ABSTRACT
Background
As a 6‐month waiting period is required to receive exception points to prioritize patients with hepatocellular carcinoma (HCC) for liver transplantation, prompt addition to the waitlist is critical in access to LT.
Methods
Using Epic Cosmos data on patients with HCC referred for LT 1/2018‐10/2024, we used modified Poisson regression to calculate rates of waitlisting. Center‐level and individual (socioeconomic, geographic, and insurance) factors were measured among those who progressed.
Results
Among 11,422 HCC patients referred for LT at 70 centers, with median age 63 [IQR: 58, 68], 71.5% initiated evaluation and, of those who began evaluation, 57.6% were waitlisted for LT. Of those referred, patients who were older (age 70+ vs. 51‐60; RR 0.77, 95% CI: 0.65–0.90, p < 0.001), on Medicaid (0.83, 95% CI: 0.71–0.97, p = 0.02), never‐married (0.82, 95% CI: 0.73–0.91, p < 0.001), or low SES (Q4: 0.87, 95% CI: 0.77–0.97, p = 0.002) had lower rates of waitlisting. Among waitlisted patients, median time from referral was 3.3 months [IQR: 2.0, 5.3]. Despite adjustment for patient level covariates, there was high center‐level variation in rate of waitlisting within 12 months; 13% of centers listed patients at a rate ≥ 20% below the national median.
Conclusion
Only a fraction of referred patients with HCC are waitlisted for LT. High variation in access to waitlisting based on non‐clinical factors suggests barriers to waitlisting that must be addressed. Centers should focus on interventions to reduce barriers to waitlisting in patients with HCC.
Keywords: hepatocellular carcinoma, liver transplantation, referral, evaluation, waitlist
Abbreviations
- aRR
adjusted relative rate
- HCC
hepatocellular carcinoma
- LT
liver transplantation
- MASLD
metabolic‐dysfunction associated steatotic liver disease
- OPTN
Organ Procurement and Transplantation Network
- RUCA
Rural‐Urban Commuting Area
- SVI
Social Vulnerability Index
1. Introduction
Hepatocellular carcinoma (HCC) incidence in the United States tripled between 1980 and 2015, peaking at 10 cases per 100 000 before declining modestly to 9.2 per 100 000 by 2019. Despite this recent decline, mortality continues to rise with 30 090 projected deaths in 2025 [1]. Five‐year survival for all HCC patients stands at just 23%–27% [2], but reaches 70%–88% for those successfully undergoing liver transplantation (LT)—underscoring the life‐saving potential of early referral and treatment access [3, 4, 5]. Importantly, use of locoregional therapy for downstaging of HCC has broadened the pool of patients who could undergo curative therapy by LT, with a survival benefit [6, 7].
Access to transplantation involves navigating the transplant system of referral, evaluation, and waitlisting—necessary steps to be eligible for an offer. Importantly, since 2015, the United Network for Organ Sharing (UNOS) has required a six‐month waiting period after waitlisting before HCC patients receive MELD exception points—an effort to balance access to LT for those with and without HCC [8]. Barriers to LT for patients on the waitlist have been well studied as the organ procurement and transplantation network (OPTN) collects national data for every patient listed for transplantation [9, 10, 11, 12]. Our knowledge is limited on what happens to patients prior to waitlisting. The interval between referral and waitlisting represents a critical juncture where potentially curable patients are lost to tumor progression, hepatic decompensation, psychosocial barriers, and systemic failures in care coordination. Previously, a VA study using patients from 6 centers found that evaluation within 30 days of referral reduced time from referral to listing by 29 days and time to transplant by 115 days, with a mortality benefit for those with early evaluation [13]. More recently, a study of two transplant centers in Georgia found that patients from areas with more than 20% of their residents below the poverty line were 38% less likely to undergo evaluation for transplant after referral [14]. In this same study, Medicaid patients, when compared with those with private insurance, were 47% less likely to undergo evaluation. A national study would help us to better understand how patients may be limited in access to LT, and offer opportunities to improve policy.
In our study of United States electronic health record data, we measured access to evaluation and waitlisting in patients with HCC referred for LT in Epic Cosmos [15]. We looked at demographic, socioeconomic, and geographic data to understand factors associated with reduced access at each step. Using center level data, we measured the variation among transplant centers in rates of evaluation of those referred and waitlisting among those evaluated.
2. Methods
2.1. Study Population
Data for this study were obtained from Epic Cosmos, a centralized, deidentified electronic health record (EHR) database created and maintained by Epic Systems Corporation (Verona, WI) [15]. As of October 2025, this database included more than 300 million patient records from 1762 hospitals and 40 700 clinics across the United States, representing patients from all 50 states, Washington D.C., and select foreign countries. Cosmos uses HIPAA‐defined limited data sets submitted by participating healthcare providers nationwide. Cosmos data have been previously demonstrated to be representative of the general population of U.S. transplant recipients [16]. The dataset includes 70 LT centers, which represents 47% of the total number of LT centers in the US.
2.2. Cumulative Incidence of Progression through the LT Process
We identified patients aged 18–75 who were both diagnosed with HCC and referred for LT between 1/2018 and 10/2024. Patients were excluded if their earliest recorded date of diagnosis of HCC was after their date of referral for LT, along with patients with a prior liver transplantation. Patients were administratively censored on 10/1/2025. We constructed cumulative incidence curves of time from referral to waitlisting, treating recorded non‐progression as a competing event (death, deferral, not followed, ineligibility).
2.3. Patient Level Factors Associated With Progression
Since almost all patients who were waitlisted reached this milestone within 12 months of referral, we analyzed waitlisting within 12 months of referral as a binary outcome using modified Poisson regression, adjusting for referral age, sex, reported race, Hispanic ethnicity, preferred language, marital status, insurance type, diagnoses, body mass index, MELD at referral/evaluation, congestive heart failure, cardiovascular disease, diabetes, pulmonary disease, renal disease, smoking status, substance use, alcohol use, era (2018‐2019/2020‐2022/2023‐2024), Rural‐Urban Commuting Area (RUCA) categories (Metropolitan/Micropolitan/Rural and Small town/Other), and socioeconomic social vulnerability index (SVI) categories (quartiles). Socioeconomic SVI is a deprivation metric developed by the Centers for Disease Control that provides insight into the impact of social determinants of health. Patients are given a vulnerability index between 0 and 1 by Epic Cosmos based on ZIP code data on a patient's current residence. Similarly, we separately analyzed the outcomes of evaluation within 6 months among referred patients and waitlisting within 12 months among patients who initiated evaluation. Time to progression was estimated by assessing only patients who progressed from referral to waitlisting and measuring the median time and interquartile range among those who were waitlisted. To better understand limitations at each stage, this analysis was repeated for referral to evaluation initiation, evaluation initiation to waitlisting, and waitlisting to LT.
2.4. Recorded Non‐Progression to Waitlisting
To better understand the reasons for non‐progression prior to waitlisting, we identified recorded reasons among centers where >80% of patients had specific documentation as some of the centers record many non‐progression events as “unspecified”. These recorded reasons are entered into a patient's chart when they are removed from consideration in the Epic module for transplantation.
2.5. Center‐Level Variation in Access to Waitlisting
To examine center‐level variation in access to waitlisting, we extended the models above using hierarchical mixed‐effects Poisson regression with center‐level random intercept. Centers with center‐specific rate ratios (exponentiated random intercept) <0.80 were categorized as low‐performing, whereas centers with adjusted rate ratios >1.2 were categorized as high‐performing. To model the relationship between referral volume and each outcome, transplant centers were categorized into small (≤80 referrals), medium (≤ 200 referrals), and large (> 200 referrals) referral‐volume groups based on tertiles of volume of referrals at each center.
2.6. Statistical Analysis
All analyses were performed using R Studio 4.2.1.
3. Results
3.1. Study Population
A total of 11 422 patients diagnosed with HCC were referred for LT at 70 centers. The median age of the cohort was 63 years (IQR: 58–68 years) and patients were predominately White (77%), male (74%), English‐speaking (85%), and residing in metropolitan areas (84%). Clinical indications for these referrals were diverse, led by Viral Hepatitis + Alcohol (20%), Alcohol and metabolic‐dysfunction associated steatotic liver disease (MASLD) (16%), and MASLD (14%) of the cohort (Table 1).
TABLE 1.
Patients with HCC referred for transplantation between 1/2018‐10/2024 in Epic Cosmos.
| Characteristic | N = 11,422 a |
|---|---|
| Age category | |
| 18–30 | 76 (0.7%) |
| 31–40 | 241 (2.1%) |
| 41–50 | 634 (5.6%) |
| 51–60 | 2753 (24%) |
| 61–70 | 6314 (55%) |
| 70+ | 1404 (12%) |
| Race | |
| White | 8178 (77%) |
| Black or African American | 992 (9.3%) |
| Asian | 593 (5.6%) |
| American Indian or Alaska Native | 336 (3.1%) |
| Other Race | 583 (5.5%) |
| Ethnicity | |
| Not Hispanic or Latino | 8447 (81%) |
| Hispanic or Latino | 1970 (19%) |
| Sex | |
| Male | 8417 (74%) |
| Female | 3005 (26%) |
| BMI | |
| 20–24 | 2044 (18%) |
| <20 | 316 (2.8%) |
| 25–29 | 3545 (31%) |
| 30–34 | 2538 (22%) |
| 35–39 | 1124 (9.8%) |
| 40+ | 1855 (16%) |
| Disease category | |
| MASLD | 1606 (14%) |
| Hepatitis B | 284 (2.5%) |
| Hepatitis C | 1222 (11%) |
| Alcohol | 1279 (11%) |
| Viral hepatitis + Alcohol | 2313 (20%) |
| Alcohol + MASLD | 1831 (16%) |
| Viral Hepatitis + MASLD | 900 (7.9%) |
| Other | 1987 (17%) |
| MELD at Referral | |
| 10–20 | 4938 (56%) |
| <10 | 1632 (18%) |
| 21–30 | 1265 (14%) |
| 31–40 | 447 (5.0%) |
| 40+ | 612 (6.9%) |
| AFP | |
| <20 | 4943 (70%) |
| 20–99 | 1027 (15%) |
| 100–499 | 616 (8.8%) |
| ≥500 | 433 (6.2%) |
| Congestive heart failure | 849 (7.8%) |
| Cardiovascular disease | 881 (8.1%) |
| Diabetes mellitus | 5052 (46%) |
| Lung disease | 2474 (23%) |
| Chronic kidney disease | 1576 (14%) |
| Smoking status | |
| Current | 1789 (16%) |
| Former | 4481 (39%) |
| Never/No | 3841 (34%) |
| Unknown | 1311 (11%) |
| Substance use | |
| Never | 3305 (29%) |
| No | 2435 (21%) |
| Unknown | 3074 (27%) |
| Not Currently | 1675 (15%) |
| Yes | 933 (8.2%) |
| Alcohol use | |
| Never | 1302 (11%) |
| No | 2218 (19%) |
| Unknown | 1245 (11%) |
| Not Currently | 5536 (48%) |
| Yes | 1121 (9.8%) |
| Marital status | |
| Married/domestic partner | 6514 (57%) |
| No marital history | 2555 (22%) |
| Previously married | 2118 (19%) |
| Unknown/other | 235 (2.1%) |
| Insurance at diagnosis | |
| Miscellaneous/other | 7367 (64%) |
| Medicare | 3010 (26%) |
| Medicaid | 1045 (9.1%) |
| Socioeconomic SVI quartile | |
| Q1 (0%–25%) | 2598 (23%) |
| Q2 (26–50%) | 2623 (23%) |
| Q3 (51–75%) | 2924 (26%) |
| Q4 (76–100%) | 3227 (28%) |
| Region | |
| Midwest | 2309 (20%) |
| Northeast | 2058 (18%) |
| South | 4928 (43%) |
| West | 2127 (19%) |
| Preferred language | |
| English | 9747 (85%) |
| Spanish | 755 (6.6%) |
| Other | 920 (8.1%) |
| Rural‐urban commuting area (RUCA) | |
| Metropolitan | 9574 (84%) |
| Micropolitan | 1081 (9.5%) |
| Small town/rural | 754 (6.6%) |
| COVID‐19 Era | |
| 2018–2019 | 2749 (24%) |
| 2020–2022 | 5012 (44%) |
| 2023–2024 | 3661 (32%) |
an (%); Median (Q1, Q3).
Abbreviation: MASLD metabolic‐dysfunction associated steatotic liver disease, RUCA rural‐urban commuting area, SVI social vulnerability index.
3.2. Access to Waitlisting of Those Referred
3.2.1. Demographic, Socioeconomic, and Geographic
Of the referred patients, 71.5% underwent evaluation for LT. Among patients who were evaluated, 57.6% were waitlisted for LT (Figure 1). The median time to waitlisting among referred patients who were eventually waitlisted was 3.3 months [IQR: 2.0, 5.3] (Figure 2A). Overall, patients aged 70 years and older were 23% less likely to be waitlisted compared to those aged 51–60 years, with a relative risk (RR) of 0.77 (95% CI: 0.65–0.90, p < 0.001, Table 2). Younger patients, aged 18–30, had higher rates of waitlisting (RR: 1.50, 95% CI: 0.88–2.39, p = 0.11), though not statistically significant. There was no difference between males and females in likelihood of waitlisting (RR = 0.95, 95% CI: 0.86–1.04, p = 0.26). Patients insured under Medicaid had statistically significantly lower waitlisting rates than those with miscellaneous/other insurance (RR: 0.83, 95% CI: 0.71–0.97, p = 0.02). Patients who were never married had much lower rates of waitlisting compared to those who were currently married, with a RR of 0.82 (95% CI: 0.73–0.91, p < 0.001). Patients in the lowest socioeconomic quartile (Q4; 76%–100%) exhibited reduced rates of waitlisting compared to the highest quartile (Q1; 0%–25%), with an RR of 0.87 (95% CI: 0.77–0.97, p = 0.015). Similarly, patients in Q3 (51%–75%) had a reduced RR of 0.85 (95% CI: 0.76–0.96, p = 0.005). Further, patients with Alcohol‐associated Cirrhosis (RR: 0.74, 95% CI: 0.62–0.89, p < 0.001), Viral Hepatitis with alcohol (RR: 0.79, 95% CI: 0.68–0.93, p = 0.004), and Hepatitis C (RR: 0.80, 95% CI: 0.66–0.96, p = 0.02) had lower relative rates of waitlisting, when compared to patients with MASLD. Current Smokers (RR: 0.67, 95% CI: 0.57–0.78, p < 0.001), those endorsing alcohol use (RR: 0.72, 95% CI: 0.60–0.85, p < 0.001), and comorbid congestive heart failure (RR: 0.78, 95% CI: 0.66–0.92, p = 0.003) or cardiovascular disease (RR: 0.80, 95% CI: 0.68–0.94, p = 0.006) had lower rates of waitlisting. Patients diagnosed with cirrhosis were far more likely to be waitlisted, compared to those without cirrhosis (RR: 1.62, 95% CI: 1.10–2.51, p = 0.021). Patients in 2020‐2022 (RR: 0.87, 95% CI: 0.79–0.96, p = 0.007), as well as those in 2023‐2024 (RR: 0.80, 95% CI: 0.72–0.90, p < 0.001) had lower rates of waitlisting, compared with those patients listed in 2018–2019 (Table 2).
FIGURE 1.
Rate of progression to evaluation and waitlisting of those diagnosed with HCC and referred for liver transplantation.
FIGURE 2.
Accounting for the competing risk of attrition, cumulative incidence of progression to (A) waitlisting from referral, (B) evaluation from referral, and (C) waitlisting from evaluation initiation.
TABLE 2.
Rate ratios among patients diagnosed with HCC in access to: Waitlisting, from referral; evaluation initiation, from referral; waitlisting, from evaluation initiation.
| Referral to waitlisting | Referral to evaluation | Evaluation to waitlisting | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Characteristic | RR | 95% CI | p‐value | RR | 95% CI | p‐value | RR | 95% CI | p‐value |
| Age (ref: 51–60) | |||||||||
| 18–30 | 1.50 | (0.88–2.39) | 0.111 | 1.18 | (0.75–1.76) | 0.481 | 1.07 | (0.58–1.80) | 0.823 |
| 31–40 | 1.04 | (0.77–1.40) | 0.789 | 1.04 | (0.80–1.32) | 0.748 | 1.04 | (0.76–1.39) | 0.787 |
| 41–50 | 1.03 | (0.85–1.24) | 0.746 | 1.01 | (0.87–1.17) | 0.851 | 1.05 | (0.85–1.29) | 0.640 |
| 51–60 | Ref | Ref | Ref | ||||||
| 61–70 | 0.97 | (0.87–1.07) | 0.504 | 1.00 | (0.93–1.08) | 0.937 | 0.96 | (0.86–1.07) | 0.428 |
| 70+ | 0.77 | (0.65–0.90) | < 0.001 | 0.88 | (0.78–1.00) | 0.045 | 0.84 | (0.70–0.99) | 0.041 |
| Sex (ref: Male) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Male | Ref | Ref | Ref | ||||||
| Female | 0.95 | (0.86–1.04) | 0.259 | 0.99 | (0.92–1.07) | 0.832 | 0.94 | (0.85–1.04) | 0.252 |
| Insurance at Diagnosis (ref: Private/Other) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Private/Other | Ref | Ref | Ref | ||||||
| Medicare | 1.05 | (0.96–1.16) | 0.258 | 1.00 | (0.93–1.08) | 0.975 | 1.05 | (0.95–1.17) | 0.364 |
| Medicaid | 0.83 | (0.71–0.97) | 0.020 | 0.92 | (0.70–1.20) | 0.527 | 0.87 | (0.73–1.03) | 0.123 |
| Race (ref: White) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| White | Ref | Ref | Ref | ||||||
| Black or African American | 0.89 | (0.76–1.05) | 0.182 | 1.02 | (0.90–1.14) | 0.773 | 0.90 | (0.75–1.07) | 0.232 |
| Asian | 1.02 | (0.83–1.23) | 0.822 | 1.03 | (0.87–1.22) | 0.724 | 1.04 | (0.83–1.27) | 0.727 |
| American Indian or Alaska Native | 0.94 | (0.74–1.18) | 0.608 | 0.95 | (0.79–1.14) | 0.616 | 1.00 | (0.77–1.29) | 0.970 |
| Other Race | 0.95 | (0.80–1.15) | 0.604 | 0.99 | (0.86–1.16) | 0.934 | 1.02 | (0.84–1.25) | 0.813 |
| Ethnicity (ref: non‐Hispanic) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Non‐Hispanic | Ref | Ref | Ref | ||||||
| Hispanic or Latino | 1.08 | (0.94–1.24) | 0.257 | 1.03 | (0.93–1.15) | 0.553 | 1.02 | (0.87–1.18) | 0.833 |
| BMI (ref: 20–24) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| < 20 | 0.78 | (0.54–1.08) | 0.238 | 0.92 | (0.72–1.24) | 0.555 | 0.78 | (0.53–1.11) | 0.180 |
| 20–24 | Ref | Ref | Ref | ||||||
| 25–29 | 1.07 | (0.95–1.21) | 0.289 | 1.05 | (0.96–1.16) | 0.283 | 1.02 | (0.89–1.16) | 0.796 |
| 30–34 | 1.09 | (0.95–1.24) | 0.213 | 1.06 | (0.96–1.17) | 0.267 | 1.02 | (0.88–1.18) | 0.794 |
| 35–39 | 1.03 | (0.88–1.22) | 0.684 | 1.07 | (0.95–1.21) | 0.275 | 0.97 | (0.81–1.15) | 0.702 |
| 40+ | 1.04 | (0.89–1.22) | 0.627 | 1.05 | (0.93–1.28) | 0.290 | 1.00 | (0.84–1.17) | 0.960 |
| MELD at Referral1 / Evaluation2 (ref: 10–20) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| < 10 | 1.04 | (0.94–1.16) | 0.400 | 1.00 | (0.92–1.09) | 0.960 | 1.05 | (0.93–1.17) | 0.435 |
| 10–20 | Ref | Ref | Ref | ||||||
| 21–30 | 0.88 | (0.78–1.00) | 0.048 | 0.96 | (0.87–1.05) | 0.340 | 0.91 | (0.79–1.03) | 0.150 |
| 31–40 | 0.76 | (0.61–0.94) | 0.013 | 0.88 | (0.75–1.02) | 0.090 | 0.82 | (0.64–1.02) | 0.083 |
| 40+ | 0.93 | (0.78–1.00) | 0.403 | 0.99 | (0.87–1.13) | 0.887 | 0.93 | (0.77–1.10) | 0.422 |
| AFP at Referral1 / Evaluation2 (ref: <20) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| <20 | Ref | Ref | Ref | ||||||
| 20–99 | 0.94 | (0.84–1.06) | 0.347 | 0.99 | (0.90–1.08) | 0.295 | 0.92 | (0.80–1.04) | 0.186 |
| 100–499 | 0.78 | (0.67–0.91) | 0.002 | 0.93 | (0.83–1.05) | 0.238 | 0.86 | (0.72–1.01) | 0.064 |
| ≥ 500 | 0.44 | (0.34–0.56) | < 0.001 | 0.73 | (0.62–0.85) | < 0.001 | 0.64 | (0.49–0.83) | 0.001 |
| Comorbidities (Yes vs. No) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Congestive heart failure | 0.78 | (0.66–0.92) | 0.003 | 0.96 | (0.85–1.08) | 0.459 | 0.86 | (0.69–1.02) | 0.100 |
| Cardiovascular disease | 0.80 | (0.68–0.94) | 0.006 | 0.96 | (0.85–1.07) | 0.461 | 0.85 | (0.71–1.01) | 0.068 |
| Diabetes | 0.94 | (0.87–1.03) | 0.187 | 1.00 | (0.94–1.07) | 0.948 | 0.94 | (0.84–1.03) | 0.108 |
| Pulmonary disease | 0.93 | (0.84–1.03) | 0.169 | 0.95 | (0.88–1.03) | 0.224 | 0.97 | (0.86–1.08) | 0.560 |
| Kidney disease | 1.30 | (1.17–1.44) | <0.001 | 1.09 | (1.00–1.19) | 0.043 | 1.18 | (1.05–1.32) | 0.004 |
| Smoking status (ref: Never/No) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Never/No | Ref | Ref | Ref | ||||||
| Current | 0.67 | (0.57–0.78) | <0.001 | 0.88 | (0.79–0.98) | 0.019 | 0.75 | (0.63–0.88) | 0.001 |
| Former | 1.02 | (0.93–1.12) | 0.680 | 0.90 | (0.82–0.98) | 0.017 | 1.01 | (0.92–1.12) | 0.821 |
| Unknown | 1.02 | (0.89–1.17) | 0.776 | 0.98 | (0.88–1.09) | 0.713 | 1.03 | (0.89–1.20) | 0.648 |
| Substance use (ref: Never/No) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Never/No | Ref | Ref | Ref | ||||||
| Unknown | 0.98 | (0.87–1.10) | 0.767 | 1.05 | (0.94–1.15) | 0.392 | 0.98 | (0.86–1.11) | 0.804 |
| Not currently | 1.00 | (0.88–1.13) | 0.951 | 1.00 | (0.90–1.10) | 0.992 | 1.01 | (0.88–1.16) | 0.858 |
| Yes | 0.90 | (0.75–1.06) | 0.207 | 0.99 | (0.87–1.13) | 0.876 | 0.91 | (0.75–1.09) | 0.335 |
| Alcohol use (ref: Never/No) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Never/No | Ref | Ref | Ref | ||||||
| Unknown | 0.90 | (0.70–1.17) | 0.440 | 0.98 | (0.85–1.13) | 0.771 | 1.03 | (0.85–1.25) | 0.723 |
| Not currently | 1.02 | (0.93–1.12) | 0.709 | 0.99 | (0.92–1.07) | 0.780 | 1.03 | (0.93–1.14) | 0.540 |
| Yes | 0.72 | (0.60–0.85) | <0.001 | 0.87 | (0.76–0.99) | 0.031 | 0.88 | (0.73–1.05) | 0.163 |
| Marital status (ref: Married) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Married | Ref | Ref | Ref | ||||||
| No marital history | 0.82 | (0.73–0.91) | 0.001 | 0.94 | (0.86–1.02) | 0.161 | 0.88 | (0.78–1.00) | 0.052 |
| Previously married | 0.89 | (0.79–0.99) | 0.034 | 0.99 | (0.91–1.07) | 0.758 | 0.95 | (0.84–1.07) | 0.413 |
| Unknown/other | 0.76 | (0.50–1.18) | 0.217 | 0.88 | (0.65–1.15) | 0.364 | 0.87 | (0.55–1.30) | 0.538 |
| RUCA (ref: Metropolitan) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Metropolitan | Ref | Ref | Ref | ||||||
| Micropolitan | 0.93 | (0.80–1.06) | 0.247 | 0.99 | (0.89–1.10) | 0.815 | 0.95 | (0.81–1.11) | 0.517 |
| Small town/rural | 0.88 | (0.74–1.04) | 0.138 | 0.89 | (0.78–1.02) | 0.095 | 0.94 | (0.77–1.12) | 0.488 |
| Preferred language (ref: English) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| English | Ref | Ref | Ref | ||||||
| Spanish | 1.10 | (0.89–1.28) | 0.380 | 0.99 | (0.86–1.14) | 0.934 | 1.10 | (0.90–1.34) | 0.340 |
| Other | 0.98 | (0.77–1.23) | 0.835 | 1.01 | (0.83–1.21) | 0.935 | 0.94 | (0.73–1.20) | 0.608 |
| SVI socioeconomic quartile (ref: Q1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Q1 (0%–25%) | Ref | Ref | Ref | ||||||
| Q2 (26%–50%) | 0.90 | (0.80–1.00) | 0.058 | 1.01 | (0.92–1.11) | 0.815 | 0.89 | (0.79–1.01) | 0.066 |
| Q3 (51%–75%) | 0.85 | (0.76–0.96) | 0.005 | 0.96 | (0.87–1.06) | 0.405 | 0.88 | (0.78–1.00) | 0.055 |
| Q4 (76%–100%) | 0.87 | (0.77–0.97) | 0.015 | 0.97 | (0.88–1.07) | 0.481 | 0.91 | (0.80–1.03) | 0.171 |
| Disease category (ref: MASLD) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| MASLD | Ref | Ref | Ref | ||||||
| Hepatitis B | 1.13 | (0.85–1.51) | 0.377 | 1.01 | (0.79–1.28) | 0.944 | 1.11 | (0.81–1.50) | 0.456 |
| Hepatitis C | 0.80 | (0.66–0.96) | 0.019 | 0.90 | (0.78–1.04) | 0.167 | 0.82 | (0.67–1.01) | 0.066 |
| Alcohol | 0.74 | (0.62–0.89) | <0.001 | 0.95 | (0.83–1.09) | 0.491 | 0.76 | (0.63–0.92) | 0.006 |
| Viral hepatitis + alcohol | 0.79 | (0.68–0.93) | 0.004 | 0.96 | (0.85–1.08) | 0.492 | 0.80 | (0.68–0.95) | 0.012 |
| Alcohol + MASLD | 1.00 | (0.87–1.14) | 0.998 | 1.07 | (0.96–1.19) | 0.209 | 0.94 | (0.81–1.08) | 0.373 |
| Viral hepatitis + MASLD | 1.19 | (1.02–1.41) | 0.028 | 1.05 | (0.93–1.21) | 0.393 | 1.10 | (0.93–1.31) | 0.240 |
| Other | 1.07 | (0.93–1.27) | 0.339 | 1.03 | (0.92–1.17) | 0.554 | 1.05 | (0.90–1.22) | 0.566 |
| Cirrhosis status (ref: No) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| No | Ref | Ref | Ref | ||||||
| Yes | 1.62 | (1.10–2.51) | 0.021 | 1.42 | (1.05–1.92) | 0.030 | 1.19 | (0.79–1.90) | 0.427 |
| COVID‐19 era (ref: 2018–2019) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 2018–2019 | Ref | Ref | Ref | ||||||
| 2020–2022 | 0.87 | (0.79–0.96) | 0.007 | 0.93 | (0.85–1.00) | 0.056 | 0.92 | (0.82–1.03) | 0.136 |
| 2023–2024 | 0.80 | (0.72–0.90) | < 0.001 | 0.91 | (0.83–0.99) | 0.028 | 0.89 | (0.79–1.01) | 0.056 |
Abbreviations: MASLD, metabolic‐dysfunction associated steatotic liver disease; RUCA, rural‐urban commuting area; SVI social vulnerability index.
3.2.2. Recorded Non‐Progression to Waitlisting
Of the 26 (37%) centers meeting the threshold for documentation of recorded non‐progression, the most common reason for removal was “Does Not Meet Criteria,” accounting for 23% of the cases. Other reasons for non‐progression included the determination that a patient was not a candidate (7.3%), death (7.9%), patient's decision (17%), financial/insurance complications (6.8%), medical contraindication (6.6%), the candidates condition deteriorated or was too sick to transplant (3.1%), substance abuse (4.4%), referred too early (3%), and unable to contact patient (2.5%). The remaining patients (8.8%) had recorded non‐progression for other reasons (Table 3).
TABLE 3.
Reasons for recorded non‐progression to waitlisting among those referred for liver transplantation for HCC.
| Reason for non‐progression | N = 2023 |
|---|---|
| Does not meet criteria | 471 (23%) |
| Patient choice | 334 (17%) |
| Unspecified | 200 (9.9%) |
| Patient died | 159 (7.9%) |
| Not a candidate | 147 (7.3%) |
| Financial/Insurance complications | 137 (6.8%) |
| Medical contraindication | 134 (6.6%) |
| Substance abuse | 90 (4.4%) |
| Candidate condition deteriorated, too sick for transplant | 63 (3.1%) |
| Referred too early | 60 (3.0%) |
| Unable to contact patient | 50 (2.5%) |
| Other | 178 (8.8%) |
3.2.3. Center‐Level Variation
Center‐level variability in waitlisting rates was substantial; 9 LT centers were low‐performing, while 11 centers were high‐performing, compared to the national median (Figure 3A). Waitlisting was not lower at low volume centers (RR: 0.90, 95% CI: 0.78–1.04, p = 0.150), or medium volume centers (RR: 1.00, 95% CI: 0.93–1.07, p = 0.99), when compared to high volume. Rates of waitlisting within 6 months for patients referred to centers with a high proportion of Medicaid patients was no different from patients at low proportion centers (≥ 20% vs. < 20%, RR: 0.97, 95% CI: 0.86–1.08, p = 0.55).
FIGURE 3.
Center level variation in access to waitlisting among those with HCC referred for liver transplantation.
3.3. Access to Evaluation of Those Referred
3.3.1. Demographic, Socioeconomic, and Geographic
Median time to evaluation initiation among patients who initiated evaluation was 0.9 months [IQR: 0.4, 1.5] (Figure 2B). Patients who were older 70+ were less likely to initiate evaluation (RR: 0.88, 95% CI: 0.78–1.00, p = 0.045), when compared to patients age 51–60. Current smokers, compared with non/never smokers were less likely to initiate evaluation (RR: 0.88, 95% CI: 0.79–0.98, p = 0.019) and those who reported alcohol use, compared with non/never users, had lower rates of evaluation initiation (RR: 0.87, 95% CI: 0.76–0.99, p = 0.031).
3.3.2. Center‐Level Variation
There was noted center‐level variability in rates of evaluation initiation; 5 LT centers were low‐performing, while 6 centers were high‐performing centers compared to the national median (Figure 3B). Evaluation initiation rates were not statistically significantly different for centers with medium and low volume or for center with high proportion of Medicaid patients.
3.4. Access to Waitlisting of Those Who Initiated Evaluation
3.4.1. Demographic, Socioeconomic, and Geographic
Median time to waitlisting among those who initiated evaluation and were waitlisted was 2.2 months [IQR: 1, 4] (Figure 2C). No differences were observed in waitlisting rates between male and female sex (RR: 0.94, 95% CI: 0.85–1.04, p = 0.252). Patients residing in micropolitan (RR: 0.95, 95% CI: 0.81–1.11, p = 0.517) and small‐town/rural areas (RR: 0.94, 95% CI: 0.77–1.12, p = 0.488) did not show statistically significant differences compared to metropolitan residents. Patients with alcoholic cirrhosis (RR: 0.76, 95% CI: 0.63–0.92, p = 0.006) and combined viral hepatitis and alcohol (RR: 0.80, 95% CI: 0.68–0.95, p = 0.012) had lower relative rates of waitlisting, when compared to patients with MASLD.
3.4.2. Center‐Level Variation
Center‐level variability, was much lower between centers for waitlisting among patients who initiated evaluation with no high‐ or low performing centers noted in the analysis (Figure 3C). Waitlisting rates were no different among centers based on volume or percentage of Medicaid patients.
4. Discussion
In this national study of access to the LT waitlist for over 11 000 patients in Epic Cosmos who were referred for LT at 70 centers across the US, we found substantial variation in rates of evaluation and waitlisting based on center‐level, socioeconomic, and patient‐level factors. Among patients referred for LT, only 71.5% proceeded to evaluation and 57.6% of those who initiated evaluation were waitlisted. Of those referred, relative rates of waitlisting were 23% lower in patients who were older [age 70+ vs. 51–60], 17% lower for patients on Medicaid, 18% lower for patients who were never married, and 13% lower for patients with low socioeconomic status. There was also high variability associated with center‐level factors. Some transplant centers performed substantially better or worse in waitlisting within twelve months of referral. 13% of centers listed patients at a rate ≥ 20% below the national median, and 16% listed patients at ≥ 20% above the national median. Importantly, time to waitlisting varied substantially, underscoring the urgency of addressing barriers in the early steps of the LT process to improve access for patients with HCC.
Our results align with findings from prior studies in highlighting systemic issues in equitable access to LT. As previously reported in a VA study of 6 centers (50% of patients with HCC), there was variability in time to evaluation for patients undergoing LT evaluation, including a survival benefit to early evaluation [13]. Our study demonstrates substantial variability in rate of evaluation and waitlisting among individual centers included in our study, despite controlling for the differences in case mix between centers. A study of two centers in Georgia looked at evaluation for patients with HCC and found a higher rate of evaluation initiation than our study, although their study only included early stage HCC [14]. For the first time, we have found that these barriers are found nationally, and we have further identified additional individual risk factors, center specific factors, and substantial variation in center level rate of progression from referral to waitlisting. Further, we identified reasons for non‐progression among patients referred, but not waitlisted. We noted that 11% of these patients died or clinically deteriorated while awaiting waitlisting, and only 30% did not meet criteria or were determined to not be a candidate. Importantly, financial and insurance complications were noted as a factor for non‐progression in 7% of patients who did not progress to waitlisting.
Importantly, our study focused on all patients referred for LT with HCC and not just those within Milan Criteria. As Lai et al. reported in 2019, the survival benefit for LT is not limited to those within Milan criteria, in fact those who are outside Milan criteria with a response to local regional therapy benefit most from LT in their model [17]. Limiting referral to only early stage patients misses an opportunity to transplant those who would benefit most. We add to these studies by performing a national study that assessed 70 LT centers (47% of US LT centers) and focused on attrition and variation in access to the waitlist.
Very likely, our findings are in part mediated by HCC stage at detection. Tumor size and number and stage of cancer are unavailable in the Cosmos database, and thus were not used in this study. Patients with lower socioeconomic status are more likely to be diagnosed with HCC at a later stage [18]. Since tumor and staging information are unavailable, we cannot guarantee that all patients were eligible for LT; nevertheless, the fact that all these patients were referred for LT by a medical provider suggests they were deemed a potential candidate. The substantial differences seen in this national study based on socioeconomic factors as well as social support are unlikely to be mediated entirely by cancer stage, and illustrate substantial pre‐waitlisting barriers to LT.
Importantly, our data rely on a deidentified EHR repository, which omits granular variables like individual psychosocial readiness, and precise center‐level organizational barriers. The Epic Cosmos database is limited to participating Epic centers, which may not represent all populations in the US. While referral volume and center performance were included, the categorization of high‐ and low‐performing centers may neglect unmeasured structural challenges within certain transplant regions. When looking at reasons for non‐progression, there was variability among completeness at centers, so we opted to use only the centers with good documentation. There is likely center level variability in why patients are removed, although this figure represents 37% of the patients in the study who were noted to have non‐progression. We also note that, while this study attempted to include all patients with HCC referred for liver transplantation, if a referring physician (regardless of whether that physician is within the Epic system) recommends that a patient be evaluated for transplantation, but neither the patient nor the physician contacts the transplant team, the referral will not be included in our study. All referrals to Epic Cosmos centers, including those from non‐Epic centers, are captured. Furthermore, confounding factors such as socioeconomic status and insurance type may remain incompletely adjusted despite robust statistical approaches; however, these biases likely amplify rather than diminish the significance of our findings regarding disparities in LT access. Fortunately, the extensive national deidentified dataset provides insights into systemic barriers, representing diverse populations and geographically diverse centers across the U.S.
In conclusion, our findings highlight substantial patient‐, center‐, and system‐level barriers that limit equitable access to timely waitlisting for patients referred for LT with HCC. Addressing these inequities, especially among older patients, those with lower socioeconomic status and Medicaid insurance, is critical for improving access to curative therapies and reducing HCC mortality. Policies aimed at standardizing early steps in the LT process, such as rapid evaluation protocols or community‐based outreach programs for disadvantaged populations, should be prioritized. Additionally, efforts to reduce center‐level variability, including resource allocation and practice standardization, may improve waitlisting rates across diverse settings. Further studies are needed to better understand the intersection of individual factors and center‐level practices, and to develop targeted interventions for equitable LT access.
Conflicts of Interest
The authors declare no conflicts of interest.
Acknowledgments
This work was supported by grant number T32CA193111 (Donnelly, Eitan) from the National Cancer Institute. The analyses described here are the responsibility of the authors alone and do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the U.S. Government.
Data Availability Statement
Data sharing not applicable to this article as data was derived from Epic Cosmos and only those from participating institutions have access to this data.
References
- 1. Singal A. G., Llovet J. M., Yarchoan M., et al., “AASLD Practice Guidance on Prevention, Diagnosis, and Treatment of Hepatocellular Carcinoma,” Hepatology 78, no. 6 (2023): 1922–1965, 10.1097/HEP.0000000000000466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Yilma M., Houhong Xu R., and Saxena V., “Survival Outcomes Among Patients With Hepatocellular Carcinoma in a Large Integrated US Health System,” JAMA Network Open 7, no. 9 (2024): e2435066, 10.1001/jamanetworkopen.2024.35066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mazzaferro V., Regalia E., Doci R., et al., “Liver Transplantation for the Treatment of Small Hepatocellular Carcinomas in Patients with Cirrhosis,” New England Journal of Medicine 334, no. 11 (1996): 693–700, 10.1056/NEJM199603143341104. [DOI] [PubMed] [Google Scholar]
- 4. Chen J. W., Kow L., Verran D. J., et al., “Poorer Survival in Patients Whose Explanted Hepatocellular Carcinoma (HCC) Exceeds Milan or UCSF Criteria. An Analysis of Liver Transplantation in,” HCC in Australia and New Zealand HPB 11, no. 1 (2009): 81–89, 10.1111/j.1477-2574.2009.00022.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Shimamura T., Goto R., Watanabe M., Kawamura N., and Takada Y., “Liver Transplantation for Hepatocellular Carcinoma: How Should We Improve the Thresholds?,” Cancers 14, no. 2 (2022): 419, 10.3390/cancers14020419. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Tabrizian P., Holzner M. L., Mehta N., et al., “Ten‐Year Outcomes of Liver Transplant and Downstaging for Hepatocellular Carcinoma,” JAMA Surgery 157, no. 9 (2022): 779–788, 10.1001/jamasurg.2022.2800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mazzaferro V., Citterio D., Bhoori S., et al., “Liver Transplantation in Hepatocellular Carcinoma after Tumour Downstaging (XXL): A Randomised, Controlled, Phase 2b/3 Trial,” Lancet Oncology 21, no. 7 (2020): 947–956, 10.1016/S1470-2045(20)30224-2. [DOI] [PubMed] [Google Scholar]
- 8. Heimbach J. K., Hirose R., Stock P. G., et al., “Delayed Hepatocellular Carcinoma Model for End‐stage Liver Disease Exception Score Improves Disparity in Access to Liver Transplant in the United States,” Hepatology 61, no. 5 (2015): 1643–1650, 10.1002/hep.27704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Ishaque T., Massie A. B., Bowring M. G., et al., “Liver Transplantation and Waitlist Mortality for HCC and Non‐HCC Candidates Following the 2015 HCC Exception Policy Change,” American Journal of Transplantation: Official Journal of the American Society of Transplantation and the American Society of Transplant Surgeons 19, no. 2 (2019): 564–572, 10.1111/ajt.15144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Saberi B., Gurakar A., Tamim H., et al., “Racial Disparities in Candidates for Hepatocellular Carcinoma Liver Transplant After 6‐Month Wait Policy Change,” JAMA Network Open 6, no. 11 (2023): e2341096, 10.1001/jamanetworkopen.2023.41096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Wehrle C. J., Kusakabe J., Akabane M., et al., “Expanding Selection Criteria in Deceased Donor Liver Transplantation for Hepatocellular Carcinoma: Long‐Term Follow‐up of a National Registry and 2 Transplant Centers,” Transplantation 108, no. 12 (2024): 2386–2395, 10.1097/TP.0000000000005097. [DOI] [PubMed] [Google Scholar]
- 12. Cron D. C., Mazur R. D., Bhan I., Adler J. T., and Yeh H., “Sex and Size Disparities in Access to Liver Transplant for Patients With Hepatocellular Carcinoma,” JAMA Surgery 159, no. 11 (2024): 1291–1298, 10.1001/jamasurg.2024.3498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. John B. V., Schwartz K., Scheinberg A. R., et al., “Evaluation Within Thirty Days of Referral for Liver Transplantation Is Associated with Reduced Mortality: A Multi‐Center Analysis of Patients Referred Within the VA Health System,” Transplantation 106, no. 1 (2022): 72–84, 10.1097/TP.0000000000003615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Ross‐Driscoll K., Ayuk‐Arrey A. T., Lynch R., et al., “Disparities in Access to Liver Transplant Referral and Evaluation Among Patients With Hepatocellular Carcinoma in Georgia,” Cancer Research Communications 4, no. 4 (2024): 1111–1119, 10.1158/2767-9764.CRC-23-0541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Epic Cosmos Accessed April 16, 2024, https://cosmos.epic.com/.
- 16. Strauss A. T., Terlizzi K., Orandi B., et al., “Evaluating the Representativeness and Validity of Cosmos as a Novel, Large‐Scale, Real‐World Data Source for Liver Transplant Research,” Liver Transplantation: Official Publication of the American Association for the Study of Liver Diseases and the International Liver Transplantation Society 32(2026): 419–427, 10.1097/LVT.0000000000000728. [DOI] [PubMed] [Google Scholar]
- 17. Lai Q., Vitale A., Iesari S., et al., “Intention‐to‐Treat Survival Benefit of Liver Transplantation in Patients With Hepatocellular Cancer,” Hepatology 66, no. 6 (2017): 1910–1919, 10.1002/hep.29342. [DOI] [PubMed] [Google Scholar]
- 18. Vaz J., Hagström H., Eilard M. S., Rizell M., and Strömberg U., “Socioeconomic Inequalities in Diagnostics, Care and Survival Outcomes for Hepatocellular Carcinoma in Sweden: A Nationwide Cohort Study,” Lancet Regional Health—Europe 52 (2025): 101273, 10.1016/j.lanepe.2025.101273. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data sharing not applicable to this article as data was derived from Epic Cosmos and only those from participating institutions have access to this data.