Abstract
Purpose
Previous studies have demonstrated improved outcomes at high-volume colorectal surgery centers, however the benefit for patients who live far from such centers has not been assessed relative to local, low-volume facilities.
Methods
The 2010–2015 National Cancer Database (NCDB) was queried for patients with stage I-III colon adenocarcinoma undergoing treatment at a single center. A ‘local, low-volume’ cohort was constructed of 12,768 patients in the bottom quartile of travel distance at the bottom quartile of institution surgical volume and a ‘travel, high-volume’ cohort of 11,349 patients in the top quartile of travel distance at the top quartile of institution surgical volume.
Results
In unadjusted analysis, patients in the travel cohort had improved rates of positive resection margins (3.7% vs. 5.5%, p<0.001), adequate lymph node harvests (92% vs. 83.6%, p<0.001), and 30- (2.2% vs. 3.9%, p<0.001) and 90-day mortality (3.7% vs. 6.4%, p<0.001). On multivariable logistic regression analysis adjusting for patient demographic, tumor, and facility characteristics, the cohorts demonstrated equivalent overall survival (HR: 0.972, p=0.39), with improved secondary outcomes in the ‘travel’ cohort of adequate lymph node harvesting (OR: 0.57, p<0.001), and 30- (OR 0.79, p=0.019) and 90-day mortality (OR 0.80, p=0.004).
Conclusion
For patients with stage I-III colon cancer, traveling to high-volume institutions compared to local, low- volume centers does not convey an overall survival benefit. However, given advantages including 30- nd 90-day mortality and adequate lymph node harvest, nuanced patient recommendations should consider both these differences and the unquantified benefits to local care, including cost, travel time, and support systems.
Keywords: colorectal, surgery, high-volume, local, low-volume, location
Introduction:
The provision of cancer care in the United States has been evaluated through the use of large datasets, contributing to our understanding of oncologic outcomes and quality of care on a national scale (1–5). These datasets have been leveraged to assess regional and facility differences in the care of complex oncologic diseases. Such studies have examined both the distances patients travel to receive their care and the outcomes involving care received at high-volume, and/or academic facilities (2–4, 6, 7). However, differences in outcomes between patients who travel to receive care at high-volume centers and those who receive care at local, low-volume facilities have been incompletely explored for diseases with perceivably more straightforward management, such as colon cancer (1, 8).
Colon cancer is a prevalent and morbid disease process (9). Localized disease is managed with surgical resection, and the pathologic stage dictates the need for adjuvant chemotherapy. Recent improvements in chemotherapy and minimally invasive approaches to resection have led to improved survival with decreased morbidity (10). While these two aspects of treatment are widespread, variability in adherence is considerable. In this study we aim to compare the overall survival and oncologic outcomes of patients who seek care at low-volume, local hospitals with patients who travel long distances to high-volume centers. We hypothesize that there will be no difference in overall survival but there will be improved short-term outcomes for patients with stages I-III colon cancer who travel to high-volume centers compared to those who receive care locally at low-volume centers.
Methods:
Data Source:
This is a retrospective study of patients with adenocarcinoma of the colon using the National Cancer Database (NCDB) 2010–2015. This review has been approved by the Duke University Institutional Review Board. The NCDB is the product of collaboration between the American College of Surgeons and the Commission on Cancer, and it provides aggregate oncologic outcomes data on greater than 70% of newly diagnosed cancers from more than 1500 cancer centers.
Patient Selection:
Patients were included in the study cohort if they were treated for a single primary, pathologically staged I-III colon adenocarcinoma. Patients with multiple primaries, those with appendiceal tumors, and those with resections that included the rectum, were excluded from the analysis. In the NCDB ‘distance traveled’ indicates the miles from the patient’s home zip code to the treating center’s zip code. The variable used is CROW FLY, which is straight line measurement of the center of the zip codes and is a surrogate for travel distance. Hospital case volume is a recorded data field for each participating center.
Statistical Analysis:
Patients included in this study were stratified by distance traveled to their surgical center and the case volume of the treating facility. A “local” cohort was comprised of patients who traveled the shortest distance (<25th percentile in miles traveled) to a low-volume center (<25th percentile in case volume). A “travel” cohort was comprised of patients who traveled the furthest distance (>75th percentile in miles traveled) to a high-volume center (>75th percentile in case volume) (Figure 1). The remainder of the patients were excluded from this analysis as their data did not contribute to the central aim of the study. Patient, tumor, operative, and facility characteristics were compared using Wilcoxon rank-sum test for continuous variables and Pearson’s Chi squared for discrete variables.
Figure 1:
Representation of cohort generation. Travel: high-volume long distance compared to local: low- volume, short travel distance.
To adjust for confounding from selection and indication bias, multivariable logistic regression was performed to understand the adjusted risk of oncologic outcomes. Cox-proportional hazard modeling was performed to measure adjusted mortality for patients who traveled to high-volume centers compared to the those who received treatment at local, low-volume centers. P-values of <0.05 were deemed significant. Statistical analysis was performed in R version 3.4.0 (Vienna, Austria).
Results:
Of the combined local/travel cohort of 24,117 patients treated for pathologic stage I-III colon cancer, 12,768 were included in the ‘local’ and 11,349 in the ‘travel’ cohorts (Table 1). Those who received their care locally were older (local median age: 70, Interquartile Range (IQR): 60–80; travel median age: 67, IQR: 58–78, p<0.001), female (local 53.3% vs travel 49.4%, p<0.001), and non-white (local black race: 15.3%, local ‘other’ race: 4.4% vs travel black race: 8.8%, travel ‘other’ race: 2.8%, p<0.001). Patients who were highly educated were also more likely to receive care locally (local above median: 55% vs. travel above median 48.9%, p<0.001), as well as those with incomes above the median (local: 66.6% vs travel: 58.9%, p<0.001). Patients with private insurance coverage were the only subset more likely to travel to a high-volume center (local: 27.9% vs travel: 39.4%, p <0.001) (Table 1).
Table 1:
Patient demographic, facility and tumor characteristics for the cohort
| N | Local (N=12768) |
Travel (N=11349) |
P-value | |
|---|---|---|---|---|
| Age (years) Median/IQR | 24117 | 60/70/80 | 57/67/76 | <0.001 |
| Gender | 24117 | <0.001 | ||
| Male | 46.7% (5957) | 50.6% (5742) | ||
| Female | 53.3% (6811) | 49.4% (5607) | ||
| Year of Diagnosis | 24117 | <0.001 | ||
| 2010 | 16.2% (2070) | 14.6% (1652) | ||
| 2011 | 16.9% (2163) | 14.9% (1693) | ||
| 2012 | 16.6% (2116) | 16.7% (1892) | ||
| 2013 | 16.7% (2129) | 17.9% (2027) | ||
| 2014 | 17.2% (2193) | 17.8% (2025) | ||
| 2015 | 16.4% (2097) | 18.2% (2060) | ||
| Race | 23942 | <0.001 | ||
| White | 80.3% (10204) | 88.5% (9934) | ||
| Black | 15.3% (1946) | 8.8% (984) | ||
| Other | 4.4% (565) | 2.8% (309) | ||
| Charlson-Deyo Score | 24117 | 0.156 | ||
| 0 | 67.7% (8643) | 69.0% (7829) | ||
| 1 | 22.8% (2917) | 22.1% (2509) | ||
| 2 | 6.5% (834) | 6.1% (687) | ||
| 3 | 2.9% (374) | 2.9% (324) | ||
| Education Above Median | 23348 | <0.001 | ||
| Above Median | 56.2% (6943) | 48.8% (5364) | ||
| Below Median | 43.8% (5406) | 51.2% (5635) | ||
| Income Level | 23353 | <0.001 | ||
| Below Median | 33.4% (4123) | 40.2% (4418) | ||
| Above Median | 66.6% (8231) | 59.8% (6581) | ||
| Insurance Status | 24117 | <0.001 | ||
| Medicaid | 6.9% (875) | 3.3% (370) | ||
| Medicare | 59.8% (7634) | 52.8% (5995) | ||
| No Insurance | 3.4% (431) | 2.5% (283) | ||
| Other Government | 0.8% (96) | 1.0% (116) | ||
| Private Insurance | 27.9% (3560) | 39.4% (4473) | ||
| Unknown | 1.3% (172) | 1.0% (112) | ||
| Distance (miles) Median/IQR | 24117 | 1.1/1.8/2.7 | 23.6/34.5/58.7 | <0.001 |
| Facility Type | 21567 | <0.001 | ||
| Community | 55.5% (6284) | 0.0% (0) | ||
| Comprehensive | 34.7% (3931) | 45.4% (4655) | ||
| Academic | 9.8% (1109) | 54.6% (5588) | ||
| Facility Location | 23557 | <0.001 | ||
| New England | 8.6% (1080) | 2.9% (322) | ||
| Middle Atlantic | 13.7% (1722) | 13.2% (1451) | ||
| South Atlantic | 19.0% (2384) | 22.3% (2455) | ||
| East North Central | 25.2% (3158) | 18.0% (1979) | ||
| East South Central | 4.3% (534) | 12.8% (1409) | ||
| West North Central | 6.0% (754) | 12.5% (1380) | ||
| West South Central | 8.1% (1019) | 13.3% (1465) | ||
| Mountain | 4.1% (510) | 1.6% (181) | ||
| Pacific | 11.0% (1378) | 3.4% (376) | ||
| Pathologic Stage of Tumor | 24117 | <0.001 | ||
| 1 | 24.5% (3127) | 27.2% (3087) | ||
| 2 | 37.2% (4756) | 35.0% (3975) | ||
| 3 | 38.3% (4885) | 37.8% (4287) | ||
| Extent of Surgery | 24117 | <0.001 | ||
| Partial Colectomy/Segmental | 35.5% (4529) | 35.3% (4011) | ||
| Subtotal Colectomy | 62.8% (8018) | 61.6% (6989) | ||
| Total Colectomy | 1.7% (221) | 3.1% (349) | ||
| Contiguous Organ Resection | 24117 | 0.648 | ||
| With Contiguous Organ | 6.1% (775) | 5.9% (673) | ||
| Without Contiguous Organ | 93.9% (11993) | 94.1% (10676) | ||
| Surgical Approach | 24117 | <0.001 | ||
| Another facility | 2.8% (356) | 6.7% (757) | ||
| Converted to Open | 5.9% (758) | 6.0% (686) | ||
| Laparoscopic | 31.4% (4010) | 44.3% (5026) | ||
| Open | 57.1% (7292) | 38.2% (4339) | ||
| Robotic | 2.8% (352) | 4.8% (541) | ||
| Days from diagnosis to Surgery Median/IQR | 23759 | 0/7/23 | 1/14/30 | <0.001 |
| Any Adjuvant Therapy | 23719 | 0.004 | ||
| No | 68.7% (8605) | 66.9% (7491) | ||
| Yes | 31.3% (3923) | 33.1% (3700) | ||
| Adjuvant Chemotherapy (specific) | 23803 | <0.001 | ||
| No | 69.1% (8696) | 67.1% (7519) | ||
| Yes | 30.9% (3896) | 32.9% (3692) | ||
| Days from Surgery to Chemotherapy Median/IQR | 7470 | 34/45/60 | 34/44/57 | 0.081 |
| Adequate Lymph Node Harvest (12+) | 24036 | <0.001 | ||
| Adequate Harvest | 86.3% (10983) | 92.0% (10403) | ||
| Inadequate Harvest | 13.7% (1743) | 8.0% (907) | ||
| Surgical Margins | 23987 | <0.001 | ||
| Negative | 94.5% (11993) | 96.3% (10881) | ||
| Positive | 5.5% (697) | 3.7% (416) | ||
| Hospital Length of Stay (days) Median/IQR | 21589 | 4/6/8 | 4/5/7 | <0.001 |
| 30-Day Unplanned Readmission | 23908 | <0.001 | ||
| No | 94.7% (12025) | 95.7% (10727) | ||
| Yes | 5.3% (674) | 4.3% (482) | ||
| 30-Day Mortality | 19676 | <0.001 | ||
| No | 96.1% (10155) | 97.8% (8910) | ||
| Yes | 3.9% (412) | 2.2% (199) | ||
| 90-Day Mortality | 19539 | <0.001 | ||
| No | 93.6% (9834) | 96.3% (8694) | ||
| Yes | 6.4% (673) | 3.7% (338) |
Overall case volumes and distance traveled are variable by region (Table 2). Patients meeting our ‘local’ criteria traveled <3.7 miles, with a median of 1.8 miles (IQR: 1.1–2.7), while those in the ‘travel’ cohort traveled >17.5 miles, with a median of 34.5 miles (IQR: 23.6–58.7). The ‘low-volume’ centers of the bottom quartile performed between 0–33 colon resections per year, and the ‘high-volume’ centers of the highest quartile performed between 76–289 per year. Patients in New England, East North Central, Mountain, and Pacific more likely to receive care locally (p<0.001) (Table 1).
Table 2:
Median case volume per year, and median distance traveled by Region
| Median Case Volume Per year | Median Distance Traveled (miles) | Region |
|---|---|---|
| 37 | 5.7 | Northeast |
| 62 | 6.2 | Mid-Atlantic |
| 55 | 8.4 | Southeast |
| 46 | 7.1 | Great Lakes |
| 54 | 13.3 | South |
| 56 | 10.6 | Midwest |
| 49 | 10.5 | West |
| 43 | 8.7 | Mountain |
| 45 | 6.2 | Pacific |
Patient comorbidity status was not associated with likelihood of receiving local care. The healthiest patients with Charlson-Deyo Score (CDCC) of 0 (local: 67.7% vs. travel: 69.0%), and the sickest patients, CDCC of 3 (local: 2.9% vs travel: 2.9%), were equally distributed between the two cohorts (p=0.156).
Regarding operative technique, patients who required a total colectomy were more likely to travel to a high-volume center (local: 1.7% vs. travel: 3.1%, p<0.001). Resection of contiguous organs were equally likely to be performed between groups (p=0.648). Open surgical techniques were more likely to be used locally (local: 57.1% vs travel: 38.2%). Laparoscopic (local: 31.4% vs. travel: 44.3%) and robotic (local: 2.8% vs. travel: 4.8%) approaches were more likely to be used in the travel cohort (p<0.001) (Table 1).
Regarding oncologic outcomes, those in the travel cohort had a longer interval between diagnosis and resection (local median: 7, IQR: 0–23, vs. travel median: 14, IQR: 1–30, p<0.001). However, the interval between resection and adjuvant chemotherapy was equivalent between the two groups (local median: 45, IQR: 34–60 vs. travel median: 44, IQR: 34–57, p=0.081). The travel cohort was more likely to have negative surgical margins (local: 94.5% vs. travel: 96.3%, p<0.001) and adequate lymph node harvest as defined by 12 or more nodes examined (local: 83.6% vs. travel: 92.0%, p<0.001). Hospital length of stay was shorter for those who traveled to high-volume centers (local median: 6, IQR 4–8 vs. travel median: 5, IQR 4–7, p<0.001) and readmissions were lower (local: 5.3% vs. travel: 4.3%, p<0.001).
Unadjusted 30-day (local: 3.9% vs. travel: 2.2%, p<0.001), and 90-day (local: 6.4% vs. travel: 3.7%, p<0.001) mortality was lower in the cohort who traveled to high-volume centers (Table 1).
Adjusted analysis:
Survival analysis was performed using Cox-proportional hazard modeling while adjusting for appropriate clinical and facility characteristics. Overall survival was equivalent between groups (HR: 0.972, 95% CI: 0.91–1.04, p=0.39) (Table 3).
Table 3:
Adjusted Survival: Low-volume, low- distance as reference.
| Variables | Hazard Ratio | Lower 95% CI | Upper 95% CI | p-value |
|---|---|---|---|---|
| Adjusted Overall Survival | 0.972 | 0.91 | 1.037 | 0.39 |
| Age | 1.041 | 1.038 | 1.045 | <0.001 |
| Sex (Male as reference) | 0.829 | 0.779 | 0.882 | <0.001 |
| 2010 | 0.877 | 0.784 | 0.981 | 0.022 |
| 2011 | 0.927 | 0.829 | 1.036 | 0.18 |
| 2012 | 0.98 | 0.877 | 1.095 | 0.722 |
| 2013 | 0.966 | 0.862 | 1.084 | 0.559 |
| Race-Black (White as reference) | 1.043 | 0.942 | 1.155 | 0.415 |
| Race-Other (White as reference) | 0.657 | 0.528 | 0.818 | <0.001 |
| Insurance-Medicare (Medicaid as reference) | 0.742 | 0.63 | 0.873 | <0.001 |
| Insurance-No Insurance (Medicaid as reference) | 0.904 | 0.708 | 1.154 | 0.417 |
| Insurance-Other Government (Medicaid as reference) | 0.788 | 0.534 | 1.164 | 0.231 |
| Insurance-Private Insurance (Medicaid as reference) | 0.577 | 0.488 | 0.682 | <0.001 |
| Insurance-Unknown (Medicaid as reference) | 0.843 | 0.612 | 1.16 | 0.294 |
| Path Stage I (Stage III as reference) | 0.269 | 0.243 | 0.298 | <0.001 |
| Path Stage II (Stage III as reference) | 0.4 | 0.37 | 0.432 | <0.001 |
| Extent of Surgery-Subtotal Colectomy (Segmental as reference) | 1.021 | 0.955 | 1.091 | 0.551 |
| Extent of Surgery-Total Colectomy (Segmental as reference) | 1.558 | 1.277 | 1.902 | <0.001 |
| Approach-Converted to Open | 0.99 | 0.753 | 1.3 | 0.941 |
| Approach-Laparoscopic | 0.686 | 0.533 | 0.883 | 0.003 |
| Approach-Open | 0.962 | 0.75 | 1.234 | 0.759 |
| Approach-Robotic | 0.694 | 0.495 | 0.973 | 0.034 |
| Receipt of Adjuvant Chemotherapy - Omitted as Reference | 0.278 | 0.169 | 0.457 | <0.001 |
| Inadequate Node Harvest - Adequate as reference | 1.263 | 1.153 | 1.384 | <0.001 |
| Positive Margins- Negative Margins as Reference | 2.408 | 2.17 | 2.672 | <0.001 |
| Length of Stay (increasing) | 1.024 | 1.022 | 1.026 | <0.001 |
| Readmission- No Readmission as Reference | 1.476 | 1.311 | 1.661 | <0.001 |
Multivariable logistic regression was done to model the association between travel to high-volume centers and oncologically relevant outcomes. Treatment at a local, low-volume center was used as the comparative reference. Travel to a high-volume center was associated with greater likelihood of adequate nodal harvest (OR: 0.57, 95% CI: 0.50–0.64, p<0.001), 30-day mortality (OR: 0.79, 95% CI:0.65–0.93, p=0.019), and 90-day mortality (OR: 0.80, 95% CI: 0.68–0.93, p=0.004). No difference was seen in rates of readmission, (OR: 0.89, 95% CI: 0.78–1.01, p=0.067), positive margins (OR: 94, 95% CI: 0.78–1.14, p=0.54), or adjuvant chemotherapy (OR: 0.94, 95% CI: 0.84–1.06, p=0.321) between the two groups (Table 4).
Table 4:
Adjusted Oncologic Outcomes: Low-volume, low- distance as reference.
| Adjusted Oncologic Outcomes | Odds Ratio | Lower 95% CI | Upper 95% CI | p-value |
|---|---|---|---|---|
| Adequate Nodes | 0.565 | 0.498 | 0.642 | <0.001 |
| Positive Surgical Margins | 0.943 | 0.781 | 1.139 | 0.543 |
| Readmission | 0.886 | 0.778 | 1.009 | 0.067 |
| Receipt of Adjuvant Chemotherapy | 0.941 | 0.835 | 1.061 | 0.321 |
| Days to Chemo | 0.941 | 0.835 | 1.061 | 0.321 |
| 30-day Mortality | 0.794 | 0.654 | 0.963 | 0.019 |
| 90-day Mortality | 0.797 | 0.683 | 0.929 | 0.004 |
Discussion:
In this study we assess the oncologic impact of traveling to high-volume centers for resection of stage I-III colon adenocarcinoma. Rates of appropriate nodal harvest and both 30- and 90-day mortality were superior for patients who traveled to high-volume centers (11). Overall survival and other oncologic outcomes were equivalent. This is counter to published studies examining esophageal, pancreatic, and rectal resections where travel to high-volume centers was universally associated with improved overall survival. To our knowledge, the volume-distance relationship has not been previously explored in cancers where the operative technique and adjuvant therapy regimens are considered straightforward, such as colon cancer (1).
For several cancers requiring complex resection, including rectal, pancreatic, and esophageal, it has been established that the benefits of traveling to a high-volume center are significantly greater than the potential benefits of receiving care locally at a low-volume center. Using methodology initially described at Duke, Xu et al. found superior short-term outcomes and overall survival (HR: 0.66, p<0.001) for patients with rectal adenocarcinoma who chose to travel to high-volume centers (2). Lidsky et al. described both an oncologic and survival advantage for patients who travelled to high-volume centers for management of pancreatic cancer (HR: 0.75, p=0.002) (4). Speicher et al. demonstrated an increased likelihood of undergoing esophageal resection for esophageal cancer when traveling to a high-volume center, and improved 5-year survival compared to those who received their care locally (3). These findings provide evidence for the value of centralization of complex oncologic care, with each noting that caution must be exercised to not increase disparities for those with barriers to travel. Our study requires more complex interpretation, weighing which outcomes are most important to measure in cancer care. While equivalent overall survival is a compelling argument for receiving care locally, the value of improved short-term mortality and lymph node harvest at high-volume centers cannot be minimized and should be explored in other datasets.
Notably, multiple studies have established the correlation between improved overall and disease-free survival and both adequacy of lymph node harvest and negative surgical margin status. A systematic review of studies totaling 61,371 patients showed that the number of lymph nodes evaluated after resection was correlated with improved survival of patients with stage II and III colon cancer (12). In a prospective study of 1,100 patients undergoing surgery for colon cancer, Khan et al. demonstrated higher rates of 3- and 5-year disease-free survival in patients with R0 resections compared to those with R1 status (13). In a study of 1,343 patients with primary colon cancer, Lee et al. found that adequacy of lymph node harvest was an independent risk factor for disease-free survival (14). Similarly, in a study of high-risk patients with stage II and III colon cancer, Le Voyer et al. demonstrated an improvement in disease-specific survival with an increase in number of resected lymph nodes (15). The NCDB does not document disease free survival it can be extrapolated that improved quality of resection with regards to nodal harvest and negative margins can be associated with extended time to disease recurrence ultimately impacting patient quality of life, and delaying the expense of additional therapy.
Bos et al. examined the relationship between hospital case volume and outcomes for colorectal cancer in a population-based study in the Netherlands (8). A cohort of patients was generated from their population-based cancer registry from 2005–2011. Their aim was to assess if hospital volume determined surgical care characteristics and 30-day mortality. Surgical care characteristics included laparoscopic vs. open vs. converted approach and anastomotic leak requiring re-intervention or readmission within two-months. Hospital volume for colon resection were stratified into <50, 50–74, 75–99, and >100. While minimum case volume is established for rectal cancer in the Netherlands, there is no established threshold in colon cancer (16). Regarding surgical approach, high-volume centers were more likely to use a laparoscopic approach and converted to open procedures with lower frequency. No difference in anastomotic leak was seen in this population. As seen in our study, adjusted 30-day mortality was higher in low-volume centers. However, both unadjusted 1-, 3- and 5-year survival, as well as the adjusted survival (HR) were similar among hospital volumes. Corroborating results between the NCDB and the Dutch cancer registry improve the acceptability of the findings of comparable overall survival between our local and travel cohorts and augment the need to further investigate the variability leading to improved short-term mortality in low-volume centers.
A distinguishing aspect of our study compared to the studies on rectal, pancreatic, and esophageal cancers is the adjustment for use of minimally invasive surgery (2–4). Minimally invasive surgery (MIS) has become the standard of care for colorectal cancer and has become widely incorporated across hospital types(17–19). Benefits include shorter length of stay, decreased postoperative complications, and improved short-term morbidity and mortality. High-volume and academic institutions have more rapidly adopted these practices and consistently show increased use of MIS as well as decreased conversion to open rates compared to community centers. Adjusting for use of MIS may explain equivalent outcomes between the local and travel cohorts.
Limitations:
This study was a retrospective analysis of data from the NCDB 2010–2015, and is subject to the same limitations as other large database retrospective analyses. A lack of granularity in the data such as specific comorbidities, including the American Society of Anesthesiologist (ASA) class or a frailty score, surgeon-specific case volume, re-admission to non-index hospitals, and categories of post-operative complications limit the multivariate analysis to include only more broad surrogates of these data. Decision making surrounding facility selection is the summation of multiple factors including referrals, affiliate institutions, insurance, proximity to closest facility, and the decision to bypass a nearby facility. While these variables are not included in the NCDB, the distribution of impact from these confounders are assumed to be equal across our comparative groups such that their contribution is negligible. The lack of clinical variables such as measures of quality of life and physical performance also limit our analysis of factors influencing travel. Overall survival is a common endpoint in retrospective trials, however, many oncologic studies use disease-free survival as the endpoint of choice. Our analysis is limited to overall survival, as the NCDB does not provide adequate reports of disease recurrence to calculate disease-free survival. Human error in data reporting, both at the level of the providers at each of the participating centers as well as the data entry specialists who are the final handlers of data, is another limitation. “Distance,” as determined by the NCDB, is a linear measurement between the patient’s zip code and the hospital zip code, and may not reflect travel time with accuracy or the impact of rural versus urban contexts, although it has been deemed a suitable surrogate for comparison (20–23). The strengths of this study also stem from the scope of data from the NCDB, a nationwide database, which includes data from over 1500 accredited institutions. The homogeneity of the data collected as well as the large sample size make our results generalizable to the U.S. population suffering from stage I-III colon cancer and allowed for more robust thresholds for statistical significance.
Conclusions:
This retrospective study exploring the relationship between overall survival and oncologic outcomes for stage I-III colon cancer found equivalent overall survival for patients receiving care locally at low-volume centers and those who travel to high-volume centers. However, several short-term outcomes were superior for those who traveled to high-volume centers, including adequate lymph node harvest and both 30- and 90-day mortality. The study results lend to nuanced recommendations. We acknowledge the potential benefits of receiving care locally in travel time and cost savings for families given there is no sacrifice of survival benefit. However, these recommendations are made with trepidation given the demonstrated short-term mortality implications and adequacy of oncologic resection. While prospective randomized trials could shed light on these concerns, ethical and cost considerations preclude this study design. As with all conversations regarding management of cancer, this study supports detailed communication amongst society groups and surgeons regarding costs and benefits of location of care for stage I-III colon cancer.
Footnotes
Financial Disclosures: None
References
- 1.Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relation between surgical volume and mortality. The New England journal of medicine. 1979;301(25):1364–9. [DOI] [PubMed] [Google Scholar]
- 2.Xu Z, Becerra AZ, Justiniano CF, Boodry CI, Aquina CT, Swanger AA, et al. Is the Distance Worth It? Patients With Rectal Cancer Traveling to High-Volume Centers Experience Improved Outcomes. Diseases of the colon and rectum. 2017;60(12):1250–9. [DOI] [PubMed] [Google Scholar]
- 3.Speicher PJ, Englum BR, Ganapathi AM, Wang X, Hartwig MG, D’Amico TA, et al. Traveling to a High-volume Center is Associated With Improved Survival for Patients With Esophageal Cancer. Ann Surg. 2017;265(4):743–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lidsky ME, Sun Z, Nussbaum DP, Adam MA, Speicher PJ, Blazer DG 3rd. Going the Extra Mile: Improved Survival for Pancreatic Cancer Patients Traveling to High-volume Centers. Annals of surgery. 2016. [DOI] [PubMed] [Google Scholar]
- 5.Lin CC, Bruinooge SS, Kirkwood MK, Olsen C, Jemal A, Bajorin D, et al. Association Between Geographic Access to Cancer Care, Insurance, and Receipt of Chemotherapy: Geographic Distribution of Oncologists and Travel Distance. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2015;33(28):3177–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Aquina CT, Becerra AZ, Justiniano CF, Xu Z, Boscoe FP, Schymura MJ, et al. Surgeon, Hospital, and Geographic Variation in Minimally Invasive Colectomy. Annals of surgery. 2018. [DOI] [PubMed] [Google Scholar]
- 7.Stitzenberg KB, Sigurdson ER, Egleston BL, Starkey RB, Meropol NJ. Centralization of cancer surgery: implications for patient access to optimal care. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(28):4671–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bos AC, van Erning FN, Elferink MA, Rutten HJ, van Oijen MG, de Wilt JH, et al. No Difference in Overall Survival Between Hospital Volumes for Patients With Colorectal Cancer in The Netherlands. Diseases of the colon and rectum. 2016;59(10):943–52. [DOI] [PubMed] [Google Scholar]
- 9.Siegel RL, Fedewa SA, Anderson WF, Miller KD, Ma J, Rosenberg PS, et al. Colorectal Cancer Incidence Patterns in the United States, 1974–2013. J Natl Cancer Inst. 2017;109(8). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jemal A, Ward EM, Johnson CJ, Cronin KA, Ma J, Ryerson B, et al. Annual Report to the Nation on the Status of Cancer, 1975–2014, Featuring Survival. J Natl Cancer Inst. 2017;109(9). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Huo YR, Phan K, Morris DL, Liauw W. Systematic review and a meta-analysis of hospital and surgeon volume/outcome relationships in colorectal cancer surgery. J Gastrointest Oncol. 2017;8(3):534–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chang GJ, Rodriguez-Bigas MA, Skibber JM, Moyer VA. Lymph node evaluation and survival after curative resection of colon cancer: systematic review. J Natl Cancer Inst. 2007;99(6):433–41. [DOI] [PubMed] [Google Scholar]
- 13.Khan MA, Hakeem AR, Scott N, Saunders RN. Significance of R1 resection margin in colon cancer resections in the modern era. Colorectal Dis. 2015;17(11):943–53. [DOI] [PubMed] [Google Scholar]
- 14.Lee SY, Kim CH, Kim YJ, Kim HR. Prognostic impact of the length of the longitudinal resection margin in colon cancer. Colorectal Dis. 2017;19(7):634–40. [DOI] [PubMed] [Google Scholar]
- 15.Le Voyer TE, Sigurdson ER, Hanlon AL, Mayer RJ, Macdonald JS, Catalano PJ, et al. Colon cancer survival is associated with increasing number of lymph nodes analyzed: a secondary survey of intergroup trial INT-0089. J Clin Oncol 2003;21(15):2912–9. [DOI] [PubMed] [Google Scholar]
- 16.Standard Surgical Treatments 2.0., (2015).
- 17.Nelson H, Sargent DJ, Wieand HS, Fleshman J, Anvari M, Stryker SJ, et al. A comparison of laparoscopically assisted and open colectomy for colon cancer. N Engl J Med. 2004;350(20):2050–9. [DOI] [PubMed] [Google Scholar]
- 18.Salem JF, Gummadi S, Marks JH. Minimally Invasive Surgical Approaches to Colon Cancer. Surg Oncol Clin N Am. 2018;27(2):303–18. [DOI] [PubMed] [Google Scholar]
- 19.Braga M, Vignali A, Gianotti L, Zuliani W, Radaelli G, Gruarin P, et al. Laparoscopic versus open colorectal surgery: a randomized trial on short-term outcome. Ann Surg. 2002;236(6):759–66; disscussion 67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Phibbs CS, Luft HS. Correlation of travel time on roads versus straight line distance. Medical care research and review : MCRR. 1995;52(4):532–42. [DOI] [PubMed] [Google Scholar]
- 21.McGuirk MA, Porell FW. Spatial patterns of hospital utilization: the impact of distance and time. Inquiry : a journal of medical care organization, provision and financing. 1984;21(1):84–95. [PubMed] [Google Scholar]
- 22.Williams AP, Schwartz WB, Newhouse JP, Bennett BW. How many miles to the doctor? The New England journal of medicine. 1983;309(16):958–63. [DOI] [PubMed] [Google Scholar]
- 23.Raman V, Adam MA, Turner MC, Moore HG, Mantyh CR, Migaly J. Disparity of Colon Cancer Outcomes in Rural America: Making the Case to Travel the Extra Mile. J Gastrointest Surg. 2019;23(11):2285–93. [DOI] [PubMed] [Google Scholar]