Abstract
Background
In March 2023, the Composite Allocation Score (CAS) and continuous distribution replaced the Lung Allocation Score (LAS) for lung transplantation in the United States in part to reduce the role of geography. We examined whether ZIP-code-level socioeconomic drivers of waitlisted patients were associated with time to transplant or waitlist attrition, defined as death or removal due to being too sick.
Methods
Candidates waitlisted between January 1, 2021, and November 29, 2025 in the Scientific Registry of Transplant Recipients were included. Place-based measures were the Social Deprivation Index (SDI) and Local Internet Availability (LIA). Fine-Gray competing risks regression models were used. The primary outcome was time from listing to transplant or waitlist attrition.
Results
Among 15,178 patients, 6178 were from the LAS era and 9000 during CAS. The SDI did not differ between eras (43.0 during LAS and 45.0 during CAS (p =0.18)), and LIA modestly decreased (8.4% during LAS and 6.3% during CAS (p <0.001)). In the CAS era, candidates from low LIA areas, when compared to moderate/high LIA areas, were more likely to have obstructive lung disease, reside outside major metropolitan counties, and travel farther to their listing center. After multivariate modeling, there was no association between SDI or LIA and time from listing to transplant or waitlist attrition.
Discussion
In this early evaluation of CAS, place-driven factors at the zip-code-level were not associated with lung transplant waitlist outcomes after CAS implementation. However, descriptive differences by candidate LIA areas warrant continued monitoring and longer-term follow-up.
Keywords: lung transplant, allocation, composite allocation score, lung allocation score, social deprivation index
Introduction
Lung transplantation is an established, life-saving therapy for patients with end-stage lung disease. However, the demand for donor lungs continues to exceed supply. Therefore, donor lung allocation systems play a critical role in determining which patients receive transplants and when. In the United States, lung allocation policies are designed to balance urgency, utility, and equity across diverse patient populations1. Since allocation policies directly influence access to a scarce resource, even subtle structural features of these systems can produce unintended disparities.
Historically, lung allocation in the United States was governed by the Lung Allocation Score (LAS), which prioritized candidates based on a combination of predicted waitlist mortality and post-transplant survival benefit2. Although the LAS was an advance over an earlier time-based allocation system, concerns emerged regarding place-driven inequities and the extent to which candidate location influenced the likelihood of transplant3. Thus, on March 9, 2023, the Organ Procurement and Transplantation Network (OPTN) implemented the Composite Allocation Score (CAS) and continuous distribution. CAS uses a points-based framework that incorporates multiple domains, including medical urgency, post-transplant survival, candidate biology (including blood type compatibility), patient access, and placement efficiency, which captures donor-to-transplant hospital travel distance, with the goal of improving consistency and equity in organ allocation4. Notably, CAS and continuous distribution’s de-emphasis of geography includes the placement efficiency domain, potentially lessening the role of place-based factors tied to travel burden or regional healthcare infrastructure. This change provides the conceptual basis for our comparison between eras.
Where a patient lives is recognized as an important social driver of health5, 6. Place-based socioeconomic conditions, such as social deprivation, have been associated with disparities across a wide range of health outcomes7, 8, 9. In organ transplantation, these factors may influence healthcare through multiple mechanisms, such as availability of specialty services and transportation barriers. This includes travel to distant transplant centers, a burden common for patients with more complex conditions or those living in resource-limited areas. Our prior work examining place-based factors post-lung transplant relied on county-level measures of socioeconomic drivers10, which demonstrated an association with post-transplant outcomes. Early evaluations of CAS’s pre-transplant equity implications have focused on individual factors11. It is essential to understand whether place-based socioeconomic conditions continue to influence waitlist outcomes under CAS in order to assess if the new system achieved its goal to reduce geographic importance without introducing unintended disparities.
Beyond the established measures of social drivers, such as the Social Deprivation Index (SDI), Local Internet Availability (LIA) is an increasingly important dimension of healthcare equity. It has been associated with reduced healthcare engagement and disparities in accessing specialty care, especially in disadvantaged populations12. It may affect transplant-related outcomes through patient communication with transplant teams, telehealth visits, and follow-up care.
The objective of this study was to evaluate the association between ZIP-code level social drivers and local internet availability with time to lung transplantation or waitlist attrition among adult candidates listed during the LAS and CAS eras. We hypothesized that place-based factors would not be associated with waitlist outcomes after the transition to CAS.
Methods
Data
This study used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donors, wait-listed candidates, and transplant recipients in the US, submitted by the members of the Organ Procurement and Transplantation Network. The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors. The SRTR standard analysis files released in March 2026 were used. Candidate zip codes and LAS supplemental files were used to incorporate place-based variables.
We included adult (≥18 years) candidates listed for lung-only transplant between January 1, 2021, and November 29, 2025. This period was selected to include data before and after CAS implementation on March 9, 2023. Exclusion criteria were prior organ transplant, multiple organ listing, candidates who primarily reside outside the U.S., individuals traveling to the U.S. exclusively for transplantation, missing or invalid zip code, and repeat listings during the study period (Supplemental Figure 1). This study was approved by the Cleveland Clinic Institutional Review Board (#23–012).
Outcome
The primary outcome was time from listing to lung transplant with waitlist attrition (death or removal due to being too sick) treated as a competing risk. The candidate cohort censoring date (November 30, 2025) was used for censoring and follow-up was truncated at 15 months. In addition, candidates listed during the LAS era who were still on the waitlist at the time of CAS implementation (March 9, 2023) were censored as event-free at that time.
Exposures
We linked SRTR data to two place-level measures selected a priori, using a candidate’s five-digit ZIP code of residence provided in the SRTR standard analysis files:
-
1.
Social Deprivation Index (SDI, 2019), developed by the Robert Graham Center, is a composite tool that uses seven American Community Survey (ACS) demographic indicators covering poverty, education, single parent households, rental housing, overcrowded housing, transportation, and employment, to measure neighborhood-level socioeconomic disadvantage, linked at the candidate’s residential zip code level using the ZIP Code Tabulation Area (ZCTA)-zip crosswalk13, scaled 0–100 (least to most deprived).
-
2.
Local Internet Availability (LIA) (AHRQ Social Determinants of Health SDOH Database) was measured as the percentage of households with no internet availability; 2021 and 2023 values were used for the LAS and CAS cohorts, respectively14. We divided candidates into low and moderate/high LIA groups based on the median overall percentage of households without internet availability, with candidates residing in areas at or above the median classified as low LIA.
Covariates
We used Rural Urban Continuum Codes (RUCC 2023) from the US Department of Agriculture to classify the degree of urbanization of residential counties15. In addition, the distance between each candidate’s permanent zip code to their listing center was calculated as the geodetic distance between the zip code centroids in nautical miles (NM).
Models were adjusted for candidate-level demographic and clinical characteristics including sex, age, race/ethnicity, height, body mass index category, ABO blood type, highest education level, primary insurance type, and native lung disease diagnosis group, extracorporeal membrane oxygenation (ECMO), ventilation, RUCC, candidate-to-center distance, and listing center average annual listing volume.
Statistics
Continuous variables were summarized as medians with interquartile ranges, and categorical variables as frequencies and percentages. Between-group differences were compared using Wilcoxon rank-sum and chi-square tests, respectively. Absolute differences for continuous SDOH measures were estimated using Hodges–Lehmann location shifts and 95% confidence intervals.
We evaluated associations between place-level exposure and outcomes using Fine-Gray competing risks regression models, accounting for the competing risks of lung transplant and waitlist attrition. Nonlinear associations were modeled using restricted cubic splines with three knots at the 25th, 50th, and 75th percentiles. Hazard ratios and 95% confidence intervals were plotted across the 5th to 95th percentile range of exposures. All models were adjusted for covariates described above. To formally evaluate if the association between LIA and outcomes differed across eras, we tested an era-by-internet-availability interaction term. Models were further stratified by native lung disease groups. See supplement for full details. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).
Missing data
Data completeness was high, with <3% missingness for all variables and a maximum of 2.5% for education. To address this, we used multivariate imputation by chained equations (MICE) to create five imputed datasets including all candidate-level and area-level variables. Analyses were performed on each dataset, and estimates were pooled to generate final effect estimates.
Results
Patient population
A total of 15,178 patients met inclusion criteria, comprising 6178 listed during the LAS era and 9000 during the CAS era. Table 1 summarizes demographic and clinical characteristics of lung transplant candidates.
Table 1.
Characteristics of Lung Transplant Candidates, Before and After Lung CAS
| Factor | Total(N=15,178) | LAS era(N=6178) | Lung CAS era(N=9000) | p-value |
|---|---|---|---|---|
| Age at listing (years) | <0.001c | |||
| 18–39 | 945 (6.2) | 425 (6.9) | 520 (5.8) | |
| 40–49 | 1204 (7.9) | 535 (8.7) | 669 (7.4) | |
| 50–59 | 3191 (21.0) | 1416 (22.9) | 1775 (19.7) | |
| 60–69 | 7550 (49.7) | 2962 (47.9) | 4588 (51.0) | |
| ≥70 | 2288 (15.1) | 840 (13.6) | 1448 (16.1) | |
| Sex | 0.45c | |||
| Female | 6079 (40.1) | 2452 (39.7) | 3627 (40.3) | |
| Male | 9099 (59.9) | 3726 (60.3) | 5373 (59.7) | |
| Race/ethnicity* | 0.034c | |||
| Non-Hispanic White | 10,462 (69.9) | 4374 (70.8) | 6088 (69.2) | |
| Non-Hispanic Black | 1564 (10.4) | 609 (9.9) | 955 (10.9) | |
| Non-Hispanic Other and Multi-racial | 730 (4.9) | 317 (5.1) | 413 (4.7) | |
| Hispanic | 2216 (14.8) | 878 (14.2) | 1338 (15.2) | |
| Education* | 0.001c | |||
| High school or less | 6080 (41.1) | 2381 (39.3) | 3699 (42.3) | |
| Some college | 3781 (25.5) | 1600 (26.4) | 2181 (24.9) | |
| College or more | 4943 (33.4) | 2075 (34.3) | 2868 (32.8) | |
| Primary insurance* | <0.001c | |||
| Private | 6183 (40.7) | 2654 (43.0) | 3529 (39.2) | |
| Medicare | 7066 (46.6) | 2751 (44.5) | 4315 (48.0) | |
| Medicaid | 1365 (9.0) | 558 (9.0) | 807 (9.0) | |
| Other | 561 (3.7) | 215 (3.5) | 346 (3.8) | |
| BMI* | 0.35c | |||
| <18.5 | 435 (2.9) | 174 (2.8) | 261 (2.9) | |
| 18.5–24.9 | 5081 (33.6) | 2046 (33.3) | 3035 (33.8) | |
| 25–29.9 | 6007 (39.7) | 2501 (40.7) | 3506 (39.1) | |
| 30–34.9 | 3400 (22.5) | 1346 (21.9) | 2054 (22.9) | |
| ≥35 | 199 (1.3) | 79 (1.3) | 120 (1.3) | |
| Disease etiology | <0.001c | |||
| Group A (Obstructive lung disease) | 3006 (19.8) | 1119 (18.1) | 1887 (21.0) | |
| Group B (pulmonary vascular disease) | 642 (4.2) | 322 (5.2) | 320 (3.6) | |
| Group C (cystic fibrosis) | 235 (1.5) | 104 (1.7) | 131 (1.5) | |
| Group D (restrictive lung disease) | 11,295 (74.4) | 4633 (75.0) | 6662 (74.0) | |
| Diabetes* | 2974 (19.6) | 1146 (18.6) | 1828 (20.3) | 0.007c |
| On ECMO | 810 (5.3) | 386 (6.2) | 424 (4.7) | <0.001c |
| On ventilator | 489 (3.2) | 264 (4.3) | 225 (2.5) | <0.001c |
| On life support* | 1378 (9.1) | 587 (9.5) | 791 (8.8) | 0.13c |
| Blood type | 0.60c | |||
| A | 5690 (37.5) | 2299 (37.2) | 3391 (37.7) | |
| AB | 532 (3.5) | 213 (3.4) | 319 (3.5) | |
| B | 1736 (11.4) | 732 (11.8) | 1004 (11.2) | |
| O | 7220 (47.6) | 2934 (47.5) | 4286 (47.6) | |
| LAS at listing (match) | 39.7 [35.3, 50.1] | 39.7 [35.3, 50.1] | n/a | |
| Center listing volume in study period | 358.0 [250.0, 485.0] | 364.0 [250.0, 485.0] | 358.0 [250.0, 485.0] | 0.70b |
| Center listing average volume per year | 59.7 [41.7, 80.8] | 60.7 [41.7, 80.8] | 59.7 [41.7, 80.8] | 0.72b |
| Distance from residence to listing center (NM)* | 42.0 [15.4, 105.8] | 41.5 [15.6, 106.4] | 42.3 [15.2, 105.6] | 0.57b |
| RUCC (2023)* | 0.27c | |||
| Metro - 1M+ population | 8845 (58.8) | 3637 (59.4) | 5208 (58.5) | |
| Metro - 250K-1M population | 2748 (18.3) | 1125 (18.4) | 1623 (18.2) | |
| Metro - <250 K population | 1311 (8.7) | 528 (8.6) | 783 (8.8) | |
| Nonmetro, adjacent to a metro area | 1492 (9.9) | 597 (9.8) | 895 (10.0) | |
| Nonmetro, not adjacent to a metro area | 634 (4.2) | 233 (3.8) | 401 (4.5) | |
| Social Deprivation Index (2019) | 44.0 [21.0, 69.0] | 43.0 [22.0, 68.0] | 45.0 [21.0, 70.0] | 0.18b |
| Pct. of households with no internet access* | 7.2 [4.3, 11.2] | 8.4 [5.2, 12.9] | 6.3 [3.8, 9.9] | <0.001b |
BMI: Body Mass Index; CAS: Composite Allocation Score; ECMO: extracorporeal membrane oxygenation; K: thousand; LAS: Lung Allocation Score; M: million; OPTN: Organ Procurement & Transplantation Network; SDI: Social Deprivation Index; NM: Nautical Miles
Data not available for all subjects. Missing values: Race/ethnicity = 206; Education = 374; Citizenship status = 2; Primary insurance = 3; BMI = 56; Diabetes = 10; On life support = 2; Distance from residence to listing center (NM) = 9; RUCC (2023) = 148; Pct. of households with no internet access = 8. Statistics presented as Median [P25, P75], N (column %). p-values: b=Wilcoxon Rank Sum test, c=Pearson's chi-square test.
The proportion of female patients was 39.7% in the LAS era and 40.3% in the CAS era. Patients listed during CAS were significantly older at listing, with a higher proportion aged 60–69 compared to the LAS group (47.9% LAS vs 51.0% CAS; p<0.001). The majority identified as non-Hispanic White (70.8% vs 69.2%, LAS vs CAS, respectively), followed by Hispanic (14.2% vs 15.2%) and non-Hispanic Black (9.9% vs 10.9%). Educational attainment differed slightly between eras, with a higher proportion reporting a high-school education or less after CAS implementation (39.3% LAS vs 42.3% CAS).
From the LAS era to CAS era, Medicare coverage increased from 44.5% to 48.0%, while private insurance decreased from 43.0% to 39.2%. The distribution of native lung diseases was consistent across eras, with restrictive lung disease (Group D) accounting for the majority of listings (75.0% LAS vs 74.0% CAS) and obstructive lung disease (Group A) comprising 18.1% and 21.0%, respectively. Markers of medical acuity differed between eras: ECMO use (6.2% LAS vs. 4.7% CAS, p<0.001) and mechanical ventilation (4.3% vs. 2.5%, p<0.001) were lower in the CAS era.
Waitlist outcomes
During the study period, patients were followed from listing until transplant, waitlist attrition, or censoring at the end date. Time to event reflected a competing risks framework, where transplantation and waitlist attrition were treated as mutually exclusive competing events. Supplemental Figure 2 displays the unadjusted distribution of waitlist outcomes by allocation era. The cumulative incidence of lung transplant and waitlist attrition was similar between LAS and CAS eras. Median follow-up time was slightly longer for candidates listed in the LAS era (median [IQR]: 1.05 [0.36, 3.25] vs. 0.85 [0.26, 2.63] months).
Place-based differences
The SDI did not differ between eras (43.0 during LAS and 45.0 during CAS (p =0.18)). LIA was significantly lower for candidates listed in the CAS era (8.4% during LAS and 6.3% during CAS; p<0.001), with a Hodges-Lehmann estimate of a difference of −1.9 percentage points (95% CI: −2.1, −1.8, Supplemental Figure 3).
Candidates were divided into low and moderate/high LIA groups based on the median percentage of households without internet availability (Table 2). Candidates residing in low LIA areas were more likely to be younger, Non-Hispanic Black, have a high school education or less, and covered by Medicaid compared to those in moderate/high LIA areas. Low LIA candidates more likely had obstructive lung disease (Group A), while those in moderate/high LIA areas more likely had restrictive lung disease (Group D). Candidates with low LIA were also less likely to reside in a large metropolitan county with 1M+ population (44.2% for low LIA vs. 73.6% for moderate/high LIA) and had a longer median distance from their residence to the listing center (59.1NM for low LIA vs. 28.8NM for moderate/high LIA). Additionally, candidates from areas with lower LIA were more likely from centers with lower listing volumes.
Table 2.
Characteristics of Lung Transplant Candidates, by Placed-based Internet Access
| Factor | Moderate/High Internet Access(N=7580) | Low Internet Access(N=7590) | p-value |
|---|---|---|---|
| Pct. of households with no internet access | 4.3 [2.9, 5.6] | 11.1 [8.8, 14.9] | --- |
| Age at listing (years) | <0.001c | ||
| 18–39 | 443 (5.8) | 500 (6.6) | |
| 40–49 | 581 (7.7) | 622 (8.2) | |
| 50–59 | 1474 (19.4) | 1717 (22.6) | |
| 60–69 | 3771 (49.7) | 3775 (49.7) | |
| ≥70 | 1311 (17.3) | 976 (12.9) | |
| Sex | 0.005c | ||
| Female | 2951 (38.9) | 3125 (41.2) | |
| Male | 4629 (61.1) | 4465 (58.8) | |
| Race/ethnicity* | <0.001c | ||
| Non-Hispanic White | 5253 (70.5) | 5207 (69.4) | |
| Non-Hispanic Black | 630 (8.4) | 933 (12.4) | |
| Non-Hispanic Other and Multi-racial | 506 (6.8) | 223 (3.0) | |
| Hispanic | 1067 (14.3) | 1145 (15.3) | |
| Education* | <0.001c | ||
| High school or less | 2454 (33.2) | 3625 (48.9) | |
| Some college | 1896 (25.7) | 1880 (25.4) | |
| College or more | 3039 (41.1) | 1902 (25.7) | |
| Primary insurance* | <0.001c | ||
| Private | 3345 (44.1) | 2835 (37.4) | |
| Medicare | 3463 (45.7) | 3601 (47.5) | |
| Medicaid | 510 (6.7) | 852 (11.2) | |
| Other | 260 (3.4) | 301 (4.0) | |
| BMI* | <0.001c | ||
| <18.5 | 239 (3.2) | 196 (2.6) | |
| 18.5–24.9 | 2683 (35.5) | 2394 (31.7) | |
| 25–29.9 | 2997 (39.7) | 3008 (39.8) | |
| 30–34.9 | 1547 (20.5) | 1851 (24.5) | |
| ≥35 | 87 (1.2) | 112 (1.5) | |
| Disease etiology | <0.001c | ||
| Group A (Obstructive lung disease) | 1335 (17.6) | 1671 (22.0) | |
| Group B (pulmonary vascular disease) | 305 (4.0) | 337 (4.4) | |
| Group C (cystic fibrosis) | 127 (1.7) | 106 (1.4) | |
| Group D (restrictive lung disease) | 5813 (76.7) | 5476 (72.1) | |
| Diabetes* | 1466 (19.4) | 1507 (19.9) | 0.42c |
| On ECMO | 395 (5.2) | 415 (5.5) | 0.48c |
| On ventilator | 246 (3.2) | 242 (3.2) | 0.84c |
| On life support* | 658 (8.7) | 718 (9.5) | 0.095c |
| Blood type | 0.12c | ||
| A | 2825 (37.3) | 2864 (37.7) | |
| AB | 283 (3.7) | 248 (3.3) | |
| B | 901 (11.9) | 834 (11.0) | |
| O | 3571 (47.1) | 3644 (48.0) | |
| Center listing volume in period | 364.0 [251.0, 485.0] | 356.0 [226.0, 486.0] | <0.001b |
| Center listing average volume per year | 60.7 [41.8, 80.8] | 59.3 [37.7, 81.0] | <0.001b |
| Distance from residence to listing center (NM)* | 28.8 [14.1, 81.1] | 59.1 [19.6, 128.6] | <0.001b |
| RUCC (2023)* | <0.001c | ||
| Metro - 1M+ population | 5512 (73.6) | 3328 (44.2) | |
| Metro - 250K-1M population | 1314 (17.5) | 1433 (19.0) | |
| Metro - <250 K population | 419 (5.6) | 892 (11.8) | |
| Nonmetro, adjacent to a metro area | 172 (2.3) | 1318 (17.5) | |
| Nonmetro, not adjacent to a metro area | 72 (0.96) | 562 (7.5) | |
| Social Deprivation Index (2019) | 27.0 [13.0, 48.0] | 61.0 [39.0, 82.0] | <0.001b |
| Allocation policy at time of listing | <0.001c | ||
| LAS era | 2478 (32.7) | 3698 (48.7) | |
| CAS era | 5102 (67.3) | 3892 (51.3) |
BMI: Body Mass Index; CAS: Composite Allocation Score; ECMO: extracorporeal membrane oxygenation; K: thousand; LAS: Lung Allocation Score; M: million; OPTN: Organ Procurement & Transplantation Network; SDI: Social Deprivation Index; NM: Nautical Miles
Data not available for all subjects. Missing values: Race/ethnicity = 206; Education = 374; Citizenship status = 2; Primary insurance = 3; BMI = 56; Diabetes = 10; On life support = 2; Distance from residence to listing center (NM) = 9; RUCC (2023) = 148; Pct. of households with no internet access = 8. Statistics presented as Median [P25, P75], N (column %). p-values: b=Wilcoxon Rank Sum test, c=Pearson's chi-square test.
Modeling place-based factors and outcomes
In both LAS and CAS eras, the SDI was not significantly associated with either time to transplant or waitlist attrition, with adjusted sub-distribution hazard ratios (sHRs) near 1.0 across the full SDI range (Supplemental Figure 4). By contrast, a higher percentage of households without internet availability were associated with a modestly higher likelihood of transplant in the LAS era (p=0.003) but not in the CAS era (p=0.31). No significant associations were observed for waitlist attrition in either period (Figure 1). A formal test for interaction between allocation era and internet availability was not statistically significant for either outcome (p=0.083 for transplant, p=0.57 for attrition), indicating no association by allocation era.
Figure 1.
Adjusted association between area-level percentage of households with no internet access and listing outcomes, before and after CAS. CAS: Composite Allocation Score; LAS: Lung Allocation Score. Estimates presented are adjusted for sex, age, race/ethnicity, height, body mass index category, ABO blood type, highest education level, primary insurance type, native lung disease diagnosis group, on extracorporeal membrane oxygenation (ECMO), ventilation, RUCC, candidate-to-center distance, and listing center average annual listing volume. Lung transplant and waitlist attrition were treated as mutually exclusive competing risks. To consider the nonlinear effects of area-level percentage of households with no internet access, we used restricted cubic spline terms with 3 knots placed at the 25th, 50th, and 75th percentiles. The adjusted subdistribution hazard ratios displayed range from the observed 5th to the 95th percentiles.
In a sub analysis, among candidates with group A lung disease the association between low LIA and likelihood of transplant followed a similar direction to the overall cohort during the LAS era, though was not statistically significant (p=0.068; Supplemental Figure 5). No significant association was observed in the CAS era or for waitlist attrition in either era. No significant association was observed among candidates with group B lung disease (Supplemental Figure 6). Among candidates with group D lung disease, likelihood of transplant was similar to the overall cohort in the LAS era (p=0.006) but not CAS (p=0.23). A higher percentage of candidates with low LIA were associated with significantly higher waitlist attrition in the CAS era (p=0.026), but not LAS (p=0.93; Supplemental Figure 7). Candidates with group C lung disease were not included because of small sample size.
Results were consistent when allocation era was modeled as a time-varying covariate, with a significant interaction between allocation era and LIA (p=0.031, Supplemental Figure 8).
Discussion
Cohort overview
In this nationally representative cohort, several characteristics differed modestly between LAS and CAS eras, including education level, insurance type, and markers of medical acuity (ECMO use and mechanical ventilation). Waitlist outcomes (time to transplant vs. attrition) were similar between both allocation eras, and the median time to event was not significantly different.
Place-based findings
Candidates from low LIA areas modestly declined between the LAS and CAS eras, and candidates residing in these areas differed in clinically meaningful ways from those in higher LIA areas. Low LIA candidates were more likely to have obstructive lung disease, reside outside major metropolitan counties, and travel farther to their listing center. These differences identify a candidate population that warrants monitoring as more data become available.
Modeling that included the SDI and LIA as separate exposures were not associated with waitlist outcomes during the CAS era. The absence of associations during the pre-transplant waitlisted stage suggests that individual clinical characteristics, center-level practices, and upstream referral patterns to the transplant center may play an influential role in waitlist outcomes.
Prior work
Evidence from other solid-organ transplant populations highlights that place-based factors can shape transplant outcomes. In kidney transplantation, community-level socioeconomic factors among waitlisted patients have been associated with longer time to transplant16. Similarly, in liver transplantation, place-based social drivers have been linked to disparities in waitlist mortality17. Our findings in lung transplantation did not mirror this pattern which may represent differences in disease acuity and trajectory, prioritization mechanisms, or waitlist management specific to this population.
Prior literature has described limited internet access as a barrier to healthcare engagement, associated with reduced utilization of preventative and specialty services among socioeconomically disadvantaged populations18, 19. Since the candidates in our study were also more likely to reside outside major metropolitan counties and travel farther to their listing center, this association may partly reflect broader geographic patterning instead of an actual effect of connectivity itself.
Strengths
Using a national dataset from the SRTR, we assessed waitlist outcomes for lung transplant candidates across the United States, resulting in generalizable findings. The integration of these data with SDI ZIP-code level data allowed for finer place-based detail and better characterization of candidates than previous analyses at the county-level. Incorporating Fine-Gray competing-risk models strengthened our analytic approach by appropriately treating transplant and waitlist attrition as mutually exclusive outcomes. Multiple imputation also minimized the risk of bias due to missing data.
Limitations
We recognize there are limitations to this study. First, as an observational study, we were unable to establish causation, and there remains a possibility of unmeasured confounding even after adjustment for clinical and demographic characteristics. The lower proportion of candidates on ECMO or mechanical ventilation in the CAS-era cohort increases the possibility of residual confounding by clinical acuity, which could affect both transplant likelihood and waitlist attrition. We also did not include measures of donor organ availability or geographic region, which have the potential to influence transplant likelihood and could cause residual confounding beyond the covariates accounted for in our models. We did not include transplant-center as a fixed effect given the large number of centers relative to the number of events, which would have resulted in unstable parameter estimation.
Second, the SRTR only captures the most recent ZIP code reported by the transplant center and does not account for residential history, which may contribute to patients’ health outcomes20, 21. Third is the ecological bias in which measures such as SDI and LIA observed at the area level may not hold at the individual level22. Additional limitations include the modifiable areal unit problem and Uncertain Geographic Context Problem (UGCoP), where ZIP code-defined neighborhoods may not align with actual neighborhood environments, resulting in area misclassification, and the outcome may vary depending on the geographic boundaries used23, 24. Fourth, SRTR data availability begins at waitlisting, meaning that socioeconomic and place-based inequities occurring earlier in the transplant journey, including referral and evaluation, were not included. Finally, maximum potential accrual periods differed between the two cohorts; however, potential observation bias was minimized by utilizing time-to-event analyses with follow-up administratively truncated at 15 months, yielding closely aligned median follow-up times. Taken together, these factors highlight the need for future studies with longer term follow-up using granular neighborhood indicators to better characterize the influence of community conditions on transplant outcomes, and for registries that link transplant referral evaluation data to better understand disparities that occur prior to transplant listing.
Conclusion
The area-based SDI among pre-lung transplant waitlisted candidates did not change from the LAS to CAS eras. Candidates residing in lower LIA areas were modestly less represented on the CAS era waitlist and differed from candidates residing in higher LIA areas such as with higher rates of obstructive lung disease and greater distance to their listing center. These differences were not associated with waitlist outcomes. Whether the differences observed translate to meaningful outcome differences over a longer period of follow-up remains an important question for future study.
Acknowledgments and disclosures
The authors have no disclosures regarding conflict-of-interests. The research reported in this publication was supported by the National Heart, Lung, and Blood Institute under award number R01HL164660. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The data reported here have been supplied by the Hennepin Healthcare Research Institute (HHRI) as the contractor for the Scientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government.
Declaration of Generative AI and AI-assisted technologies in the writing process
The author(s) declare that no AI or AI-assisted technologies were used in the writing or editing of this manuscript.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Supplemental data associated with this article can be found in the online version at doi:10.1016/j.jhlto.2026.100694.
Appendix A. Supplemental material
Supplementary material
References
- 1.Egan T.M. How should lungs be allocated for transplant? Semin Respir Crit Care Med. 2018;39(2):126–137. doi: 10.1055/s-0037-1620265. Epub 2018 Mar 26. PMID: 29579766. [DOI] [PubMed] [Google Scholar]
- 2.Egan T.M., Edwards L.B. Effect of the lung allocation score on lung transplantation in the United States. J Heart Lung Transplant. 2016;35(4):433–439. doi: 10.1016/j.healun.2016.01.010. Epub 2016 Jan 15. PMID: 26922274. [DOI] [PubMed] [Google Scholar]
- 3.Benvenuto L.J., Anderson D.R., Kim H.P., et al. From the Columbia University Lung Transplant Program. Geographic disparities in donor lung supply and lung transplant waitlist outcomes: A cohort study. Am J Transplant. 2018;18(6):1471–1480. doi: 10.1111/ajt.14630. Epub 2018 Jan 27. PMID: 29266733. [DOI] [PubMed] [Google Scholar]
- 4.Alderete I.S., Medina C.K., Halpern S.E., et al. Implications of the composite allocation score system for lung transplantation in the United States: review of the new system. Semin Thorac Cardiovasc Surg. 2025;37(3):329–333. doi: 10.1053/j.semtcvs.2024.09.001. Epub 2024 Sep 23. PMID: 39322124. [DOI] [PubMed] [Google Scholar]
- 5.Richardson D.B., Volkow N.D., Kwan M.P., et al. Medicine. Spatial turn in health research. Science. 2013;339(6126):1390–1392. doi: 10.1126/science.1232257. PMID: 23520099; PMCID: PMC3757548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.U.S. Department of Health and Human Services, Office of Disease Prevention and Health Promotion. Social Determinants of Health. Healthy People 2030. Accessed https://odphp.health.gov/healthypeople/priority-areas/social-determinants-health.
- 7.Mupfudze T.G., Martinez A., Noreen S.M., et al. Individual-level social determinants of health and disparities in access to kidney transplant and waitlist mortality. PLoS One. 2024;19(8) doi: 10.1371/journal.pone.0308407. PMID: 39167588; PMCID: PMC11338441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jackson W.E., Lopez R., Forman L.M., et al. Marked variation in disease acuity and outcomes on the liver transplant waiting list by sociodemographic characteristics. Liver Transpl. 2025;31(5):584–595. doi: 10.1097/LVT.0000000000000502. Epub 2024 Oct 3. PMID: 39356520. [DOI] [PubMed] [Google Scholar]
- 9.Amdani S., Tang A., Schold J.D. Children from socioeconomically disadvantaged communities present in more advanced heart failure at the time of transplant listing. J Heart Lung Transplant. 2023;42(2):150–155. doi: 10.1016/j.healun.2022.09.007. Epub 2022 Sep 18. PMID: 36270922. [DOI] [PubMed] [Google Scholar]
- 10.Tsuang W.M., Lopez R., Tang A., et al. Place-based heterogeneity in lung transplant recipient outcomes. Am J Transplant. 2022;22(12):2981–2989. doi: 10.1111/ajt.17170. Epub 2022 Aug 25. PMID: 35962587; PMCID: PMC11362981. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Park S.Y., LaVanchy R., Mott N.M., et al. Racial, ethnic, and sex disparities in lung transplant waitlist outcomes before and after composite allocation score implementation. JHLT Open. 2025;11 doi: 10.1016/j.jhlto.2025.100441. PMID: 41439229; PMCID: PMC12721069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cuadros D.F., Moreno C.M., Miller F.D., et al. Assessing access to digital services in health care-underserved communities in the United States: a cross-sectional study. Mayo Clin Proc Digit Health. 2023;1(3):217–225. doi: 10.1016/j.mcpdig.2023.04.004. PMID: 40206610; PMCID: PMC11975688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Social deprivation index (SDI). Robert Graham Center - Policy Studies in Family Medicine & Primary Care. (2018, November 5). Retrieved February 28, 2022, from https://www.graham-center.org/rgc/maps-data-tools/sdi/social-deprivation-index.html.
- 14.Social Determinants of Health Database. Agency for Healthcare Research and Quality, Rockville, MD. https://www.ahrq.gov/sdoh/data-analytics/sdoh-data.html. Retrieved July 26, 2023.
- 15.Rural-Urban Continuum Codes 2023. United States Department of Agriculture, Economic Research Service, Washington, DC. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes.aspx. Retrieved February 6, 2024.
- 16.Li Y., Menon G., Kim B., et al. Residential Neighborhood disadvantage and access to kidney transplantation. JAMA Netw Open. 2025;8(12) doi: 10.1001/jamanetworkopen.2025.49679. PMID: 41468017; PMCID: PMC12754678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zhou K., Lit A., Kuo L.S., et al. Neighborhood-level social determinants of health and waitlist mortality for liver transplantation: the liver outcomes and equity index. Transplantation. 2024;108(7):1558–1569. doi: 10.1097/TP.0000000000004888. Epub 2024 Jun 20. PMID: 38049937; PMCID: PMC11150328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wakeman M., Buckman D.W., El-Toukhy S. Disparities in digital health care use in 2022. JAMA Netw Open. 2025;8(4) doi: 10.1001/jamanetworkopen.2025.5359. PMID: 40244585; PMCID: PMC12550827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hung P., Granger M., Boghossian N., et al. Dual barriers: examining digital access and travel burdens to hospital maternity care access in the United States, 2020. Milbank Q. 2023;101(4):1327–1347. doi: 10.1111/1468-0009.12668. Epub 2023 Aug 23. PMID: 37614006; PMCID: PMC10726888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tsuang W.M., Curtis J. Towards more precise neighborhood determinants of health: collecting residential histories as an important next step. J Gen Intern Med. 2024;39(13):2576–2577. doi: 10.1007/s11606-024-08884-7. Epub 2024 Jun 27. PMID: 38937367; PMCID: PMC11436546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Tsuang W.M., Curtis J. Residential history, the electronic health record, and clinical medicine: A descriptive study in a lung transplant cohort. Health Place. 2024;89 doi: 10.1016/j.healthplace.2024.103306. Epub 2024 Jun 28. PMID: 38943794; PMCID: PMC11920707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wakefield J. Ecologic studies revisited. Annu Rev Public Health. 2008;29:75–90. doi: 10.1146/annurev.publhealth.29.020907.090821. [DOI] [PubMed] [Google Scholar]
- 23.Gehlke C.E., Biehl K. Certain effects of grouping upon the size of the correlation coefficient in census tract material. J Am Stat Assoc. 1934;29(185A):169–170. doi: 10.2307/2277827. [DOI] [Google Scholar]
- 24.Kwan M.P. The uncertain geographic context problem. Ann Assoc Am Geogr. 2012;102(5):958–968. doi: 10.1080/00045608.2012.687349. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary material