Abstract
There is a broad range of patient travel distances to reach a lung transplant hospital in the United States. Whether patient travel distance is associated with waitlist outcomes is unknown. We present a cohort study of patients listed between January 1, 2006 and May 31, 2017 using the Scientific Registry of Transplant Recipients. Travel distance was measured from the patient’s permanent zip code to the transplant hospital using SAS URL access to Google Maps, and assessed using multivariable competing risk regression models. There were 22,958 patients who met inclusion criteria. Median travel distance was 69.7 miles. Among patients who traveled >60 miles, 41.2% bypassed a closer hospital and sought listing at a more distant hospital. In the adjusted models, when compared to patients who traveled ≤60 miles, patients who traveled >360 miles had a 27% lower subhazard ratio (SHR) for waitlist removal (SHR 0.73, 95% CI: 0.60, 0.89, p =0.002), 16% lower subhazard for waitlist death (SHR 0.84; 95% CI 0.73–0.95, p =0.07), and 13% increased likelihood for transplant (SHR 1.13, 95% CI: 1.07, 1.20, p <0.001). Many patients bypassed the nearest transplant hospital, and longer patient travel distance was associated with favorable waitlist outcomes.
1. Introduction
The resources and expertise required to support lung transplantation are concentrated at approximately 75 hospitals across the United States (U.S.), reflecting the regionalization of the specialty (1). This regionalization of care leads to a wide range of patient travel distances in order to reach a transplant hospital. Typically, patients must travel to a lung transplant hospital to undergo rigorous medical and psychosocial evaluations to determine eligibility for listing, be available at a moment’s notice when a donor organ is identified for the transplant operation, and travel frequently to the transplant hospital for longitudinal outpatient care.
As part of an increasing interest in the role of patient travel distance for healthcare, recent studies in liver and renal transplantation have used patient zip codes collected in national transplant registries to calculate travel distance and measure the impact on outcomes (2–6). These studies are also responsive to the Final Rule set by the U.S. Department of Health and Human Services specifying that geography should have a minimal role in access to organ transplant (7). While recent attention has focused on geography in donor lung allocation (8), less attention has focused on the role of geography in patient access to the transplant hospital. There has not been a recent examination of patient travel patterns and waitlist outcomes in lung transplant, and it is unknown if patients routinely seek listing at the closest lung transplant hospital.
Our goal was to test for an association between patient travel distance and the likelihood of waitlist removal, waitlist death, or lung transplant. Recognizing that travel distance is only part of a complex patient journey, we also assessed if patients bypassed a geographically closer transplant hospital to reach a more distant hospital, and captured social determinants of health through the inclusion of additional registries. Google Maps was leveraged to measure the drive distance from a patient’s permanent zip code to transplant hospital. Recent work in other solid organ transplants has shown an association between longer travel distance and favorable waitlist outcomes as socioeconomically advantaged patients were more able to travel to a chosen hospital (2–4). Our hypothesis was that longer travel distance for lung transplant would be associated with favorable waitlist outcomes.
2. Methods
2.1. Data
We led a retrospective cohort study of adult (≥18 years) lung only waitlisted patients in the U.S. between January 1, 2006 and May 31, 2017 in the Scientific Registry of Transplant Recipients (SRTR). The SRTR captures clinical and demographic data on all solid organ transplants in the U.S., and has the advantage of additional ascertainment of deaths from several federal databases which increases its accuracy (9). The data reported here were supplied by the SRTR as the contractor for the Organ Procurement and Transplantation Network and have a cohort censoring date of May 31, 2018. A patient seeking lung transplant would first appear in the SRTR upon official listing at a transplant hospital. A patient evaluated but declined by a transplant hospital and never listed would not be captured in the SRTR. Patients undergoing re-transplant, missing zip code, or traveling from outside the U.S. for transplant were excluded. This study was declared exempt by the Cleveland Clinic Institutional Review Board (19–1432). Portions of this study were presented at the American Transplant Congress (10), and our approach using Google Maps has been described in a separate publication on post-transplant outcomes (11).
We linked the SRTR with additional registries to include the median annual income for each zip code as captured by the U.S. Census (12), the metro population size from the U.S. Department of Agriculture (13), and the Distressed Communities Index (DCI) from the Economic Innovation Group (14). The DCI combines metrics such as unemployment, education level, poverty rate, median income, and housing vacancies in order to calculate an index score for each zip code from 0 (no distress) to 100 (severe distress).
2.2. Travel distance
Patient travel distance was measured via road driving distance from a patient’s permanent zip code to the transplant hospital using SAS URL access to Google Maps (15). Travel distance was categorized a priori into four categories we estimated would reflect patient experiences: ≤60 miles approximated one hour of travel, >60–180 miles approximated 1–3 hours of travel, >180–360 miles approximated 3–6 hours of travel, and >360 miles approximated a day or more of travel. Patients traveling from Hawaii, Alaska, or Puerto Rico were in the >360 miles category.
2.3. Hospital Bypass
To identify the closest hospital for each patient, we selected the five closest lung transplant hospital zip codes from the patient’s permanent zip code using linear distance and evaluated driving distances to those five hospitals. We defined hospital bypass two ways. First, a bypass was travel to a transplant hospital further than the nearest transplant hospital (e.g. a patient in a large city may have access to several transplant hospitals and may not seek care at the closest one). Second, we also wanted to measure how often patients made a concerted effort to travel to another community for transplant care (e.g. a patient travels from one city to another). Therefore, we also defined another type of bypass as when a patient passed the closest transplant hospital and traveled >60 miles from the closest hospital to the hospital of listing. All transplant hospitals actively reported to the SRTR during the study period.
2.4. Outcomes
We used competing risks regression models to evaluate the association between travel distance and the following outcomes: waitlist death with transplant as a competing risk, transplant with waitlist death as a competing risk, and waitlist removal with both transplant and death as competing risks. Removal from waitlist was counted only for reasons of being too sick for transplant or other reasons. In addition, transplant with both death and waitlist removal (except due to improved health) as competing risks was also evaluated. Competing risks analysis avoids the bias of informative censoring and is often indicated in transplant outcomes research over the Kaplan Meier approach (16).
2.5. Statistics
We evaluated patient characteristics within travel distance groups including counts and percentages for categorical variables, and means and standard deviations or medians and interquartile ranges for continuous variables. ANOVA, Kruskal-Wallis, and Pearson’s chi-square tests were used where appropriate. We estimated the 1-year Kaplan-Meier transplant estimates by hospital and compared the median estimates across distance groups.
Patient characteristics were selected a priori for inclusion in the adjusted models which were from the SRTR risk adjustment model for waitlist mortality (17) and our previous experience. These include sex, age at listing, race, Body Mass Index (BMI), ABO blood group, height, and Lung Allocation Score (LAS). The LAS is a quantitative measure of medical urgency used to prioritize adult waitlisted patients using over a dozen clinical variables (18). We also included United Network for Organ Sharing defined native lung disease categories which were: obstructive disease (primarily emphysema), fibrotic disease (primarily pulmonary fibrosis), cystic disease (primarily cystic fibrosis), or vascular disease (primarily pulmonary hypertension). We adjusted for socioeconomic differences with insurance type (private, Medicare, Medicaid, Veterans Affairs (VA), or other), highest education level (some college or higher vs less), and the DCI, an index score representing the economic state of a zip code.
Recognizing that multi-listed patients have an increased likelihood for transplant (19), we performed a sensitivity analysis of our model looking at the exclusion of multi-listed patients and a second sensitivity analysis with inclusion of transplants performed at the second listing hospital. We included the first listing in our primary analysis.
Data were missing for the following variables: BMI (0.3%), distressed community index (2%), and education (6%). We used multiple imputation (SAS proc MI) with the Markov Chain Monte Carlo method and a single chain to impute 5 datasets with complete data. The multiple imputation included the following variables: sex, age, race, BMI, native lung disease, insurance status, education level, ABO blood group, DCI quintile, patient height, and LAS at the time of transplant. All models were fitted on each of the 5 imputed datasets, and parameter estimates were combined using SAS MIanalyze.
Two-sided P values of 0.05 or less were considered significant. All data were analyzed with Linux SAS, version 9.4 (SAS Institute, Cary, North Carolina). Graphics were produced with R version 3.5.1 (The R Foundation for Statistical Computing, Vienna, Austria).
3. Results
3.1. Patients
There were 22,958 patients who met inclusion criteria, and more than half (54%) were listed at hospitals >60 miles from their household zip code (Figure 1). The median travel distance was 69.7 miles from patient zip code to transplant hospital. Among patients that bypassed the closest hospital by any distance, the median difference in travel distance from their closest hospital to the listing hospital was 5, 28, 113 and 456 miles in each of the distance groups ≤60, >60–180, >180–360, >360 miles, respectively. Next, using our second definition of bypass we found that 41.2% (5,111/12,391) sought listing at a hospital >60 miles away from the closest transplant hospital. Within travel distance groups, patients who bypassed the nearest hospital by >60 miles were 17.3%, 57.3%, and 90.6%, of the >60–180, >180–360, and >360 miles groups, respectively.
Figure 1. STROBE Diagram: Study Cohert Development.
Strengthening the reporting of observational studies in epidemiology (STROBE) diagram: Patients meeting inclusion and exclusion criteria are displayed.
Patients who traveled ≤60 miles went to hospitals with a median 38.5 transplants per year and 1-year transplant estimate of 73%, while those who traveled >360 miles while bypassing a hospital within 60 miles of their home zip code chose hospitals with a median 77 transplants per year and 1-year Kaplan-Meier transplant estimate of 79%.
We observed important clinical differences by travel distance (Table 1). Women (38% in >360 miles vs 45.6% in ≤60 miles) and African Americans (5.8% in >360 miles vs 13.1% in ≤60 miles) were less likely to travel longer distances for listing. Patients in the longest distance category sought care at transplant hospitals with a higher median annual volume (56.0 in >360 miles vs 38.5 in ≤60 miles). Fibrotic lung disease patients were more likely to seek listing at a further transplant hospital (60.4% in >360 miles vs 56.1% in ≤60 miles) and obstructive lung disease patients less likely (23.1% in >360 miles vs 29.3% in ≤60 miles). The mean LAS ranged from 41.9 (>60–180 miles category) to 43.5 (>360 miles category). Patients in longer distance categories were more likely to be multi-listed. Among those concurrently multi-listed at 2 hospitals, 63% traveled further to their second listing hospital. Among those in the ≤60 miles group, 86% traveled further for their second listing, while among those in the >360 miles group, 29% traveled further for their second listing.
Table 1.
Characteristics of study cohort by travel distance
| <=60 (N=10,567) |
>60–180 (N=6,768) |
>180–360 (N=3,472) |
>360 (N=2,151) |
p value | |
|---|---|---|---|---|---|
| Average travel distance in miles | 23.9 | 105.0 | 242.0 | 580.0 | |
| Bypassed nearest lung transplant hospital (by any distance) | 3,161 (29.9) | 3,906 (57.7) | 2,785 (80.2) | 2,057 (95.6) | <0.001c |
| Bypassed nearest lung transplant hospital (by >60 miles) | --- | 1,173 (17.3) | 1,990 (57.3) | 1,948 (90.6) | |
| Stayed within Donation Service Area* | 9,925 (93.9) | 4,258 (62.9) | 813 (23.4) | 127 (5.9) | <0.001c |
| Clinical variables | |||||
| Women | 4,818 (45.6) | 2,843 (42.0) | 1,491 (42.9) | 817 (38.0) | <0.001c |
| Race | <0.001c | ||||
| White | 7,927 (75.0) | 5,890 (87.0) | 3,020 (87.0) | 1,832 (85.2) | |
| African American | 1,386 (13.1) | 431 (6.4) | 234 (6.7) | 125 (5.8) | |
| Other | 1,254 (12.8) | 447 (6.6) | 218 (6.3) | 194 (9.0) | |
| Age in years at listing | 55.6±12.6 | 55.6±12.5 | 54.6±13.4 | 55.5±13.7 | <0.001a |
| Height in centimeters | 168.9±10.0 | 169.6±10.0 | 169.7±10.0 | 170.8±9.8 | <0.001a |
| Native lung disease | <0.001c | ||||
| Group A - obstructive disease | 3,101 (29.3) | 2,185 (32.3) | 1,035 (29.8) | 496 (23.1) | |
| Group B - vascular disease | 453 (4.3) | 233 (3.3) | 113 (3.3) | 82 (3.8) | |
| Group C - cystic disease | 1,083 (10.2) | 723 (10.7) | 424 (12.2) | 273 (12.7) | |
| Group D - fibrotic disease | 5,928 (56.1) | 3,637 (53.7) | 1,900 (54.7) | 1,300 (60.4) | |
| Body Mass Index | 25.7±4.8 | 25.6±4.6 | 25.3±4.6 | 25.2±4.5 | <0.001a |
| Lung Allocation Score at transplant | 42.4±15.5 | 41.9±15.6 | 42.1±16.4 | 43.5±16.6 | <0.001a |
| Listed at multiple centers | 224 (2.1) | 168 (2.5) | 79 (2.3) | 72 (3.3) | 0.007c |
| Among multi-listed only, travel further for second listing | 192 (85.7) | 103 (61.3) | 28 (35.4) | 21 (29.2) | <0.001c |
| Center variables | |||||
| Median annual transplant center volume | 38.5 [23.5, 62.0] | 35.5 [21.5, 56.0] | 51.0 [30.0, 100.0] | 56.0 [33.0, 100.0] | <0.001b |
| Kaplan-Meier transplant estimate for center at 1 year | 0.73 [0.62,0.79] | 0.72 [0.63,0.80] | 0.72 [0.65,0.82] | 0.75 [0.70,0.82] | <0.001b |
| Socioeconomic variables | |||||
| Median zip code annual salary | 63690 [49325,83513] | 51284 [42276.5, 64569.5] | 50712 [41157, 63211] | 54175 [42961, 68893] | <0.001b |
| Metro area population >= 1 million | 9,089 (86.0) | 1,872 (27.7) | 1,057 (30.4) | 712 (33.1) | <0.001c |
| College or higher education level | 5,712 (54.1) | 3,314 (49.0) | 1,830 (52.7) | 1,340 (62.3) | |
| Insurance status | <0.001c | ||||
| Private | 5,984 (56.6) | 3,539 (52.3) | 1,960 (56.5) | 1,162 (54.0) | |
| Medicare | 3,564 (33.7) | 2,612 (38.6) | 1,256 (36.2) | 720 (33.5) | |
| Medicaid | 888 (8.4) | 535 (7.9) | 216 (6.2) | 78 (3.6) | |
| Veterans Affairs | 71 (0.67) | 58 (0.86) | 31 (0.89) | 178 (8.3) | |
| Other | 58 (0.55) | 24 (0.35) | 9 (0.26) | 13 (0.60) | |
| Distressed Community Index Score | 36.0±28.1 | 49.0±27.9 | 48.3±28.8 | 43.2±28.1 | <0.001a |
| Distressed Community Index Quintile | <0.001c | ||||
| 1 - Prosperous | 4,115 (39.4) | 1,330 (20.1) | 716 (21.3) | 564 (27.1) | |
| 2 - Comfortable | 2,257 (21.6) | 1,350 (20.4) | 717 (21.3) | 482 (23.2) | |
| 3 - Mid tier | 1,734 (16.6) | 1,387 (21.0) | 654 (19.4) | 391 (18.8) | |
| 4 - At risk | 1,255 (12.0) | 1,398 (21.1) | 664 (19.7) | 357 (17.2) | |
| 5 - Distressed | 1,084 (10.4) | 1,146 (17.3) | 618 (18.3) | 285 (13.7) | |
| Secular trends | |||||
| 2006 to 2009 | 2,971 (28.1) | 1,977 (29.2) | 1,100 (31.7) | 651 (30.3) | |
| 2010 to 2013 | 3,659 (34.6) | 2,402 (35.5) | 1,322 (38.1) | 802 (37.3) | |
| 2014 to May-2018 | 3,937 (37.3) | 2,389 (35.3) | 1,050 (30.2) | 698 (32.5) | |
Statistics presented as Mean ± SD, Median [P25, P75] or N (column %).
p-values:
ANOVA,
Kruskal-Wallis test,
Pearson’s chi-square test.
There are 58 donation service areas in the United States; each is served by an organ procurement organization.
We evaluated socioeconomic characteristics which could influence a patient’s ability to travel. Median zip code income was highest for patients traveling ≤60 miles ($63,690), followed by those traveling >360 miles ($54,175), and the lowest income was among patients traveling >180–360 miles ($50,712). For the DCI, patients traveling ≤60 miles were from the least distressed communities (36.0), followed by those traveling >360 miles (43.2), and the most distressed communities were among patients traveling >180–360 miles (49.0). As travel distance increased the likelihood of college level education or higher increased (62.3% in >360 miles vs 54.1% in ≤60 miles), and having Medicaid insurance decreased (3.6% in >360 miles vs 8.4% in ≤60 miles).
3.2. Linear vs driving distances
We compared linear and driving distance calculations. In the ≤60 miles linear distance category 10.4% of patients were re-categorized into a longer distance category after using Google Maps to calculate driving distance. In the >60–180 linear miles category 16.8% and in the >180–360 linear miles category 18.5% of patients were re-categorized into a longer distance category after using Google Maps. No patients were moved to a shorter distance category after calculating driving distance.
3.3. Competing risks analysis
Unadjusted cumulative incidence estimates for waitlist death with transplant as a competing risk at 1 year for ≤60, >60–180, >180–360 and >360 miles categories were 11%, 9%, 10% and 9%, respectively (p =0.03). Unadjusted cumulative incidence estimates for transplant with death as a competing risk at 1 year for ≤60, >60–180, >180–360 and >360 miles categories were 67%, 69%, 70% and 72% (p <0.001), and unadjusted cumulative incidence for waitlist removals were 5.2%, 4.7%, 3.9%, and 3.5% (p <0.001).
In the adjusted competing risks model, longer travel distance (>360 miles) was associated with a 16% reduction in the risk for waitlist death (subhazard ratio (SHR), 0.84; 95% CI 0.73–0.95) when compared to the shortest travel distance (≤60 miles), although the overall effect of travel distance group was not significant (p =0.07). Patients traveling further distances were more likely to undergo transplant when considering death and waitlist removal as competing risks (>360 miles SHR 1.13, 95% CI: 1.07, 1.20, p <0.001) and were less likely to experience wait list removal (>360 miles SHR 0.73, 95% CI: 0.60, 0.89, p =0.002) (Table 2). In Figure 2, transplant was analyzed with only death as a competing risk. The inference with both approaches is similar. In our sensitivity analyses: a) excluding multi-listed patients and b) including transplants performed at the second listing hospital, we found similar results (Supporting Tables 1 and 2).
Table 2.
Adjusted* models of waitlist outcomes
| Patient travel distance in miles | Subhazard Ratio for Waitlist Removal with Transplant and Death as Competing Risks (95% CI) |
p value | Subhazard Ratio for Transplant with Death and Waitlist Removal as a Competing Risk (95% CI) |
p value | Subhazard Ratio for Death with Transplant as a Competing Risk (95% CI) |
p value |
|---|---|---|---|---|---|---|
| ≤60 | Reference | 0.002 | Reference | <0.001 | Reference | 0.07 |
| >60–180 | 1.03 (0.92, 1.15) | 1.01 (0.98, 1.05) | 0.95 (0.88, 1.04) | |||
| >180–360 | 0.86 (0.75, 0.998) | 1.04 (0.999, 1.09) | 0.96 (0.87, 1.07) | |||
| >360 | 0.73 (0.60, 0.89) | 1.13 (1.07, 1.20) | 0.84 (0.73, 0.95) |
Models adjusted for age at listing, LAS, sex, BMI, race, insurance (Private, Medicare, Veterans Affairs, Medicaid, or other), education (some college vs. less), ABO blood group, native lung disease, distressed community index, and patient height.
Figure 2. Adjusted models of waitlist outcomes.
Competing risks regression models for the following outcomes: waitlist removal with both transplant and death as competing risks, waitlist death with transplant as a competing risk, and transplant with waitlist death as a competing risk.
4. Discussion
4.1. Waitlist outcomes
When comparing long travel distance (>360 miles) to short travel distance (≤60 miles) in our adjusted competing risks model, we observed a 16% lower subhazard for waitlist death (p =0.07), a 23% lower subhazard for waitlist removal (p =0.002), and 13% higher likelihood of transplant (p <0.001). Patients traveling the longest distance were more likely to be male, White, have a higher LAS, pulmonary fibrosis as the underlying diagnosis, at least a college level education, and were less likely to have public insurance. Among patients who traveled >60 miles, 41.2% sought listing at a hospital which was >60 miles away from the closest transplant hospital. Patients who traveled further for listing sought hospitals with higher annual transplant volumes.
A patient listed for lung transplant is the result of two important sequential decisions. The first is the patient’s decision on which transplant hospital to undergo evaluation. For patients who travel long distance, this decision is also influenced by the cost of leaving their local community and their local network of physicians. Out of pocket costs of travel can include lodging, transportation, and food. Family or friends accompanying patients for the transplant operation or clinic visits likely incur additional costs such as time away from work. The second decision is made by the transplant hospital using a multidisciplinary committee composed of surgeons, physicians, and social workers who weigh a broad range of medical and psychosocial data to assess the potential benefit of transplant for each patient. Each committee uses its own criteria to approve a patient for its waitlist. Thus, the patient journey from referral for transplant, travel to transplant hospital, and evaluation and listing for transplant is a complex and heterogeneous journey of which travel distance is only one component, but nonetheless an important lens through which to understand patient access and experience.
The patient journey has been examined in depth in liver transplant where travel distance is associated with favorable or unfavorable waitlist outcomes depending on the patient population. In the United Kingdom, increased travel distance from home postcode to one of seven national liver transplant hospitals was associated with increased risk for waitlist mortality (20). Among VA patients in the U.S., increased distance to one of five VA liver transplant hospitals was also associated with worse waitlist outcomes (21). However, for the general U.S. population travel outside of a patient’s local community to a transplant hospital was associated with more favorable waitlist outcomes. Patients from socioeconomically advantaged zip codes could travel to waitlist at a chosen transplant hospital, and in doing so reduce the risk for waitlist death 22% (3–4). In our cohort, long distance travel was associated with favorable waitlist outcomes. Among patients traveling >60 miles, those in the >360 miles category had the most favorable DCI score, indicating an association between economically advantaged zip codes and the ability to travel.
4.2. Hospital Bypass
We found that many patients bypassed a closer transplant hospital to be listed at a more distant hospital. The SRTR registry does not capture why patients may have chosen to bypass a closer lung transplant hospital. There has been prior qualitative work in renal transplantation which showed that when patients select a transplant hospital major considerations include the recommendation by trusted clinicians, the perceived hospital reputation, or insurance network limitations (22). Patients may also initially seek evaluation at a local transplant hospital but may have medical comorbidities ultimately deemed too high risk for transplant, prompting an evaluation at a more distant hospital. During this time there may have been the unintended consequence of interval progression of the patient’s native lung disease.
Patients may also select a more distant hospital to increase the odds of transplant and decrease the odds of waitlist death. There has been wide variation in lung transplant rates across the U.S. (23). This is partly a reflection of an uneven geographic distribution of donor lungs and an allocation policy which prioritized donor lungs to nearby local transplant hospitals. As a result, the likelihood of a patient receiving a transplant had been heavily dependent on the geographic location of a transplant hospital (24). However, recent changes in U.S. donor allocation policy aim to reduce the role of geography in access to lung transplant (8). Our work informs the ongoing policy discussion with regards to patient travel to access the waitlist.
4.3. Social Determinants of Health
The link between a patient’s neighborhood environment and transplant outcomes has been previously reported. In renal transplantation, Schold et al. demonstrated a strong independent relationship between a patient’s home county health indicators (such as low birth rates, rate of obesity, and preventable hospital stays) and post renal transplant mortality (25). In our study we leveraged additional databases to characterize the zip code environment. Patients traveling ≤60 miles had the highest median income and favorable DCI scores when compared to patients traveling >60 miles. These were likely patients living in large cities where many lung transplant hospitals are located, and in fact for patients who traveled ≤60 miles, 86% of zip codes were in metro areas with a population >1 million. Patients in the >360 miles travel category were more likely to have a college level education or higher when compared to patients who traveled <360 miles. Presuming education level as a surrogate for health literacy, a link between health literacy and travel for lung transplant care is an area for further investigation and potential intervention. It has previously been observed that low health literacy is associated with a lower hazard for successful referral for renal transplant evaluation (26).
4.4. Regionalized care
Lung transplant remains a specialized surgery which requires intensive peri-operative care. In our cohort, patients who traveled longer distances sought hospitals with higher annual transplant volumes. Kilic et al. showed that high volume lung transplant hospitals correlated with improved short and long term post-transplant survival (27). This hospital volume and outcome relationship has been demonstrated in other complex thoracic surgeries such as pneumonectomies or esophagectomies (28). The benefit of high-volume hospitals is access to experienced surgeons and care from clinical teams who are ‘ready to rescue’ – the notion that these teams are prepared to rapidly respond to common complications of uncommon and complex procedures. Thus, the regionalization of lung transplantation can improve outcomes for patients, especially those with comorbidities known to increase peri-operative risks such as coronary disease or esophageal reflux. As a cautionary note, however, we observed that African Americans composed 13.1% (1,386/10,567) of the ≤60 miles travel category, which decreased to 6.4% (796/12,391) among those traveling >60 miles. Mooney et al. found no difference in waitlist mortality by race/ethnicity, but non White waitlisted patients had lower access to lung transplant (29). The regionalization of care must balance outcomes and access so as not to limit the potential choices racial minorities have in accessing care.
4.5. Limitations
Our findings were limited by the absence of data on the reasons for hospital bypass, highlighting the need to understand patient preferences in seeking lung transplant and more specifically the lung transplant metrics sought by patients. Given the complexity and heterogeneity of patient referral for transplant, additional data is needed on the ‘upstream’ patient journey as current national transplant registries only capture patients at the time of waitlist. Additionally, there is a gap in available data on hospital level heterogeneity in patient selection for the waitlist and how much this played a role in the bypass of a nearby hospital for a further hospital. Though there are established clinical guidelines for pre-lung transplant evaluations (30), ultimately each transplant hospital uses its own criteria to select patients for its waitlist.
We also note several data limitations. First, within a zip code there can be a broad range of socioeconomic levels. Recognizing this limitation, we merged the SRTR data with additional registries to better characterize our cohort. And our approach of using zip code as a marker of geographic location is consistent with other large registry studies in organ transplantation (2–6). Second, registry zip codes were self-reported, which may limit their accuracy. However, the standardized registry case report form specifically requests the patient’s ‘permanent zip code’ in order to avoid reporting a zip code a patient may have temporarily re-located to in the peri-operative period. Also, all registry information is submitted by trained hospital staff and not patients, increasing the likelihood of consistent interpretation of ‘permanent zip code’. It is unlikely patients or transplant hospitals would mis-represent zip code information as doing so provides no advantage. Third, though we calculated the driving distance for each patient, some may have chosen to travel by plane. We were unable to ascertain which patients traveled by plane, however we felt it was important to apply a travel distance calculation to all patients so that there was a uniform assessment of the potential patient travel distance. Fourth, elevated calculated panel reactive antigen (cPRA) levels have been associated with higher waitlist mortality (31–32). However cPRA is not a standard part of the SRTR waitlist file and therefore we were unable to include in our model. Finally, the Google Maps driving route used to estimate patient travel distance may not have been the route used by a patient. Weather, time of day, and traffic conditions all impact a patient’s actual travel distance. While other studies have used linear travel distance, we sought a more granular, patient focused approach by using Google Maps to calculate driving distance. We found that up to 18.5% of patients were moved into a longer travel distance category when using driving as compared to linear distance.
4.6. Conclusion
Longer patient travel distance to a lung transplant hospital was associated with favorable waitlist outcomes with significantly reduced subhazard for waitlist removal and higher likelihood for transplant. Many patients bypass the closest transplant hospital for care. As travel distance increased, so did patient education level. Further investigation is needed into the role of health literacy in helping patients choose a lung transplant hospital. Improved patient access to regionalized care will require investigation beyond zip codes through using neighborhood and community level factors to personalize patient care paths and increase access to transplant.
Supplementary Material
Acknowledgment
The data reported here have been supplied by SRTR as the contractor for the Organ Procurement and Transplantation Network. Data use was in accordance with the SRTR data-use agreement. 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 SRTR.
Dr. Tsuang is funded by a Career Development Award from the National Heart, Lung, and Blood Institute (K23 HL138191-02). The authors have no additional funding sources or relevant disclosures.
Abbreviations:
- BMI
Body Mass Index
- cPRA
Calculated Panel Reactive Antigen
- CI
Confidence Interva
- DCI
Distressed Community Index
- LAS
Lung Allocation Score
- SHR
SubHazard Ratio
- SRTR
Scientific Registry of Transplant Recipients
- VA
Veterans Affairs
Footnotes
Disclosure
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article.
References
- 1.Scientific Registry of Transplant Recipients, https://www.srtr.org/transplant-centers/?organ=lung&recipientType=adult&query= Accessed January 1, 2020
- 2.Axelrod DA, Dzebisashvili N, Schnitzler MA, et al. The interplay of socioeconomic status, distance to center, and interdonor service area travel on kidney transplant access and outcomes. Clin J Am Soc Nephrol. 2010;5(12):2276–2288. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kwong AJ, Mannalithara A, Heimbach J, et al. Migration of Patients for Liver Transplantation and Waitlist Outcomes. Clin Gastroenterol Hepatol. 2019;17(11):2347–2355. [DOI] [PubMed] [Google Scholar]
- 4.Dzebisashvili N, Massie AB, Lentine KL, et al. Following the organ supply: assessing the benefit of inter-DSA travel in liver transplantation. Transplantation. 2013;95(2):361–71. [DOI] [PubMed] [Google Scholar]
- 5.Cicalese L, Shirafkan A, Jennings K, et al. Increased Risk of Death for Patients on the Waitlist for Liver Transplant Residing at Greater Distance From Specialized Liver Transplant Centers in the United States. Transplantation. 2016;100(10):2146–2152. [DOI] [PubMed] [Google Scholar]
- 6.Ross K, Patzer RE, Goldberg DS, Lynch RJ. Sociodemographic Determinants of Waitlist and Posttransplant Survival Among End-Stage Liver Disease Patients. Am J Transplant 2017;17(11):2879–2889. [DOI] [PubMed] [Google Scholar]
- 7.Department of Health and Human Services. Organ procurement and transplantation network, Final Rule. Fed Regist 1998; 63: 16296. [PubMed] [Google Scholar]
- 8.Egan TM. From 6 years to 5 days for organ allocation policy change. J Heart Lung Transplant. 2018;37(5):675–677. [DOI] [PubMed] [Google Scholar]
- 9.Massie AB, Kucirka LM, Segev DL. Big data in organ transplantation: registries and administrative claims. Am J Transplant. 2014;14(8):1723–1730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tsuang W, Arrigain S, Lopez R, et al. ,, Waitlisted for Lung Transplant and the Role of Patient Travel Distance: A National Cohort Study. Am J Transplant. 2019;19(suppl 3), D174 https://onlinelibrary.wiley.com/doi/10.1111/ajt.15406 [Google Scholar]
- 11.Tsuang WM, Arrigain S, Lopez R, et al. Patient travel distance and post lung transplant survival in the United States: A cohort study. Transplantation. 2020. January 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.American Community Survey 2016. Available at external link: https://factfinder.census.gov. Accessed September 27, 2018
- 13. [December 6, 2018];United States Department of Agriculture Rural-Urban Continuum Codes. 2013 Available at External link. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes.aspx. Accessed.
- 14.EIG: Economic Innovation Group Distressed Communities Index. Available at External link http://eig.org/dci. Accessed September 20, 2018
- 15.Zdeb M, Driving distances and Times Using SAS® and Google Maps. Accessed September 20, 2018 http://support.sas.com/resources/papers/proceedings10/050-2010.pdf
- 16.Sapir-Pichhadze R, Pintilie M, Tinckam KJ, et al. Survival Analysis in the Presence of Competing Risks: The Example of Waitlisted Kidney Transplant Candidates. Am J Transplant. 2016;16(7):1958–1966. [DOI] [PubMed] [Google Scholar]
- 17.SRTR Risk Adjustment Model Documentation: Waiting list Models. https://www.srtr.org/reports-tools/waiting-list/ Accessed March 1, 2020
- 18.Egan TM, Murray S, Bustami RT, et al. Development of the new lung allocation system in the United States. Am J Transplant. 2006;6(5 Pt 2):1212‐1227. [DOI] [PubMed] [Google Scholar]
- 19.Mooney JJ, Yang L, Hedlin H, Mohabir P, Dhillon GS. Multiple listing in lung transplant candidates: A cohort study. Am J Transplant. 2019;19(4):1098‐1108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Webb GJ, Hodson J, Chauhan A, et al. Proximity to transplant center and outcome among liver transplant patients. Am J Transplant. 2019;19(1):208–220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Goldberg DS, French B, Forde KA, et al. Association of distance from a transplant center with access to waitlist placement, receipt of liver transplantation, and survival among US veterans. JAMA. 2014;311(12):1234–1243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schaffhausen CR, Bruin MJ, McKinney WT, et al. How patients choose kidney transplant centers: A qualitative study of patient experiences. Clin Transplant. 2019;33(5):e13523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kosztowski M, Zhou S, Bush E, et al. Geographic disparities in lung transplant rates. Am J Transplant. 2019;19(5):1491–1497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Benvenuto LJ, Anderson DR, Kim HP, et al. Geographic disparities in donor lung supply and lung transplant waitlist outcomes: A cohort study. Am J Transplant. 2018;18(6):1471–1480. [DOI] [PubMed] [Google Scholar]
- 25.Schold JD, Buccini LD, Kattan MW, et al. The association of community health indicators with outcomes for kidney transplant recipients in the United States. Arch Surg. 2012;147(6):520–526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Grubbs V, Gregorich SE, Perez-Stable EJ, et al. Health literacy and access to kidney transplantation. Clin J Am Soc Nephrol. 2009;4(1):195–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kilic A, Gleason TG, Kagawa H, et al. Institutional volume affects long-term survival following lung transplantation in the USA. Eur J Cardiothorac Surg. 2019. February 2. [DOI] [PubMed] [Google Scholar]
- 28.Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med. 2002;346(15):1128–1137. [DOI] [PubMed] [Google Scholar]
- 29.Mooney JJ, Hedlin H, Mohabir P, et al. Racial and ethnic disparities in lung transplant listing and waitlist outcomes. J Heart Lung Transplant. 2018;37(3):394–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Weill D, Benden C, Corris PA. A consensus document for the selection of lung transplant candidates: 2014--an update from the Pulmonary Transplantation Council of the International Society for Heart and Lung Transplantation. J Heart Lung Transplant. 2015;34(1):1–15. [DOI] [PubMed] [Google Scholar]
- 31.Barac YD, Mulvihill MS, Jawitz O, et al. Increased cPRA is associated with increased waitlist time and mortality in LTx. Ann Thorac Surg. 2020. April 3 pii: S0003–4975(20)30436–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Aversa M, Benvenuto L, Kim H, et al. Effect of Calculated Panel Reactive Antibody Value on Waitlist Outcomes for Lung Transplant Candidates. Ann Transplant. 2019;24:383–392. [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.