Abstract
Background:
Clinical practice guidelines recommend centralized care for patients with bone sarcoma. However, the relationship between the distance that patients travel to obtain care, institutional treatment volume, and survival is unknown.
Methods:
We used the National Cancer Database to examine associations between travel distance and survival among 8,432 patients with bone sarcoma diagnosed from 2004 to 2015. Associations were identified using multivariable Cox regression analyses that controlled for sociodemographic, clinical, and hospital-level factors; subgroup analyses stratified patients by histological diagnosis, tumor stage, and pediatric or adult status.
Results:
Mortality risk was lower among patients who traveled ≥50 miles (≥80.5 km) than among patients who traveled ≤10 miles (≤16.1 km) (hazard ratio [HR], 0.69 [95% confidence interval (CI), 0.63 to 0.76]). Among hospital-level factors, facility volume independently affected survival: mortality risk was lower among patients at high-volume facilities (≥20 cases per year) than at low-volume facilities (≤5 cases per year), with an HR of 0.72 (95% CI, 0.66 to 0.80). The proportion of patients who received care at high-volume facilities varied by distance traveled (p < 0.001); it was highest among patients who traveled ≥50 miles (53%) and lower among those who traveled 11 to 49 miles (17.7 to 78.9 km) (32%) or ≤10 miles (18%). Patients who traveled ≥50 miles to a high-volume facility had a lower risk of mortality (HR, 0.65 [95% CI, 0.56 to 0.77]) than those who traveled ≤10 miles to a low-volume facility. In subgroup analyses, this association was evident among patients with all 3 major histological subtypes; those with stage-I, II, and IV tumors; and adults.
Conclusions:
This national study showed that greater travel burden was associated with higher survival rates in adults, a finding attributable to patients traveling to receive care at high-volume facilities. Despite the burdens associated with travel, modification of referral pathways to specialized centers may improve survival for patients with bone sarcoma.
Level of Evidence:
Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
Bone sarcomas are rare malignancies, accounting for 0.2% of malignant neoplasms and approximately 5% of childhood malignancies1. Many clinical practice guidelines recommend that patients with bone sarcoma receive centralized care2–4 to improve treatment quality and reduce costs in places where patient referral is not mandatory5,6. However, in addition to the social, economic, psychological, and family barriers that patients must overcome to obtain diagnosis and treatment7, patients who reside far from centralized care institutions face geographic barriers to obtaining specialty care.
Poor geographic access delays diagnosis and impacts the care that patients receive8–11. For the major carcinomas, such as breast, lung, and colon cancers, greater travel distance is generally associated with more advanced stages at diagnosis7,10,12–14 and worse survival outcomes9,15–17. However, most relevant studies were performed using state-level data or small geographic units, and they used a variety of definitions for geographic access, including travel distance, travel time, and rurality9,12–15,18–20. Furthermore, little is known about the relationship between geographic access and survival in patients with bone sarcoma.
In this study, we used the U.S. national database on cancer care to test the hypothesis that greater travel burden and other institutional factors are associated with worse survival among patients with bone sarcoma.
Materials and Methods
Data Source
The data source for this study was the National Cancer Database (NCDB), a nationwide database sponsored by the American College of Surgeons (ACS) and the American Cancer Society that contains information about patterns of cancer care and treatment outcomes. The data represent approximately 70% of cancers diagnosed in the United States and include 25 million cancer cases from >1,500 facilities accredited by the ACS Commission on Cancer21–23. This study was approved by the Memorial Sloan Kettering Cancer Center institutional review board (20–033), and the study concept was approved by the Commission on Cancer.
Study Population
We reviewed NCDB data for patients with bone sarcoma diagnosed from 2004 to 2015. Patients with bone sarcoma arising from the skull, face, mandible, or vertebrae were excluded from the analysis, as were those for whom data on study variables were missing (Fig. 1). Because travel distance was calculated as the distance between the center of the patient’s residential ZIP Code and the reporting facility’s address, we excluded patients who were not diagnosed at the reporting facility or had reports from multiple facilities.
Fig. 1.
Flow diagram of study inclusion and exclusion.
Outcomes and Covariates
The primary outcome of this study was mortality or overall survival. Demographic and clinical variables included in analyses were age, sex, race or ethnicity, modified Charlson-Deyo Comorbidity Index score24, tumor stage25, community-based indicators of education and income (the estimated percentage of adults in the patient’s ZIP Code who did not graduate from high school and the median household income in the ZIP Code, both derived from 2012 American Community Survey data), insurance status, and rurality of residence. Hospital-level variables were facility type (academic or research center, comprehensive community cancer center, community cancer center, or other)26, facility volume, and predefined geographic region. Facility volume was determined by the number of patients registered in the NCDB annually and was categorized into tertiles of approximately equal size: low (≤5 cases per year), medium (6 to 19 cases per year), or high (≥20 cases per year). Finally, patient travel distance was provided in the NCDB; we categorized distance as short (≤10 miles [<16.1 km]), intermediate (11 to 49 miles [17.7 to 78.9 km]), or long (≥50 miles [≥80.5 km]).
Statistical Analysis
We performed descriptive analyses to summarize patient characteristics and used chi-square tests to identify significant differences in covariates according to travel distance. The overall survival was estimated using Kaplan-Meier methods; differences were compared using log-rank tests. Cox proportional hazards regression was used to assess the relationships of patient, clinical, and hospital-level variables with mortality; all available covariates were included in the model. Before estimating this model, we tested the proportional hazards assumption for each individual variable using the time-varying Cox model, failing to reject the null hypothesis of proportional hazards when the p value was >0.05. When the proportional hazards assumption was violated, as indicated by both the p value and visually apparent nonproportionality (i.e., of Kaplan-Meier survival curves), we regrouped the variables so that they satisfied the proportionality assumption. Additional analyses examined survival by the combined factor of travel distance and hospital volume; subgroup analyses explored this association by stratifying patients according to histological diagnosis, tumor stage, and pediatric or adult status. We report 2-sided p values; significance was defined as p < 0.05. Statistical analyses were performed using SPSS software, version 23 (IBM).
Source of Funding
This research was funded in part by NIH/NCI Cancer Center Support Grant P30 CA008748 and by grants-in-aid for overseas research fellowships from the Yasuda Medical Foundation (2018) and the Japan Society for the Promotion of Science (201860336).
Results
Baseline Characteristics
The study cohort consisted of 8,432 patients with bone sarcoma. Patient characteristics are shown in Table I. The median age was 44 years; 43.5% of patients were female and 75.5% of patients were white. Most patients (84.4%) had no comorbid illness. Nearly two-thirds (65.7%) were diagnosed at academic or research institutions. The institutions where patients were diagnosed were most commonly in the South Atlantic region (20.1%), followed by the East North Central region (16.9%) and the Middle Atlantic region (15.7%). Nearly all patients received treatment at the institution where they were diagnosed (92.0%). The histological diagnoses were osteosarcoma (32.8%), Ewing sarcoma or primitive neuroectodermal tumor (11.2%), chondrosarcoma (39.6%), and others (16.5%) (see Appendix, Supplementary Table I).
TABLE I.
Patient Demographic, Clinical, and Facility-Related Characteristics, Stratified by Distance Traveled to Facility
| Characteristic | All (N = 8,432) | Distance Traveled to Facility* | P Value | ||
|---|---|---|---|---|---|
| ≤10 Miles (N = 2,780) | 11 to 49 Miles (N = 3,544) | ≥50 Miles (N = 2,108) | |||
| Percentage of sample | 100% | 33.0% | 42.0% | 25.0% | |
| Age† (yr) | 44 (22, 62) | 47 (24, 67) | 43 (21, 61) | 41 (20, 60) | <0.001 |
| Age | <0.001 | ||||
| <18 years | 17.8% | 16.2% | 18.1% | 19.3% | |
| ≥18 years | 82.2% | 83.8% | 81.9% | 80.7% | |
| Sex | 0.030 | ||||
| Male | 56.5% | 55.8% | 58.1% | 54.6% | |
| Female | 43.5% | 44.2% | 41.9% | 45.4% | |
| Race or ethnicity | <0.001 | ||||
| White | 75.5% | 66.8% | 78.7% | 81.8% | |
| Black | 10.6% | 15.0% | 8.7% | 7.8% | |
| Hispanic | 8.9% | 11.6% | 8.2% | 6.4% | |
| Asian | 2.8% | 4.3% | 2.5% | 1.5% | |
| Other | 2.2% | 2.3% | 1.9% | 2.5% | |
| Modified Charlson-Deyo Comorbidity Index score | <0.001 | ||||
| 0 | 84.4% | 81.8% | 86.0% | 85.3% | |
| 1 | 12.1% | 13.5% | 11.3% | 11.6% | |
| ≥2 | 3.5% | 4.7% | 2.7% | 3.1% | |
| Tumor stage | <0.001 | ||||
| I | 48.0% | 45.3% | 49.1% | 49.7% | |
| II | 31.3% | 29.0% | 31.7% | 33.4% | |
| III | 2.3% | 2.4% | 1.9% | 2.4% | |
| IV | 18.6% | 23.3% | 17.3% | 14.5% | |
| Diagnosis | <0.001 | ||||
| Osteosarcoma | 32.8% | 32.9% | 32.0% | 34.1% | |
| Ewing sarcoma or primitive neuroectodermal tumor | 11.2% | 10.5% | 12.4% | 10.1% | |
| Chondrosarcoma | 39.6% | 38.1% | 39.6% | 41.6% | |
| Others | 16.4% | 18.6% | 16.0% | 14.3% | |
| Facility type | <0.001 | ||||
| Academic or research | 65.7% | 55.2% | 67.3% | 76.9% | |
| Comprehensive community | 19.6% | 25.5% | 18.7% | 13.4% | |
| Community | 2.9% | 5.6% | 2.2% | 0.3% | |
| Other | 11.8% | 13.7% | 11.8% | 9.4% | |
| Facility volume | <0.001 | ||||
| High (≥20 cases per year) | 32.9% | 18.4% | 32.0% | 53.4% | |
| Medium (6 to 19 cases per year) | 35.2% | 31.9% | 38.1% | 34.8% | |
| Low (≤5 cases per year) | 31.9% | 49.7% | 29.9% | 11.8% | |
| Insurance | <0.001 | ||||
| Private | 56.1% | 49.4% | 60.9% | 56.5% | |
| Medicare | 22.7% | 27.5% | 20.7% | 19.9% | |
| Medicaid | 14.2% | 16.0% | 12.0% | 15.6% | |
| Other government | 1.8% | 1.1% | 1.7% | 2.8% | |
| Uninsured | 5.2% | 6.0% | 4.7% | 5.2% | |
| ZIP Code-level income | <0.001 | ||||
| <$38,000 | 17.8% | 21.0% | 10.1% | 26.6% | |
| $38,000 to $47,999 | 23.2% | 21.0% | 19.0% | 33.1% | |
| $48,000 to $63,000 | 25.7% | 24.2% | 26.3% | 26.8% | |
| >$63,000 | 33.3% | 33.8% | 44.5% | 13.6% | |
| ZIP Code-level lack of high-school education | <0.001 | ||||
| <7.0% | 25.4% | 26.9% | 29.5% | 16.6% | |
| 7.0% to 12.9% | 31.2% | 29.4% | 32.3% | 32.1% | |
| 13.0% to 20.9% | 24.7% | 22.5% | 23.4% | 29.6% | |
| ≥21.0% | 18.7% | 21.2% | 14.8% | 21.7% | |
| Rurality of residence | <0.001 | ||||
| Metropolitan | 83.6% | 97.8% | 88.3% | 57.0% | |
| Suburban | 14.6% | 2.2% | 10.7% | 37.7% | |
| Rural | 1.8% | 0.0% | 1.0% | 5.3% | |
| Facility location | <0.001 | ||||
| Pacific | 11.2% | 11.8% | 10.8% | 11.3% | |
| Mountain | 6.0% | 5.3% | 6.0% | 7.2% | |
| West North Central | 9.3% | 6.9% | 6.8% | 16.5% | |
| East North Central | 16.9% | 18.6% | 17.8% | 13.1% | |
| Middle Atlantic | 15.7% | 21.3% | 16.3% | 7.1% | |
| New England | 4.4% | 5.2% | 5.5% | 1.4% | |
| West South Central | 8.9% | 8.6% | 8.4% | 10.2% | |
| East South Central | 7.5% | 4.9% | 5.9% | 13.5% | |
| South Atlantic | 20.1% | 17.4% | 22.5% | 19.7% | |
| Diagnosed and treated at reporting facility | <0.001 | ||||
| Yes | 92.0% | 88.6% | 92.6% | 95.6% | |
| No | 8.0% | 11.4% | 7.4% | 4.4% | |
With regard to distance, 1 mile is equivalent to 1.609 km.
The values are given as the median, with the interquartile range in parentheses.
Predictors of Overall Survival
The 5-year overall survival in the entire cohort was 60% (not shown); it was 53% for patients with osteosarcoma, 55% for those with Ewing sarcoma or primitive neuroectodermal tumor, 72% for those with chondrosarcoma, and 47% for those with other sarcoma types (p < 0.001). In the univariable analysis, travel distance was negatively associated with mortality; patients had a lower risk of death if they had traveled ≥50 miles (hazard ratio [HR], 0.69 [95% confidence interval (CI), 0.63 to 0.76]) or 11 to 49 miles (HR, 0.75 [95% CI, 0.70 to 0.82]) than if they had traveled ≤10 miles. In the multivariable analysis (Table II), mortality was higher among patients who were ≥18 years of age (HR, 1.88); had a comorbidity score of 1 (HR, 1.14) or ≥2 (HR, 1.63); had a tumor stage of II (HR, 2.61), III (HR, 3.27), or IV (HR, 8.43); or had insurance through Medicare (HR, 2.71) or Medicaid (HR, 1.20). The mortality risk was lower if patients were female (HR, 0.87), had Ewing sarcoma (HR, 0.76) or chondrosarcoma (HR, 0.70), or had been diagnosed at a high-volume facility (HR, 0.72 [95% CI, 0.66 to 0.80]) or a medium-volume facility (HR, 0.71 [95% CI, 0.64 to 0.79]). Facility type, ZIP Code-level income and education, rurality, and facility location were not associated with overall survival.
TABLE II.
Multivariable Cox Regression Hazards Model Evaluating the Association Between Covariates and the Risk of Overall Mortality
| Covariate | Adjusted HR* | P Value |
|---|---|---|
| Age | ||
| <18 years | Reference | |
| ≥18 years | 1.88 (1.68 to 2.11) | <0.001 |
| Sex | ||
| Male | Reference | |
| Female | 0.87 (0.81 to 0.93) | <0.001 |
| Race or ethnicity | ||
| White | Reference | |
| Hispanic | 0.89 (0.78 to 1.00) | 0.059 |
| Black | 0.80 (0.69 to 0.93) | 0.004 |
| Asian | 0.83 (0.65 to 1.06) | 0.130 |
| Other | 0.97 (0.76 to 1.25) | 0.832 |
| Modified Charlson-Deyo Comorbidity Index score | ||
| 0 | Reference | |
| 1 | 1.14 (1.03 to 1.27) | 0.013 |
| ≥2 | 1.63 (1.39 to 1.90) | <0.001 |
| Tumor stage | ||
| I | Reference | |
| II | 2.61 (2.36 to 2.89) | <0.001 |
| III | 3.27 (2.63 to 4.07) | <0.001 |
| IV | 8.43 (7.58 to 9.37) | <0.001 |
| Diagnosis | ||
| Osteosarcoma | Reference | |
| Ewing sarcoma or primitive neuroectodermal tumor | 0.76 (0.68 to 0.86) | <0.001 |
| Chondrosarcoma | 0.70 (0.63 to 0.78) | <0.001 |
| Others | 0.89 (0.80 to 0.98) | 0.021 |
| Facility type | ||
| Academic or research | Reference | |
| Not academic or research | 1.00 (0.91 to 1.10) | 0.981 |
| Facility volume | ||
| Low | Reference | |
| Medium | 0.71 (0.64 to 0.79) | <0.001 |
| High | 0.72 (0.66 to 0.80) | <0.001 |
| Insurance status | ||
| Private | Reference | |
| Medicare | 2.71 (2.48 to 2.97) | <0.001 |
| Medicaid | 1.20 (1.07 to 1.35) | 0.002 |
| Other government | 1.06 (0.80 to 1.40) | 0.676 |
| Uninsured | 1.03 (0.86 to 1.23) | 0.768 |
| ZIP Code-level income | ||
| <$38,000 | Reference | |
| $38,000 to $47,999 | 1.00 (0.89 to 1.12) | 0.927 |
| $48,000 to $63,000 | 0.89 (0.79 to 1.01) | 0.079 |
| >$63,000 | 0.89 (0.77 to 1.03) | 0.110 |
| ZIP Code-level lack of high-school education | ||
| <7.0% | Reference | |
| 7.0% to 20.9% | 0.96 (0.87 to 1.07) | 0.447 |
| ≥21.0% | 0.90 (0.77 to 1.04) | 0.146 |
| Rurality | ||
| Rural | Reference | |
| Suburban | 1.05 (0.95 to 1.17) | 0.337 |
| Metropolitan | 0.96 (0.74 to 1.25) | 0.775 |
| Facility location | ||
| Pacific | Reference | |
| Mountain | 1.06 (0.89 to 1.28) | 0.507 |
| West North Central | 1.08 (0.92 to 1.27) | 0.344 |
| East North Central | 0.96 (0.83 to 1.10) | 0.537 |
| Middle Atlantic | 0.87 (0.75 to 1.01) | 0.060 |
| New England | 1.05 (0.86 to 1.29) | 0.615 |
| West South Central | 1.11 (0.94 to 1.31) | 0.215 |
| East South Central | 1.05 (0.88 to 1.24) | 0.593 |
| South Atlantic | 1.02 (0.89 to 1.17) | 0.744 |
The values are given as the adjusted HR, with the 95% CI in parentheses.
Correlates of Travel Distance
Patient, clinical, and hospital-level characteristics stratified by patient travel distance are shown in Table I. Although tumor stage differed by travel distance (p < 0.001), the proportion of patients with metastatic disease at diagnosis was, contrary to our hypothesis, negatively associated with travel distance; 23.3% of patients with a short travel distance had a stage-IV tumor, compared with 17.3% of patients with an intermediate travel distance and 14.5% of patients with a long travel distance. Other characteristics that varied by travel distance included age, sex, race, modified Charlson-Deyo Comorbidity Index score, diagnosis, facility type and volume, insurance, income, education, rurality, and facility location. The relationship between selected characteristics and travel distance is presented in the Appendix, Supplementary Fig. 1. Among favorable prognostic factors, the greatest variation per travel distance was seen for facility volume. Notably, 53.4% of patients who traveled a long distance were diagnosed at a high-volume facility, compared with 32.0% of patients who traveled an intermediate distance and 18.4% of patients who traveled a short distance (Fig. 2). Analyses by histological subtype revealed that, among patients who traveled a short distance, the proportion diagnosed at a low-volume facility was highest for chondrosarcoma (54.8%) and lowest for osteosarcoma (36.9%).
Fig. 2.
The proportion of patients who received sarcoma care at high-volume, medium-volume, and low-volume facilities by travel distance, among those with all histological subtypes of bone sarcoma (Fig. 2-A), osteosarcoma (Fig. 2-B), Ewing sarcoma (Fig. 2-C), and chondrosarcoma (Fig. 2-D).
Combination of Travel Distance and Facility Type
We next compared patients who had traveled short distances to low-volume facilities and patients who had traveled long distances to high-volume facilities. Table III shows patient characteristics stratified by these 2 groups. Compared with the patients in the long-distance and high-volume facility group, patients in the short-distance and low-volume facility group were older (median age, 54 years compared with 40 years), were more likely to have stage-IV disease at diagnosis (28.9% compared with 13.4%) and to live in a metropolitan area (95.7% compared with 60.1%), and were less likely to have private insurance (43.1% compared with 58.4%). In an unadjusted comparison of survival, patients in the short-distance and low-volume facility group had worse overall survival than those in the long-distance and high-volume facility group (e.g., 45% compared with 64% at 5 years; p < 0.001) (Fig. 3); the same was true among patients with osteosarcoma (36% compared with 57%; p < 0.001), Ewing sarcoma (43% compared with 56%; p = 0.002), and chondrosarcoma (59% compared with 76%; p < 0.001). In another analysis, the long-distance and high-volume facility group had significantly better overall survival than the short-distance and low-volume facility group for stage-I, II, and IV tumors (see Appendix, Supplementary Fig. 2). At 5 years, for example, the overall survival was lower in the short-distance and low-volume facility group than in the long-distance and high-volume facility group for stage-I tumors (68% compared with 81%; p < 0.001), stage-II tumors (41% compared with 59%; p < 0.001), and stage-IV tumors (13% compared with 22%; p < 0.001). Notably, the overall survival was worse in the short-distance and low-volume facility group in adult patients, but not in patients younger than 18 years of age (see Appendix, Supplementary Fig. 3). In a regression analysis that adjusted for relevant covariates, patients in the long-distance and high-volume facility group had a lower risk of mortality than those in the short-distance and low-volume facility group (HR, 0.65 [95% CI, 0.56 to 0.77]) (Table IV). Moreover, in aggregate, patients who traveled a short distance to a high-volume facility had a significantly lower risk of mortality than those who traveled a long distance to a low-volume facility (HR, 0.65 [95% CI, 0.51 to 0.83]).
TABLE III.
Characteristics of Patients Who Traveled a Short Distance to a Low-Volume Facility and Patients Who Traveled a Long Distance to a High-Volume Facility
| Characteristic | Short-Distance and Low-Volume Facility Group (N = 1,382) | Long-Distance and High-Volume Facility Group (N = 1,126) | P Value |
|---|---|---|---|
| Age* (yr) | 54 (35, 73) | 40 (20, 59) | <0.001 |
| Age | <0.001 | ||
| <18 years | 10.4% | 20.6% | |
| ≥18 years | 89.6% | 79.4% | |
| Sex | 0.293 | ||
| Male | 57.4% | 55.2% | |
| Female | 42.6% | 44.8% | |
| Race or ethnicity | <0.001 | ||
| White | 75.0% | 83.9% | |
| Black | 9.9% | 7.0% | |
| Hispanic | 9.5% | 6.0% | |
| Asian | 3.5% | 1.2% | |
| Other | 2.1% | 1.9% | |
| Modified Charlson-Deyo Comorbidity Index score | <0.001 | ||
| 0 | 77.8% | 84.6% | |
| 1 | 15.6% | 12.5% | |
| ≥2 | 6.6% | 2.9% | |
| Tumor stage | <0.001 | ||
| I | 45.7% | 48.4% | |
| II | 23.3% | 36.2% | |
| III | 2.1% | 2.0% | |
| IV | 28.9% | 13.4% | |
| Diagnosis | <0.001 | ||
| Osteosarcoma | 24.4% | 35.5% | |
| Ewing sarcoma or primitive neuroectodermal tumor | 11.4% | 11.7% | |
| Chondrosarcoma | 41.9% | 39.5% | |
| Others | 22.3% | 13.3% | |
| Insurance | <0.001 | ||
| Private | 43.1% | 58.4% | |
| Medicare | 38.9% | 17.9% | |
| Medicaid | 12.3% | 16.5% | |
| Other government | 1.1% | 2.9% | |
| Uninsured | 4.6% | 4.3% | |
| ZIP Code-level income | <0.001 | ||
| <$38,000 | 17.3% | 24.9% | |
| $38,000 to $47,999 | 21.9% | 32.2% | |
| $48,000 to $63,000 | 24.3% | 28.4% | |
| >$63,000 | 36.5% | 14.5% | |
| ZIP Code-level lack of high-school education | <0.001 | ||
| <7.0% | 27.5% | 19.1% | |
| 7.0% to 12.9% | 35.0% | 32.4% | |
| 13.0% to 20.9% | 20.8% | 28.9% | |
| ≥21.0% | 16.7% | 19.6% | |
| Rurality | <0.001 | ||
| Metropolitan | 95.7% | 60.1% | |
| Suburban | 4.3% | 35.3% | |
| Rural | 0.0% | 4.6% | |
| Facility location | <0.001 | ||
| Pacific | 13.8% | 14.4% | |
| Mountain | 5.3% | 7.9% | |
| West North Central | 7.1% | 20.8% | |
| East North Central | 19.9% | 11.5% | |
| Middle Atlantic | 15.3% | 6.8% | |
| New England | 7.2% | 1.0% | |
| West South Central | 9.2% | 0.0% | |
| East South Central | 5.0% | 17.5% | |
| South Atlantic | 17.2% | 20.1% |
The values are given as the median, with the interquartile range in parentheses.
Fig. 3.
Kaplan-Meier curves comparing the overall survival between patients who traveled a short distance (≤10 miles [≤16.1 km]) to a low-volume facility (short/low-volume) and those who traveled a long distance (≥50 miles [≥80.5 km]) to a high-volume facility (long/high-volume), among those with all histological types of bone sarcoma (Fig. 3-A), osteosarcoma (Fig. 3-B), Ewing sarcoma (Fig. 3-C), and chondrosarcoma (Fig. 3-D).
TABLE IV.
Multivariable Cox Regression Hazards Model Evaluating the Association Between Covariates and the Risk of Overall Mortality in Patients Who Traveled a Short Distance to a Low-Volume Facility or a Long Distance to a High-Volume Facility
| Covariate | Adjusted HR* | P Value |
|---|---|---|
| Age | ||
| <18 years | Reference | |
| ≥18 years | 1.91 (1.53 to 2.38) | <0.001 |
| Sex | ||
| Male | Reference | |
| Female | 0.93 (0.82 to 1.05) | 0.251 |
| Race | ||
| White | Reference | |
| Hispanic | 0.67 (0.53 to 0.86) | 0.001 |
| Black | 0.60 (0.45 to 0.79) | <0.001 |
| Asian | 0.61 (0.38 to 0.96) | 0.034 |
| Other | 0.67 (0.41 to 1.12) | 0.127 |
| Modified Charlson-Deyo Comorbidity Index score | ||
| 0 | Reference | |
| ≥1 | 1.26 (1.08 to 1.45) | 0.002 |
| Stage | ||
| I | Reference | |
| II | 2.28 (1.91 to 2.72) | <0.001 |
| III | 2.49 (1.68 to 3.69) | <0.001 |
| IV | 6.71 (5.63 to 8.01) | <0.001 |
| Diagnosis | ||
| Osteosarcoma | Reference | |
| Ewing sarcoma or primitive neuroectodermal tumor | 0.91 (0.74 to 1.13) | 0.389 |
| Chondrosarcoma | 0.68 (0.57 to 0.82) | <0.001 |
| Others | 0.79 (0.66 to 0.94) | 0.009 |
| Insurance status | ||
| Private | Reference | |
| Medicare | 2.69 (2.31 to 3.13) | <0.001 |
| Medicaid | 1.25 (1.02 to 1.54) | 0.031 |
| Other government | 1.32 (0.84 to 2.06) | 0.227 |
| Uninsured | 0.92 (0.64 to 1.32) | 0.647 |
| ZIP Code-level income | ||
| >$63,000 | Reference | |
| $48,000 to $63,000 | 0.83 (0.68 to 1.01) | 0.061 |
| $38,000 to $47,999 | 0.84 (0.68 to 1.04) | 0.101 |
| <$38,000 | 0.90 (0.70 to 1.15) | 0.406 |
| ZIP Code-level lack of high-school education | ||
| <7% | Reference | |
| 7.0% to 12.9% | 1.02 (0.85 to 1.23) | 0.795 |
| 13.0% to 20.9% | 1.11 (0.89 to 1.38) | 0.361 |
| ≥21% | 1.06 (0.81 to 1.38) | 0.665 |
| Rurality | ||
| Rural | Reference | |
| Suburban | 1.04 (0.86 to 1.25) | 0.681 |
| Metropolitan | 0.91 (0.55 to 1.50) | 0.714 |
| Facility location | ||
| Pacific | Reference | |
| Mountain | 0.97 (0.72 to 1.32) | 0.861 |
| West North Central | 1.06 (0.84 to 1.36) | 0.615 |
| East North Central | 0.92 (0.73 to 1.15) | 0.450 |
| Middle Atlantic | 0.85 (0.67 to 1.08) | 0.191 |
| New England | 1.21 (0.89 to 1.64) | 0.232 |
| West South Central | 1.08 (0.77 to 1.52) | 0.655 |
| East South Central | 0.89 (0.69 to 1.16) | 0.395 |
| South Atlantic | 0.88 (0.71 to 1.10) | 0.252 |
| Travel distance per facility volume | ||
| Short distance and low volume | Reference | |
| Long distance and high volume | 0.65 (0.56 to 0.77) | <0.001 |
The values are given as the adjusted HR, with the 95% CI in parentheses.
Discussion
Centralized care for bone sarcomas at specialized institutions has been advocated for decades in many countries2,4,6. Although centralized care systems provide improved quality of diagnosis and treatment5,6, it was unknown whether the overall survival differs by geographic and socioeconomic status. Contrary to our hypothesis, greater travel distance was associated with better survival in the United States, a finding attributable to more patients receiving their diagnosis and treatment at high-volume facilities as travel distance increased. This result contradicts findings from a prior review of patients with cancer7 and studies of the major carcinoma types9,15–17. For example, in a survival analysis of patients in Australia with 20 cancer types (excluding sarcomas), geographically remote patients had a relative risk of death from cancer that was 1.35 times that of patients who lived in highly accessible areas16. In a study of 6,848 patients with rectal cancer, the mortality risk increased by 6% for each 100-km increment in distance from the nearest radiation therapy facility15. Moreover, 3 of the authors of the current study recently reported that travel distance and treatment at an academic or research center are associated with reduced mortality among patients with soft-tissue sarcoma27. In the current study, our results in patients with bone sarcoma may be related to age-related differences in the centralized care system. In a separate analysis, we found that hospital volume was more important than the distance that the patients traveled; the mortality risk was lower in patients who traveled a short distance to a high-volume facility than in patients who traveled a long distance to a low-volume facility. These data underscore the importance of receiving care for bone sarcomas at specialized centers, regardless of the travel distance involved. Notably, patients who traveled short distances more frequently received care at low-volume facilities; 95.7% of these patients lived in metropolitan areas. An additional analysis identified significant regional variation in the proportion of patients receiving care at low-volume facilities; more than half of patients in New England (53.0%) had care at low-volume facilities, but less than one-third of those in the Mountain area (26.3%), the West North Central area (28.1%), the Middle Atlantic area (24.9%), the East South Central area (29.6%), and the South Atlantic area (30.1%) did so (p < 0.001) (see Appendix, Supplementary Fig. 4). Providers should recognize these patterns and reconsider their referral pathways for sarcoma care, especially for patients living in metropolitan areas or regions where many patients receive care at low-volume facilities.
Recent reports have shown increased adherence to clinical practice guidelines in the treatment of malignant disease28. In treating bone sarcomas, physicians frequently consult the National Comprehensive Cancer Network guidelines on bone cancer, which strongly recommend that surgical excisions obtain negative margins regardless of histological subtype3. In a supplementary analysis, we found that adherence to the guideline on achieving negative margins was greater at high-volume facilities than at low-volume facilities: a negative margin was achieved in 82.5% of surgically treated patients in the long-distance and high-volume facility group, compared with 70.2% of those in the short-distance and low-volume facility group (p < 0.001) (see Appendix, Supplementary Table II). An analysis of patients with soft-tissue sarcoma in the Netherlands found that high-volume facilities were more likely than low-volume facilities to adhere to clinical practice guidelines and less likely to have patients with macroscopic residual disease28. These data demonstrate that treatment adheres to clinical practice guidelines more effectively when patients receive centralized care at specialized institutions.
The factors associated with the type of facility where patients receive care are numerous and complicated, and likely include insurance type29, primary care and other physicians’ referral patterns, the local reputation of the medical center, social support systems, and patients’ and family members’ prior experiences at the facility. Of these, insurance type is the only one for which information is available in the NCDB. In supplementary analyses, we found that most patients who received care at high-volume facilities (63% [1,736 of 2,771]) were private insurance holders (see Appendix, Supplementary Table III). However, 27.0% of patients with private insurance traveled to low-volume facilities, and many patients who had Medicare (25.0%), Medicaid (33.4%), or other government insurance (39.5%) or were uninsured (21.8%) received care at high-volume facilities. These data suggest that the choice of hospital is influenced by other factors and question whether decisions by referring providers to direct patients to specialized institutions are influenced by the socioeconomic status of each patient.
In this study, the overall survival was worse in adult patients who traveled short distances to low-volume facilities, but not in pediatric patients. The reason may be that approximately 80% of cancer care for U.S. children is provided through children’s hospitals (which provide care up to the age of 18 years)30. Currently, there are 220 registered children’s hospitals, which can be categorized as independent, free-standing children’s hospitals or pediatric units within academic medical centers30. Because the vast majority of high-volume facilities (93.7%) in this study were academic centers (see Appendix, Supplementary Table IV), most pediatric patients likely received care at high-volume facilities. Children’s hospitals are typically in major cities, and patients may travel long distances to receive care at them. There are also free-of-charge referral children’s hospitals, such as St. Jude’s, to which patients may travel across the country to receive cancer care. This may neutralize the effects of travel distance in children. As the role of children’s hospitals was not captured in the NCDB, explanations for the lack of an association in this population are beyond the capacity of the current analysis.
This study had several limitations. First, although patients may benefit from care at specialized sarcoma centers, information on whether treatment was provided at such a center is unavailable in the NCDB, and the code of the facility reporting each case is anonymized. Second, data with regard to decisions regarding travel could not be captured in the NCDB. Once we understand the factors that drive care choices, we may be able to influence them. A follow-up prospective study, with multi-institutional collaborations, would be helpful to understand how to effectively direct patients to high-volume centers. Third, information on whether the full course of multidisciplinary treatment was performed at the reporting facility was unavailable in the NCDB. Although 92% of patients were treated at the facility where they were diagnosed, some may not have completed adjuvant treatment (chemotherapy and/or radiation therapy) due to financial reasons or insurance status. Fourth, detailed information on treatment, such as surgery type, chemotherapy regimen, or radiation therapy dose, was unavailable in the NCDB. Fifth, the overall survival was chosen as the main end point because the NCDB does not include data on disease-specific mortality, although the latter would have been a better objective measure. Finally, our results may not be fully generalizable because the NCDB captures only 70% of incident cancers.
In conclusion, greater travel burden was associated with a higher likelihood of getting high-quality sarcoma care at high-volume facilities and a lower risk of mortality in patients with bone sarcoma. In adult patients, survival outcomes in patients who traveled short distances to low-volume facilities were worse than those in patients who traveled long distances to high-volume facilities, indicating the advantages of pursuing care at specialized institutions despite the possible travel burden. It is crucial that referring providers consider these findings when making recommendations about where to obtain sarcoma care.
Supplementary Material
Footnotes
Appendix
Supporting material provided by the authors is posted with the online version of this article as a data supplement at jbjs.org (http://links.lww.com/XXXXXXX).
Investigation performed at the Orthopaedic Service, Department of Surgery, Memorial Sloan Kettering Cancer Center, New York, NY
Disclosure: The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article (http://links.lww.com/XXXXXXX).
References
- 1.Howlader N, Noone AM, Krapcho M, Garshell J, Miller D, Altekruse SF, Kosary CL, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA. SEER Cancer Statistics Review 1975–2011. National Cancer Institute; 2014. [Google Scholar]
- 2.Gerrand C, Athanasou N, Brennan B, Grimer R, Judson I, Morland B, Peake D, Seddon B, Whelan J; British Sarcoma Group. UK guidelines for the management of bone sarcomas. Clin Sarcoma Res. 2016. May 4;6:7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Biermann JS, Chow W, Reed DR, Lucas D, Adkins DR, Agulnik M, Benjamin RS, Brigman B, Budd GT, Curry WT, Didwania A, Fabbri N, Hornicek FJ, Kuechle JB, Lindskog D, Mayerson J, McGarry SV, Million L, Morris CD, Movva S, O’Donnell RJ, Randall RL, Rose P, Santana VM, Satcher RL, Schwartz H, Siegel HJ, Thornton K, Villalobos V, Bergman MA, Scavone JL. NCCN guidelines insights: bone cancer, version 2.2017. J Natl Compr Canc Netw. 2017. Feb;15(2):155–67. [DOI] [PubMed] [Google Scholar]
- 4.Casali PG, Bielack S, Abecassis N, Aro HT, Bauer S, Biagini R, Bonvalot S, Boukovinas I, Bovee JVMG, Brennan B, Brodowicz T, Broto JM, Brugières L, Buonadonna A, De Álava E, Dei Tos AP, Del Muro XG, Dileo P, Dhooge C, Eriksson M, Fagioli F, Fedenko A, Ferraresi V, Ferrari A, Ferrari S, Frezza AM, Gaspar N, Gasperoni S, Gelderblom H, Gil T, Grignani G, Gronchi A, Haas RL, Hassan B, Hecker-Nolting S, Hohenberger P, Issels R, Joensuu H, Jones RL, Judson I, Jutte P, Kaal S, Kager L, Kasper B, Kopeckova K, Krákorová DA, Ladenstein R, Le Cesne A, Lugowska I, Merimsky O, Montemurro M, Morland B, Pantaleo MA, Piana R, Picci P, Piperno-Neumann S, Pousa AL, Reichardt P, Robinson MH, Rutkowski P, Safwat AA, Schöffski P, Sleijfer S, Stacchiotti S, Strauss SJ, Sundby Hall K, Unk M, Van Coevorden F, van der Graaf WTA, Whelan J, Wardelmann E, Zaikova O, Blay JY; ESMO Guidelines Committee, PaedCan and ERN EURACAN. Bone sarcomas: ESMO-PaedCan-EURACAN Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2018. Oct 1;29(Suppl 4):iv79–iv95. [DOI] [PubMed] [Google Scholar]
- 5.Blay JY, Soibinet P, Penel N, Bompas E, Duffaud F, Stoeckle E, Mir O, Adam J, Chevreau C, Bonvalot S, Rios M, Kerbrat P, Cupissol D, Anract P, Gouin F, Kurtz JE, Lebbe C, Isambert N, Bertucci F, Toumonde M, Thyss A, Piperno-Neumann S, Dubray-Longeras P, Meeus P, Ducimetière F, Giraud A, Coindre JM, Ray-Coquard I, Italiano A, Le Cesne A. Improved survival using specialized multidisciplinary board in sarcoma patients. Ann Oncol. 2017. Nov 1;28(11):2852–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Blay JY. Management of sarcoma patients: centralization in reference centers to fragmentation of systemic treatment. Curr Opin Oncol. 2018. Jul;30(4):240–2. [DOI] [PubMed] [Google Scholar]
- 7.Ambroggi M, Biasini C, Del Giovane C, Fornari F, Cavanna L. Distance as a barrier to cancer diagnosis and treatment: review of the literature. Oncologist. 2015. Dec;20(12):1378–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bosanac EM, Parkinson RC, Hall DS. Geographic access to hospital care: a 30-minute travel time standard. Med Care. 1976. Jul;14(7):616–24. [DOI] [PubMed] [Google Scholar]
- 9.Lee B, Goktepe O, Hay K, Connors JM, Sehn LH, Savage KJ, Shenkier T, Klasa R, Gerrie A, Villa D. Effect of place of residence and treatment on survival outcomes in patients with diffuse large B-cell lymphoma in British Columbia. Oncologist. 2014. Mar;19(3):283–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Massarweh NN, Chiang YJ, Xing Y, Chang GJ, Haynes AB, You YN, Feig BW, Cormier JN. Association between travel distance and metastatic disease at diagnosis among patients with colon cancer. J Clin Oncol. 2014. Mar 20;32(9):942–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Stitzenberg KB, Thomas NE, Dalton K, Brier SE, Ollila DW, Berwick M, Mattingly D, Millikan RC. Distance to diagnosing provider as a measure of access for patients with melanoma. Arch Dermatol. 2007. Aug;143(8):991–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Campbell NC, Elliott AM, Sharp L, Ritchie LD, Cassidy J, Little J. Rural and urban differences in stage at diagnosis of colorectal and lung cancers. Br J Cancer. 2001. Apr 6;84(7):910–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Huang B, Dignan M, Han D, Johnson O. Does distance matter? Distance to mammography facilities and stage at diagnosis of breast cancer in Kentucky. J Rural Health. 2009. Fall;25(4):366–71. [DOI] [PubMed] [Google Scholar]
- 14.Scoggins JF, Fedorenko CR, Donahue SM, Buchwald D, Blough DK, Ramsey SD. Is distance to provider a barrier to care for Medicaid patients with breast, colorectal, or lung cancer? J Rural Health. 2012. Jan;28(1):54–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Baade PD, Dasgupta P, Aitken JF, Turrell G. Distance to the closest radiotherapy facility and survival after a diagnosis of rectal cancer in Queensland. Med J Aust. 2011. Sep 19;195(6):350–4. [DOI] [PubMed] [Google Scholar]
- 16.Jong KE, Smith DP, Yu XQ, O’Connell DL, Goldstein D, Armstrong BK. Remoteness of residence and survival from cancer in New South Wales. Med J Aust. 2004. Jun 21;180(12):618–22. [DOI] [PubMed] [Google Scholar]
- 17.Tracey E, Hacker NF, Young J, Armstrong BK. Effects of access to and treatment in specialist facilities on survival from epithelial ovarian cancer in Australian women: a data linkage study. Int J Gynecol Cancer. 2014. Sep;24(7):1232–40. [DOI] [PubMed] [Google Scholar]
- 18.Celaya MO, Berke EM, Onega TL, Gui J, Riddle BL, Cherala SS, Rees JR. Breast cancer stage at diagnosis and geographic access to mammography screening (New Hampshire, 1998–2004). Rural Remote Health. 2010. Apr-Jun;10(2):1361. [PMC free article] [PubMed] [Google Scholar]
- 19.Parsons MA, Askland KD. Cancer of the colorectum in Maine, 1995–1998: determinants of stage at diagnosis in a rural state. J Rural Health. 2007. Winter;23(1):25–32. [DOI] [PubMed] [Google Scholar]
- 20.Wendt R, Gao Y, Miller BJ. Rural patients are at risk for increased stage at presentation and diminished overall survival in osteosarcoma. Cancer Epidemiol. 2019. Aug;61:119–23. [DOI] [PubMed] [Google Scholar]
- 21.Lin CC, Bruinooge SS, Kirkwood MK, Olsen C, Jemal A, Bajorin D, Giordano SH, Goldstein M, Guadagnolo BA, Kosty M, Hopkins S, Yu JB, Arnone A, Hanley A, Stevens S, Hershman DL. Association between geographic access to cancer care, insurance, and receipt of chemotherapy: geographic distribution of oncologists and travel distance. J Clin Oncol. 2015. Oct 1;33(28):3177–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Mohanty S, Bilimoria KY. Comparing national cancer registries: the National Cancer Data Base (NCDB) and the Surveillance, Epidemiology, and End Results (SEER) program. J Surg Oncol. 2014. Jun;109(7):629–30. [DOI] [PubMed] [Google Scholar]
- 23.Murphy CT, Galloway TJ, Handorf EA, Egleston BL, Wang LS, Mehra R, Flieder DB, Ridge JA. Survival impact of increasing time to treatment initiation for patients with head and neck cancer in the United States. J Clin Oncol. 2016. Jan 10;34(2):169–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992. Jun;45(6):613–9. [DOI] [PubMed] [Google Scholar]
- 25.Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A III, editors. AJCC Cancer Staging Manual. 7th ed. Springer; 2010. [Google Scholar]
- 26.Monson JR, Probst CP, Wexner SD, Remzi FH, Fleshman JW, Garcia-Aguilar J, Chang GJ, Dietz DW; Consortium for Optimizing the Treatment of Rectal Cancer (OSTRiCh). Failure of evidence-based cancer care in the United States: the association between rectal cancer treatment, cancer center volume, and geography. Ann Surg. 2014. Oct;260(4):625–31, discussion 631–2. [DOI] [PubMed] [Google Scholar]
- 27.Fujiwara T, Ogura K, Healey J. Greater travel distance to specialized facilities is associated with higher survival for patients with soft-tissue sarcoma: US nationwide patterns. PLoS One. 2021. Jun 4;16(6):e0252381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hoekstra HJ, Haas RLM, Verhoef C, Suurmeijer AJH, van Rijswijk CSP, Bongers BGH, van der Graaf WT, Ho VKY. Adherence to guidelines for adult (non-GIST) soft tissue sarcoma in the Netherlands: a plea for dedicated sarcoma centers. Ann Surg Oncol. 2017. Oct;24(11):3279–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ho K. The welfare effects of restricted hospital choice in the US medical care market. J Appl Econ. 2006;21(7):1039–79. [Google Scholar]
- 30.Children’s Hospital Association. Membership. Accessed 2022 Feb 11. https://www.childrenshospitals.org/about-cha/about/membership
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.