frequent instability of insurance coverage after LT, and especially the incidence of private to public change in coverage, has led to the question of whether post-LT change in insurance is independently associated with long-term outcomes [18]. In other populations of organ transplant recipients, change from private to public insurance has been associated with increased mortality hazard [19], yet the applicability of these findings to the setting of LT is unclear. Therefore, we have used the US national transplant registry to test the hypothesis that early change from private to public insurance after LT is associated with lower patient and graft survival during long-term follow-up.
Methods
The study was deemed exempt from review by the local Institutional Review Board. Data were obtained from the United Network for Organ Sharing (UNOS), including follow-up through March 2017. First-time adult (age ≥ 18) recipients of orthotopic LT were included if they had undergone deceased donor LT alone between February 2002 and March 2016, to allow for at least 1 year of follow-up on insurance status, described below. Patients older than 64 at the time of LT were excluded, as universal Medicare coverage at ages 65 and older likely confounds any influences of change in insurance coverage past this age. To focus on long-term outcomes of LT, patients who died or experienced graft failure before the first transplant anniversary were excluded. Patients with missing data on study outcomes and patients without insurance follow-up data at the first transplant anniversary (described further below) were also excluded. Due to few LT being performed in patients with insurance other than private coverage, Medicaid, or Medicare, we excluded patients if they reported any coverage by other payors at the time of LT or on the first LT anniversary. Lastly, we excluded patients whose state of residence at the time of LT could not be determined, as we could not categorize these patients according to state participation in Medicaid expansion. The primary study outcomes were patient and graft survival conditional on survival to 1 year post-transplant.
Data on insurance coverage were mandatorily reported by transplant centers at the time of LT and at the first transplant anniversary [2]. We categorized patients according to their insurance coverage at the time of transplant and at the 1-year follow-up as: (1) continuous private, (2) continuous public, (3) private to public, and (4) public to private. To distinguish between Medicaid and Medicare, supplemental analyses differentiated among patients in categories (2) and (3) by their type of public insurance at the 1-year transplant anniversary (e.g., private-to-Medicaid change vs. private-to-Medicare change). Univariate Cox and Kaplan–Meier analyses were performed to characterize the association between insurance coverage change or continuity and long-term patient and graft survival. A sub-analysis was performed for patients in states with expanded Medicaid coverage [20], whose first LT anniversary fell on or after January 2014, when ACA was fully implemented (i.e., LT performed January 2013-March 2016).
Descriptive analysis using ANOVA and Chi-square tests was performed to characterize differences in potential confounders among categories of health insurance coverage. Covariates in the analysis included recipient age, gender, race, educational attainment, work participation, ABO blood type, body mass index (BMI), laboratory MELD score at time of transplant, indication for LT, history of diabetes, dialysis, transjugular intrahepatic portosystemic shunt (TIPSS), previous abdominal surgery, ascites, and portal vein thrombosis. Additionally, we adjusted for donor risk index (DRI), dichotomized with DRI > 1.9 representing high-risk donors; [21, 22] total center volume of LT over the study period; postoperative hospital length of stay (LOS); year of LT; residence in a state that expanded Medicaid eligibility; and UNOS region where patients had undergone LT. In a supplemental analysis, we examined whether the association between insurance coverage pattern and study outcomes varied according to UNOS region.
Multivariable Cox proportional hazards models were fitted for the study outcomes to examine confounding of the association between insurance coverage and LT outcomes. Multiple imputation using chained equations was used to complete missing data for multivariable analysis. Categorical variables were imputed using multinomial logistic regression models, and continuous variables were imputed using linear regression models. Independent variables in imputation models included study outcomes of patient and graft survival (including time to death or graft failure), in addition to all study covariates. Ten imputed data sets were created, and estimates were combined across imputations as previously described [23, 24].
Results
There were 59,026 adults ages 18–64 receiving first-time isolated LT during the study period, of whom 7635 were excluded due to death or graft failure within 1 year of transplant, 12,163 were excluded due to lack of follow-up data on insurance coverage at 1 year post-transplant, 3009 were excluded due to unknown or ineligible location of residence at the time of transplant, and 1732 were excluded due to an uncommon type of insurance coverage at transplant or the 1-year anniversary. The final sample included 34,487 patients, of whom 4188 comprised the Medicaid expansion subgroup (underwent LT in a state that expanded Medicaid eligibility, with 1-year transplant anniversary in 2014 or later). In the overall cohort, the most common type of insurance coverage was continuous private coverage (58%), followed by continuous public coverage (29%), private to public change in coverage (8%), and public to private change in coverage (4%).
Patient, donor, procedure, and center characteristics were compared across categories of insurance coverage in Table 1. (Descriptive statistics according to a more detailed classification of public insurance types are presented in Supplemental Table 1). Compared to patients with continuous private insurance, those who switched from private to public insurance were distinguished by higher unemployment, lower educational attainment, greater prevalence of diabetes, longer post-transplant LOS, and greater likelihood of residence in a Medicaid expansion state. Whereas Table 1 shows the proportion of patients from each region in a given category of insurance coverage (column percentage), we also calculated the proportion of each coverage type by UNOS region (row percentage) to determine which regions had the highest rates of private to public insurance change. This change in coverage was most common in Region 1 (104/1208; 9%), Region 2 (396/3488; 11%), Region 3 (618/5754; 11%), and Region 9 (288/2337; 12%), representing New England, the Mid-Atlantic, and the Southeastern US.
Table 1.
Liver transplant recipient and donor characteristics, by insurance trajectory (public insurance types combined), for patients surviving at least 1 year post-transplant (n, %, or mean, standard deviation)
| Covariate | Missing data | Continuous private insurance (N = 20,075) |
Continuous public insurance (N = 10,128) |
Transition from private to public insurance (N = 2909) |
Transition from public to private insurance (N = 1375) |
P value* |
|---|---|---|---|---|---|---|
| Agea (y) | 0 | 52 (9) | 52 (9) | 54 (10) | 51 (10) | < 0.001 |
| Male gender | 0 | 13,989 (70%) | 6483 (64%) | 1971 (68%) | 955 (69%) | < 0.001 |
| Race | 0 | – | – | – | – | < 0.001 |
| White | 15,506 (77%) | 6466 (64%) | 2135 (73%) | 933 (68%) | ||
| Black | 1641 (8%) | 1143 (11%) | 310 (11%) | 153 (11%) | ||
| Other | 2928 (15%) | 2519 (25%) | 464 (16%) | 289 (21%) | ||
| Education | 4571 | – | – | – | – | < 0.001 |
| High school or less | 7508 (43%) | 6010 (68%) | 1350 (55%) | 756 (63%) | ||
| Some college | 4741 (27%) | 1853 (21%) | 632 (26%) | 275 (23%) | ||
| College degree | 5188 (30%) | 967 (11%) | 467 (19%) | 169 (14%) | ||
| Employed | 831 | 8255 (42%) | 591 (6%) | 599 (21%) | 142 (11%) | < 0.001 |
| BMIa (kg/m2) | 9 | 28 (6) | 29 (6) | 29 (6) | 29 (6) | < 0.001 |
| ABO blood type | 0 | – | – | – | – | 0.006 |
| A | 7573 (38%) | 3641 (36%) | 1057 (36%) | 508 (37%) | ||
| B | 2668 (13%) | 1361 (13%) | 393 (14%) | 179 (13%) | ||
| AB | 1035 (5%) | 469 (5%) | 134 (5%) | 86 (6%) | ||
| O | 8799 (44%) | 4657 (46%) | 1325 (46%) | 602 (44%) | ||
| MELD score at transplanta | 37 | 21 (10) | 22 (10) | 22 (10) | 23 (10) | < 0.001 |
| Etiology of liver failure | 0 | – | – | – | – | < 0.001 |
| Viral | 4683 (23%) | 2977 (29%) | 770 (26%) | 416 (30%) | ||
| Cryptogenic | 978 (5%) | 424 (4%) | 135 (5%) | 65 (5%) | ||
| Autoimmune | 2265 (11%) | 742 (7%) | 234 (8%) | 96 (7%) | ||
| NASH | 1331 (7%) | 510 (5%) | 206 (7%) | 71 (5%) | ||
| Alcoholic | 3381 (17%) | 2118 (21%) | 527 (18%) | 272 (20%) | ||
| HCC | 4554 (23%) | 2298 (23%) | 661 (23%) | 293 (21%) | ||
| Other | 2883 (14%) | 1059 (10%) | 376 (13%) | 162 (12%) | ||
| Comorbid conditions | – | – | – | – | – | – |
| Diabetes | 405 | 4110 (21%) | 2410 (24%) | 719 (25%) | 320 (24%) | < 0.001 |
| Dialysis | 0 | 1590 (8%) | 889 (9%) | 241 (8%) | 128 (9%) | 0.034 |
| TIPSS | 201 | 1736 (9%) | 1333 (13%) | 331 (11%) | 135 (10%) | < 0.001 |
| Previous abdominal surgery | 154 | 9096 (45%) | 4865 (48%) | 1410 (49%) | 613 (45%) | < 0.001 |
| Ascites | 287 | 14,681 (74%) | 8089 (80%) | 2230 (77%) | 1116 (82%) | < 0.001 |
| Portal vein thrombosis | 637 | 1509 (8%) | 963 (10%) | 268 (9%) | 127 (9%) | < 0.001 |
| DRI > 1.9 | 1785 | 8944 (47%) | 4646 (49%) | 1371 (49%) | 610 (47%) | 0.013 |
| Center total LT volumea (×100 cases) | 0 | 11.7 (6) | 10.5 (6) | 11.8 (6) | 11.9 (6) | < 0.001 |
| Postop LOSa (days) | 37 | 13 (17) | 15 (18) | 16 (22) | 15 (17) | < 0.001 |
| Transplant yeara | 0 | 2009 (3) | 2009 (3) | 2009 (3) | 2009 (3) | 0.240 |
| Resident in a medicaid expansion state | 0 | 11,297 (56%) | 6010 (59%) | 1766 (61%) | 846 (62%) | < 0.001 |
| UNOS/OPTN region | 0 | < 0.001 | ||||
| 1 | 632 (3%) | 422 (4%) | 104 (4%) | 50 (4%) | ||
| 2 | 1997 (10%) | 890 (9%) | 396 (14%) | 205 (15%) | ||
| 3 | 3265 (16%) | 1670 (16%) | 618 (21%) | 201 (15%) | ||
| 4 | 2198 (11%) | 884 (9%) | 198 (7%) | 101 (7%) | ||
| 5 | 2672 (13%) | 1612 (16%) | 341 (12%) | 203 (15%) | ||
| 6 | 689 (3%) | 306 (3%) | 82 (3%) | 45 (3%) | ||
| 7 | 1606 (8%) | 850 (8%) | 204 (7%) | 76 (6%) | ||
| 8 | 1590 (8%) | 790 (8%) | 189 (7%) | 86 (6%) | ||
| 9 | 1255 (6%) | 706 (7%) | 288 (8%) | 88 (6%) | ||
| 10 | 2001 (10%) | 946 (9%) | 342 (12%) | 116 (8%) | ||
| 11 | 2170 (11%) | 1052 (10%) | 207 (7%) | 204 (15%) |
BMI body mass index, MELD model for end-stage liver disease, NASH nonalcoholic steatohepatitis, HCC hepatocellular carcinoma, TIPSS transjugular intrahepatic portosystemic shunt, DRI donor risk index, LT liver transplant LOS length of stay
P-value by ANOVA for continuous variables and Chi-square test for categorical variables
Mean and standard deviation
In univariate analysis of patient survival, outcomes significantly differed across patterns of insurance coverage (Fig. 1; log-rank test p < 0.001), with the highest mortality hazard seen among patients with continuous public insurance (HR relative to continuous private insurance = 1.40; 95% CI 1.33, 1.48; p < 0.001), followed by change from private to public insurance (HR 1.30; 95% CI 1.19, 1.41; p < 0.001) and change from public to private insurance (HR 1.23; 95% CI 1.09, 1.39; p = 0.001). We observed similar differences in graft survival across these insurance categories (Supplemental Fig. 1; log-rank p < 0.001), with the highest hazard of graft failure seen among patients with continuous public insurance (HR 1.32; 95% CI 1.25, 1.39; p < 0.001), followed by private to public change (HR 1.27; 95% CI 1.17, 1.38; p < 0.001) and public to private change (HR 1.25; 95% CI 1.11, 1.41; p < 0.001). In the Medicaid expansion subgroup, the increased hazard of mortality and graft failure among patients switching from private to public insurance did not reach statistical significance on univariate analysis (mortality HR 1.39; 95% CI 0.89, 2.15; p = 0.147; graft failure HR 1.48; 95% CI 0.96, 2.29; p = 0.077).
Fig. 1.
Kaplan–Meier curves of patient survival by insurance trajectory 1 year after Liver transplantation (N = 34,487)
Results from multivariable analysis of patient survival using multiply imputed data are summarized in Table 2. Continuous public insurance, transition from private to public coverage, and transition from public to private coverage all remained associated with greater mortality hazard when compared to continuous private insurance. When disaggregating coverage according to public insurance type (Supplemental Table 2), a higher mortality hazard was associated with transition from private to Medicaid insurance (HR 1.32; 95% CI 1.14, 1.52; p < 0.001), continuous Medicaid coverage (HR 1.31; 95% CI 1.21, 1.42; p < 0.001), and continuous Medicare coverage (HR 1.28; 95% CI 1.20, 1.37; p < 0.001). In region-specific analysis, the association between private to public change and greater mortality hazard only reached statistical significance in Region 3 (Supplemental Table 3, although in 5 of the 11 regions, the magnitude of this association exceeded the HR estimated for the national cohort.
Table 2.
Multivariable Cox proportional hazards model of long-term patient survival after liver transplantation, conditional on survival to 1 year (N = 34,487)
| Covariate | HR | 95% CI | P |
|---|---|---|---|
| Continuous private insurance | Ref. | – | – |
| Continuous public insurance | 1.29 | (1.22, 1.37) | < 0.001 |
| Transition from private to public insurance | 1.17 | (1.07, 1.28) | < 0.001 |
| Transition from public to private insurance | 1.14 | (1.00, 1.29) | 0.044 |
| Age (years) | 1.01 | (1.00, 1.01) | < 0.001 |
| Male | 1.12 | (1.06, 1.18) | < 0.001 |
| Race | – | – | – |
| White | Ref. | – | – |
| Black | 1.36 | (1.25, 1.47) | < 0.001 |
| Other | 0.76 | (0.71, 0.82) | < 0.001 |
| Education | – | – | – |
| High school or less | Ref | – | – |
| Some college | 0.97 | (0.90, 1.04) | 0.358 |
| College degree | 0.89 | (0.83, 0.96) | 0.003 |
| Employed | 0.88 | (0.83, 0.94) | < 0.001 |
| BMI (kg/m2) | 1.00 | (0.99, 1.00) | 0.072 |
| ABO blood type | – | – | – |
| A | Ref. | – | – |
| B | 0.91 | (0.84, 0.98) | 0.016 |
| AB | 0.97 | (0.87, 1.09) | 0.617 |
| O | 0.97 | (0.92, 1.02) | 0.264 |
| MELD at transplant | 0.99 | (0.990, 0.996) | < 0.001 |
| Etiology | – | – | – |
| Viral | Ref. | – | – |
| Cryptogenic | 0.73 | (0.64, 0.82) | < 0.001 |
| Autoimmune | 0.61 | (0.55, 0.69) | < 0.001 |
| NASH | 0.67 | (0.59, 0.78) | < 0.001 |
| Alcoholic | 0.95 | (0.89, 1.02) | 0.193 |
| HCC | 1.23 | (1.14, 1.31) | < 0.001 |
| Other | 0.82 | (0.75, 0.90) | < 0.001 |
| Comorbid conditions | – | – | – |
| Diabetes | 1.32 | (1.25, 1.40) | < 0.001 |
| Dialysis | 1.28 | (1.15, 1.42) | < 0.001 |
| TIPSS | 1.05 | (0.97, 1.13) | 0.188 |
| Previous abdominal surgery | 1.11 | (1.05, 1.17) | < 0.001 |
| Ascites | 1.01 | (0.94, 1.07) | 0.848 |
| Portal vein thrombosis | 0.94 | (0.85, 1.04) | 0.209 |
| DRI > 1.9 | 1.24 | (1.17, 1.30) | < 0.001 |
| Center total LT volume | 1.00 | (0.99, 1.00) | 0.052 |
| Postop LOS | 1.01 | (1.00, 1.01) | < 0.001 |
| Transplant year | 0.99 | (0.98, 1.00) | 0.041 |
| Resident in medicaid expansion state UNOS/OPTN region | 0.98 | (0.91, 1.05) | 0.494 |
| 1 | 0.99 | (0.84, 1.15) | 0.849 |
| 2 | 1.18 | (1.06, 1.31) | 0.003 |
| 3 | Ref. | – | – |
| 4 | 1.07 | (0.97, 1.18) | 0.199 |
| 5 | 0.97 | (0.87, 1.08) | 0.591 |
| 6 | 0.73 | (0.62, 0.87) | < 0.001 |
| 7 | 0.95 | (0.85, 1.06) | 0.353 |
| 8 | 0.91 | (0.82, 1.02) | 0.102 |
| 9 | 1.10 | (0.97, 1.24) | 0.126 |
| 10 | 1.02 | (0.92, 1.13) | 0.734 |
| 11 | 0.99 | (0.89, 1.09) | 0.798 |
HR hazard ratio, CI confidence interval, BMI body mass index, MELD model for end-stage liver disease, NASH nonalcoholic steatohepatitis, HCC hepatocellular carcinoma, TIPSS transjugular intrahepatic portosystemic shunt, DRI donor risk index, LT liver transplant, LOS length of stay
After adjustment for coverage type, college completion and pre-transplant employment remained associated with improved post-transplant survival, while Black race was associated with greater mortality hazard (Table 2). Similarly to the findings for patient survival, the hazard of graft failure remained elevated among patients with continuous public insurance or change between private and public insurance after multivariable adjustment (Table 3). When disaggregating HRs for graft failure according to public insurance type (Supplemental Table 2), transition from private insurance to Medicaid was associated with the highest hazard of graft failure (HR 1.33; 95% CI 1.16, 1.53; p < 0.001), compared to HRs of 1.23 (95% CI 1.14, 1.32) and 1.23 (95% CI 1.15, 1.32) for continuous Medicaid and continuous Medicare coverage, respectively.
Table 3.
Multivariable Cox proportional hazards model of long-term graft survival after liver transplantation, conditional on survival to 1 year (N = 34,487)
| Covariate | HR | 95% CI | P |
|---|---|---|---|
| Continuous private insurance | Ref. | – | – |
| Continuous public insurance | 1.23 | (1.16, 1.30) | < 0.001 |
| Transition from private to public insurance | 1.17 | (1.07, 1.27) | < 0.001 |
| Transition from public to private insurance | 1.17 | (1.03, 1.31) | 0.012 |
| Age (years) | 1.00 | (1.00, 1.01) | 0.353 |
| Male | 1.11 | (1.05, 1.18) | < 0.001 |
| Race | – | – | |
| White | Ref. | – | |
| Black | 1.38 | (1.28, 1.49) | < 0.001 |
| Other | 0.79 | (0.73, 0.84) | < 0.001 |
| Education | – | – | |
| High school or less | Ref | – | |
| Some college | 0.98 | (0.92, 1.05) | 0.606 |
| College degree | 0.91 | (0.84, 0.98) | 0.010 |
| Employed | 0.91 | (0.85, 0.97) | 0.002 |
| BMI (kg/m2) | 1.00 | (0.99, 1.00) | 0.062 |
| ABO blood type | – | – | |
| A | Ref. | – | |
| B | 0.92 | (0.86, 1.00) | 0.047 |
| AB | 0.99 | (0.89, 1.11) | 0.876 |
| O | 0.98 | (0.93, 1.03) | 0.397 |
| MELD at transplant | 0.99 | (0.99, 1.00) | < 0.001 |
| Etiology | – | – | |
| Viral | Ref. | – | |
| Cryptogenic | 0.71 | (0.63, 0.81) | < 0.001 |
| Autoimmune | 0.70 | (0.63, 0.77) | < 0.001 |
| NASH | 0.68 | (0.60, 0.77) | < 0.001 |
| Alcoholic | 0.92 | (0.86, 0.97) | 0.020 |
| HCC | 1.19 | (1.11, 1.28) | < 0.001 |
| Other | 0.79 | (0.73, 0.86) | < 0.001 |
| Comorbid conditions | – | – | – |
| Diabetes | 1.28 | (1.21, 1.35) | < 0.001 |
| Dialysis | 1.20 | (1.08, 1.33) | 0.001 |
| TIPSS | 1.02 | (0.95, 1.10) | 0.578 |
| Previous abdominal surgery | 1.10 | (1.05, 1.16) | < 0.001 |
| Ascites | 1.00 | (0.94, 1.06) | 0.980 |
| Portal vein thrombosis | 0.98 | (0.89, 1.07) | 0.634 |
| DRI > 1.9 | 1.27 | (1.20, 1.33) | < 0.001 |
| Center total LT volume | 0.99 | (0.99, 1.00) | 0.016 |
| Postop LOS | 1.01 | (1.00, 1.01) | < 0.001 |
| Transplant year | .097 | (0.96, 0.98) | < 0.001 |
| Resident in medicaid expansion state | 0.96 | (0.90, 1.03) | 0.245 |
| UNOS/OPTN region | – | – | – |
| 1 | 0.99 | (0.85, 1.15) | 0.862 |
| 2 | 1.15 | (1.04, 1.28) | 0.009 |
| 3 | Ref | – | – |
| 4 | 1.05 | (0.95, 1.16) | 0.377 |
| 5 | 0.96 | (0.87, 1.07) | 0.484 |
| 6 | 0.70 | (0.59, 0.83) | < 0.001 |
| 7 | 0.99 | (0.89, 1.10) | 0.839 |
| 8 | 0.96 | (0.86, 1.07) | 0.455 |
| 9 | 1.14 | (1.02, 1.28) | 0.025 |
| 10 | 1.03 | (0.93, 1.14) | 0.565 |
| 11 | 1.02 | (0.92, 1.12) | 0.707 |
HR hazard ratio, CI confidence interval, BMI body mass index, MELD model for end-stage liver disease, NASH nonalcoholic steatohepatitis, HCC hepatocellular carcinoma, TIPSS transjugular intrahepatic portosystemic shunt, DRI donor risk index, LT liver transplant, LOS length of stay
Discussion
Patterns of discontinuous insurance coverage are emerging as important predictors of surgical outcomes, particularly in the setting of solid organ transplantation [19], where survival depends on maintaining a costly regimen of immunosuppression and specialized follow-up care. In the context of LT, recent reports highlighting increased public insurance participation tied to the ACA Medicaid expansion have raised the question whether changes from private to public coverage among LT recipients would influence clinical outcomes in this population [15, 16]. Our study demonstrates that in a national cohort of LT recipients, transition from private to public insurance after LT is associated with worse long-term outcomes after accounting for recipient, donor, and transplant center characteristics. Particularly, transition from private insurance to Medicaid was associated with elevated hazards of mortality and graft failure. Although existing data remain insufficient to confirm this pattern specifically among LT recipients affected by ACA Medicaid expansion, our results suggest that protecting continuity of coverage for LT recipients with private insurance may be important for improving outcomes in this group.
Many explanations have been advanced as to why patients with public insurance have worse survival after solid organ transplantation. Public insurance (Medicare and Medicaid) has been associated with limited access to specialty care, including access to gastroenterology and liver transplantation [3, 12, 25, 26]. Additionally, Medicaid beneficiaries face restrictions in access to more expensive medications and therapies, such as antivirals for hepatitis C [27, 28]. This is concerning as hepatitis C is a leading cause of graft loss post-liver transplant [29]. Medicaid insurance may also be a surrogate for lower socioeconomic status, and a transition to Medicaid may be associated with financial hardship due to losing a job that previously provided employer-sponsored coverage. Below age 65, transition to Medicare may also be associated with onset of disability. Above age 65, however, transition to Medicare is typically related to age-based eligibility. Therefore, our results are unlikely to be generalizable to changes in insurance coverage among patients age 65 years or older.
As many LT recipients experience loss of employment after transplant [30], so, too, many report switching insurance coverage after transplantation. In a survey by Rongey et al. 34% of respondents reported having switched health insurance post-transplant, with 19% of the respondents losing or being denied insurance at some point [17]. With expanded Medicaid eligibility in participating states after ACA implementation, our group has previously noted increasing likelihood of LT recipients to switch from private to public insurance [16]. The clinical implications of this change in coverage are controversial [18], but our present study provides some evidence that keeping private coverage is preferable to switching to Medicaid insurance with respect to long-term LT outcomes. Indeed, our analysis may have underestimated the impact of private coverage loss, as it did not capture any early (< 1 year) mortality or graft failure attributable to such changes in coverage. Yet, our sub-analysis of patients specifically affected by ACA Medicaid expansion was underpowered to confirm these findings in the particular context of private-to-Medicaid change in coverage after this policy was implemented.
A growing literature has described how implementation of the ACA has increased access to surgical care by previously uninsured patients [31–35] However, the consequences of the ACA Medicaid expansion for surgical outcomes remain unclear. Charles et al. reported a decreased risk of morbidity and mortality after cardiac surgery in Michigan Medicaid patients after the ACA Medicaid expansion [32]. However, in a 2015 survey, members of the American College of Surgeons (ACS) Board of Governors observed that after ACA implementation, a higher percentage of patients were delaying elective surgery and presenting at a later stage of disease [33]. This was believed to be due to increased premiums among some of the participants. In the context of LT, Medicaid expansion increased the number of LT candidates using Medicaid as the primary payment type for transplant [15], as well as the number of LT recipients switching to Medicaid coverage post-transplant [16]. Our results suggest that such changes are associated with worse outcomes, relative to keeping private coverage. The reasons why patients change insurance after transplant are complex, and driven by changes in labor markets, as well as patients’ health, level of education, income, location and overall socioeconomic status. Yet, given our findings, interventions that mitigate the risk of losing private coverage after LT may be warranted.
Racial and socioeconomic disparities in surgical outcomes are well documented [36–40] but the mechanisms underlying these disparities are poorly understood. Various facets of socioeconomic status have been examined in surgical outcomes research, including insurance coverage, race, educational attainment, and ZIP code or county characteristics. LT recipients represent a unique cohort of surgical patients, insofar as patients presenting for LT are required to have proof of insurance prior to listing and transplantation, and transplant patients tend to be located near a transplant center. The uninsured rate among transplant patients is about 1–2% [41], much lower than the general population [42]. Because outright lack of insurance is relatively rare among LT recipients, continuity and change in insurance coverage may be especially informative in this population. In our analysis of mortality, the HR associated with change from private to Medicaid insurance (1.32) is almost as large as the (adjusted) Black-White disparity (HR 1.35), and exceeds adjusted HRs of high DRI (1.23) and dialysis (1.28). Notwithstanding prior evidence that public coverage at transplantation is associated with worse outcomes, this result indicates that adjusting surgical quality metrics for insurance coverage at a single point in time is inherently problematic, due to the lack of information on subsequent continuity or change of coverage.
Our conclusions are constrained by some limitations of the data and analysis. Most of the insurance transitions examined in this study occurred prior to the ACA Medicaid expansion. Due to limited data in the post-Medicaid expansion period, there were too few cases of death or graft failure in this period to confirm or refute the patterns found on multivariable analysis of the overall study cohort. Similarly, outside of Region 3, which accounted for the most patients in our cohort, our UNOS region sub-analysis may have been underpowered to confirm region-specific associations between private to public change in coverage and higher mortality hazard. Nevertheless, our results suggest that the impact of private insurance loss may vary across geographical areas, and could be highest where such change in coverage is especially common. As previously noted, UNOS data on LT only includes patients who have been listed for transplant, meaning that some disparities in access to care, including disparities addressed by the ACA, could not be adequately evaluated in this data set. Additionally, our focus on Medicaid expansion participation potentially underestimates state variability in outcomes as it may relate to long-standing differenced in health care provision, public insurance eligibility criteria, or other relevant factors. Lastly, our analysis may have been biased by incorrectly entered data, exclusion of patients who were lost to follow-up within 1 year, or unobserved confounders that could have accounted for the associations between discontinuous insurance coverage and poor long-term outcomes.
In summary, transition from private to public insurance after LT was associated with worse long-term outcomes when compared to retaining private insurance coverage. Among patients who had a change from private to public insurance coverage, transition to Medicaid was associated with especially high hazards of mortality and graft failure. Transition to Medicaid participation after LT is common and related to high rates of post-transplant unemployment, as well as recent policies expanding Medicaid eligibility. As the debate on reforming the health care system continues, policies to help transplant recipients maintain appropriate insurance coverage should be considered. In the interim, it is important for clinicians to recognize post-transplant insurance changes as a significant risk factor for graft failure and mortality, and take steps to mitigate this risk, such as closer surveillance and early referral to social work for help with finding and maintaining appropriate insurance.
Supplementary Material
Abbreviations
- LT
Liver transplantation
- ESLD
End-stage liver disease
- MELD
Model for end-stage liver disease
- ACA
Patient Protection and Affordable Care Act
- US
United States of America
- UNOS
United Network for Organ Sharing
- ANOVA
Analysis of variance
- TIPSS
Transjugular intrahepatic portosystemic shunt
- DRI
Donor risk index
- LOS
Length of stay
- HR
Hazard ratios
- CI
Confidence interval
- ACS
American College of Surgeons
Footnotes
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's Disclaimer: Disclaimer The data reported in this manuscript were supplied by the United Network for Organ Sharing as the contractor for the Organ Procurement and Transplantation Network. 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 U.S. Government.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10620-018-5031-6) contains supplementary material, which is available to authorized users.
References
- 1.Martin P, DiMartini A, Feng S, Brown R, Fallon M. Evaluation for liver transplantation in adults: 2013 practice guideline by the American Association for the Study of Liver Diseases and the American Society of Transplantation. Hepatology. 2014;59:1144–1165. 10.1002/hep.26972. [DOI] [PubMed] [Google Scholar]
- 2.UNOS. Transplant trends | UNOS. https://www.unos.org/data/transplant-trends/, 2017. Accessed 21 Feb 2018.
- 3.Barritt AS, Telloni SA, Potter CW, Gerber DA, Hayashi PH. Local access to subspecialty care influences the chance of receiving a liver transplant. Liver Transplant. 2013;19:377–382. 10.1002/lt.23588. [DOI] [PubMed] [Google Scholar]
- 4.Adler JT, Dong N, Markmann JF, Schoenfeld D, Yeh H. Role of patient factors and practice patterns in determining access to liver waitlist. Am J Transplant. 2015;15:1836–1842. 10.1111/ajt.13301. [DOI] [PubMed] [Google Scholar]
- 5.Flattau A, Olaywi M, Gaglio PJ, et al. Social barriers to listing for adult liver transplantation: their prevalence and association with program characteristics. Liver Transplant. 2011;17:1167–1175. 10.1002/lt.22357. [DOI] [PubMed] [Google Scholar]
- 6.Parikh ND, Hutton D, Marrero W, Sanghani K, Xu Y, Lavieri M. Projections in donor organs available for liver transplantation in the United States: 2014–2025. Liver Transplant. 2015;21:855–863. 10.1002/lt.24136. [DOI] [PubMed] [Google Scholar]
- 7.National Coverage Determination (NCD) for Adult Liver Transplantation (260.1). https://www.cms.gov/medicare-coverage-database/details/ncd-details.aspx?NCAId=259&NcaName=Liver+Transplantation+for+Malignancies&ExpandComments=n&CommentPeriod=0&NCDId=70&ncdver=3&id=186&bc=gABAAAAAEEAAAA%3d%3d&, 2017. Accessed 24 July 2017.
- 8.Yoo HY, Thuluvath PJ. Outcome of liver transplantation in adult recipients: influence of neighborhood income, education, and insurance. Liver Transplant. 2004;10:235–243. 10.1002/lt.20069. [DOI] [PubMed] [Google Scholar]
- 9.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;223:611–620. e614 10.1016/j.jamcollsurg.2016.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Glueckert LN, Redden D, Thompson MA, et al. What liver transplant outcomes can be expected in the uninsured who become insured via the Affordable Care Act? Am J Transplant. 2013;13:1533–1540. 10.1111/ajt.12244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kemmer N, Zacharias V, Kaiser TE, Neff GW. Access to liver transplantation in the MELD era: role of ethnicity and insurance. Dig Dis Sci. 2008;54:1794 10.1007/s10620-008-0567-5. [DOI] [PubMed] [Google Scholar]
- 12.Bryce CL, Angus DC, Arnold RM, et al. Sociodemographic differences in early access to liver transplantation services. Am J Transplant Off J Am Soc Transplant Am Soc of Transpl Surg. 2009;9:2092–2101. 10.1111/j.1600-6143.2009.02737.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bryce CL, Chang CC, Angus DC, Arnold RM, Farrell M, Roberts MS. The effect of race, sex, and insurance status on time-to-listing decisions for liver transplantation. J Transplant. 2010;2010:467976 10.1155/2010/467976. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mejia A, Halff GA, Esterl R, et al. Outcome of liver transplantation in hispanics versus non-hispanics: is there a difference? Transplant Proc. 2002;34:1236–1238. 10.1016/S0041-1345(02)02799-9. [DOI] [PubMed] [Google Scholar]
- 15.Tumin D, Beal EW, Mumtaz K, et al. Medicaid participation among liver transplant candidates after the Affordable Care Act medicaid expansion. J Am Coll Surg. 2017;225:173–180. e172 10.1016/j.jamcollsurg.2017.05.004. [DOI] [PubMed] [Google Scholar]
- 16.Tumin D, Hayes D, Washburn WK, Tobias JD, Black SM. Medicaid enrollment after liver transplantation: effects of medicaid expansion. Liver Transplant. 2016;22:1075–1084. 10.1002/lt.24480. [DOI] [PubMed] [Google Scholar]
- 17.Rongey C, Bambha K, Vanness D, et al. Employment and health insurance in long-term liver transplant recipients. Am J Transplant. 2005;5:1901–1908. 10.1111/j.1600-6143.2005.00961.x. [DOI] [PubMed] [Google Scholar]
- 18.Schlansky B, Shachar C. Implications of expanded medicaid eligibility for patient outcomes after liver transplantation: caveat emptor. Liver Transplant. 2016;22:1062–1064. 10.1002/lt.24491. [DOI] [PubMed] [Google Scholar]
- 19.Tumin D, Foraker RE, Smith S, Tobias JD, Hayes D. Health insurance trajectories and long-term survival after heart transplantation. Circ Cardiovasc Qual Outcomes. 2016;9:576–584. 10.1161/circoutcomes.116.003067. [DOI] [PubMed] [Google Scholar]
- 20.@KaiserFamFound. Status of state action on the medicaid expansion decision. http://www.kff.org/health-reform/state-indicator/state-activity-around-expanding-medicaid-under-the-affordable-care-act/, 2017. Accessed 24 July 2017.
- 21.Feng S, Goodrich NP, Bragg-Gresham JL, et al. Characteristics associated with liver graft failure: the concept of a donor risk index. Am J Transplant. 2006;6:783–790. 10.1111/j.1600-6143.2006.01242.x. [DOI] [PubMed] [Google Scholar]
- 22.Beal EW, Black SM, Mumtaz K, et al. High center volume does not mitigate risk associated with using high donor risk organs in liver transplantation. Dig Dis Sci. 2017;62:2578–2585. [DOI] [PubMed] [Google Scholar]
- 23.Graham JW, Olchowski AE, Gilreath TD. How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prev Sci. 2007;8:206–213. 10.1007/s11121-007-0070-9. [DOI] [PubMed] [Google Scholar]
- 24.White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30:377–399. 10.1002/sim.4067. [DOI] [PubMed] [Google Scholar]
- 25.Rosenbaum S Medicaid payments and access to care. New Engl J Med. 2014;371:2345–2347. 10.1056/NEJMp1412488. [DOI] [PubMed] [Google Scholar]
- 26.MACPAC. Examining access to care in medicaid and CHIP. https://www.macpac.gov/wp-content/uploads/2015/01/Examining_Access_to_Care_in_Medicaid_and_CHIP.pdf, 2011. Accessed 11 Aug 2017.
- 27.Liao JM, Fischer MA. Restrictions of Hepatitis C treatment for substance-using medicaid patients: cost versus ethics. Am J Public Health. 2017;107:893–899. 10.2105/AJPH.2017.303748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Barua S, Greenwald R, Grebely J, Dore GJ, Swan T, Taylor LE. Restrictions for medicaid reimbursement of sofosbuvir for the treatment of hepatitis C virus infection in the United States. Ann Intern Med. 2015;163:215–223. 10.7326/M15-0406. [DOI] [PubMed] [Google Scholar]
- 29.Dumortier J, Boillot O, Scoazec JY. Natural history, treatment and prevention of hepatitis C recurrence after liver transplantation: past, present and future. World J Gastroenterol. 2014;20:11069–11079. 10.3748/wjg.v20.i32.11069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Beal EW, Tumin D, Mumtaz K, et al. Factors contributing to employment patterns after liver transplantation. Clin Transplant. 2017;31:e12967 10.1111/ctr.12967. [DOI] [PubMed] [Google Scholar]
- 31.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;5:136–147. 10.1016/j.jchf.2016.10.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Charles EJ, Johnston LE, Herbert MA, et al. Impact of medicaid expansion on cardiac surgery volume and outcomes. Ann Thorac Surg. 2017. 10.1016/j.athoracsur.2017.03.079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Puls MW. 2015 ACS governors survey: surgeons describe the ACA’s effects on access to care. The Bulletin. http://bulletin.facs.org/2016/05/2015-acs-governors-survey-surgeons-describe-theacas-effects-on-access-to-care/, 2016. Accessed 14 Aug 2017. [PubMed]
- 34.Giladi AM, Aliu O, Chung KC. The effect of medicaid expansion in New York State on use of subspecialty surgical procedures by medicaid beneficiaries and the uninsured. J Am Coll Surg. 2014;218:889–897. 10.1016/j.jamcollsurg.2013.12.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Al-Refaie WB, Zheng C, Jindal M, et al. Did Pre-Affordable Care Act medicaid expansion increase access to surgical cancer care? J Am Coll Surg. 2017;224:662–669. 10.1016/j.jamcollsurg.2016.12.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Haider AH, Scott VK, Rehman KA, et al. Racial disparities in surgical care and outcomes in the United States: a comprehensive review of patient, provider, and systemic factors. J Am Coll Surg. 2013;216:482–492. 10.1016/j.jamcollsurg.2012.11.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Singh TP, Givertz MM, Semigran M, DeNofrio D, Costantino F, Gauvreau K. Socioeconomic position, ethnicity, and outcomes in heart transplant recipients. Am J Cardiol. 2010;105:1024–1029. 10.1016/j.amjcard.2009.11.015. [DOI] [PubMed] [Google Scholar]
- 38.Martin M, Beekley A, Kjorstad R, Sebesta J. Socioeconomic disparities in eligibility and access to bariatric surgery: a national population-based analysis. Surg Obes Relat Dis. 2010;6:8–15. 10.1016/j.soard.2009.07.003. [DOI] [PubMed] [Google Scholar]
- 39.Allen JG, Weiss ES, Arnaoutakis GJ, et al. Insurance and education predict long-term survival after orthotopic heart transplantation in the United States. J Heart Lung Transplant. 2012;31:52–60. 10.1016/j.healun.2011.07.019. [DOI] [PubMed] [Google Scholar]
- 40.Bennett KM, Scarborough JE, Pappas TN, Kepler TB. Patient socioeconomic status is an independent predictor of operative mortality. Ann Surg. 2010;252:552–558. 10.1097/SLA.0b013e3181f2ac64. [DOI] [PubMed] [Google Scholar]
- 41.Herring AA, Woolhandler S, Himmelstein DU. Insurance status of U.S. Organ Donors and Transplant Recipients: the uninsured give, but rarely receive. Int J Health Serv. 2008;38:641–652. 10.2190/HS.38.4.d. [DOI] [PubMed] [Google Scholar]
- 42.@KaiserFamFound. Health insurance coverage of the total population. http://www.kff.org/other/state-indicator/total-population/, 2016. Accessed 15 Aug 2017.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.